Information

Neural oscillations and activity patterns

Neural oscillations and activity patterns


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Both single neurons and groups of neurons can generate oscillatory activity spontaneously. In addition, they may show oscillatory responses to perceptual input or motor output. Even memory and thinking is associated with these neural oscillations.

My question is, are these nerve responses always oscillatory in nature? Or do we find patterns in the firing of neurons (action potentials) or in the postsynaptic potentials which are truly chaotic and do not fall under any oscillatory patterns? Or is it that because of the vast amount of inter connections between neurons as well as different sections of brain, synchronous behaviour always pops us?


Frequency analysis

All time-varying signals can be decomposed into the frequency domain, e.g. using the Fourier transform. This is a common technique for analyzing brain signals, e.g. from EEG.

However, doing this transform does not mean that the underlying signals are oscillatory in nature. Even if we talk about power in some particular frequency band, that does not mean we are talking about an oscillation at that frequency. Event-related potentials are often analyzed in the time domain, but nothing would stop you from analyzing them in the frequency domain.

Packets vs oscillations

When you write "memory and thinking is associated with these neural oscillations", this can be true of frequency analyses but yet not really related to oscillations. Artur Luczak and Ken Harris have argued for evidence of "packet-based" communication in neocortex (see for example their Luczak et al 2015 review). Effectively these are bursts of activity lasting on the order of 100-100s of milliseconds. They've shown that in anesthetized states these "packets" are revealed in a slow "oscillatory" power in the EEG/LFP, which makes sense: if you have a bunch of events that last 100-100s of milliseconds, you're going to find power in the low frequency bands. However, these are not oscillations. They are discrete events. In the wake state, the same events happen, but because there are no gaps between them, you don't see an oscillation in the EEG.

More generally, this is true of any events in frequency analysis. If you add a bunch of events lasting, say, 50 milliseconds to a white noise signal, you will see an increase in power in the vicinity of 10 Hz. They need not oscillate at all.

Separating oscillations from 'power law' power-frequency relationships

Some researchers have argued that we should be more aware of evidence for and against true "oscillations" when analyzing EEG. From Donoghue et al 2020:

Following historical traditions, the vast majority of the studies examining oscillations rely on canonical frequency bands, which are approximately defined as: infraslow (<0.1 Hz), delta (1-4 Hz), theta (4-8 Hz), alpha (8-12 Hz), beta (12-30 Hz), low gamma (30-60 Hz), high-frequency activity (60-250 Hz) and fast ripples (200-400 Hz). Although most of these bands are often described as oscillations, standard approaches fail to assess whether an oscillation-meaning rhythmic activity within a narrowband frequency range-is truly present (Fig. 1a,b).

They recommend analyzing frequency-domain signals in terms of the broadband power (dominated by the lowest frequencies), the exponent of power decay with frequency (neural signals are typically in the range of 1/f or 1/f^2, but this easily generalizes to 1/f^a), and then isolating evidence for actual oscillations that deviate from these parametric fits (linear in log-log space):

In this figure, you would only identify an oscillation in the alpha 8-12 Hz range. All of the rest of the power spectrum derives from non-oscillatory time-varying neuronal activity, which still has some detectable frequency structure.

Power-law brain activity, criticality, and chaos

Another area of emphasis in the analysis of brain signals is the idea of criticality, in particular work by John Beggs. Criticality is the border between chaotic and organized (oscillatory) activity. A prediction of this theory is exactly the power-law decay with frequency (1/f^a) described above. This scheme refers to "neuronal avalanches" which are events of synchronized neural activity that share some features with the "packets" of Luczak and Harris. Waking brains tend to be near this critical point, whereas during sleep there is much more oscillatory organization and the brain is further from critical. Although criticality can be described as "near chaos", I think it's important to recognize that it is not chaos, no matter how complicated it is. It's still quite organized, just organized in a way that optimizes information processing. Oscillations that are too organized have little capacity to transmit information (consider the information transmission limitations of radio bands that humans use for television, radio, and cell phones, for example).


Beggs, J. M. (2008). The criticality hypothesis: how local cortical networks might optimize information processing. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 366(1864), 329-343.

Donoghue, T., Haller, M., Peterson, E. J., Varma, P., Sebastian, P., Gao, R.,… & Voytek, B. (2020). Parameterizing neural power spectra into periodic and aperiodic components. Nature neuroscience, 23(12), 1655-1665.

Luczak, A., McNaughton, B. L., & Harris, K. D. (2015). Packet-based communication in the cortex. Nature Reviews Neuroscience, 16(12), 745-755.


The Human Auditory System

Luc H. Arnal , . Anne-lise Giraud , in Handbook of Clinical Neurology , 2015

Neural oscillations as endogenous temporal constraints

Neural oscillations correspond to synchronous activity of neuronal assemblies that are both intrinsically coupled and coupled by a common input. It was proposed that these oscillations reflect modulations of neuronal excitability that temporally constrain the sampling of sensory information ( Schroeder and Lakatos, 2009a ). The intriguing correspondence between the size of certain speech temporal units and the frequency of oscillations in certain frequency bands ( Fig. 5.1 ) has elicited the intuition that they might play a functional role in sensory sampling (see below). Oscillations are evidenced by means of a spectrotemporal analysis of electrophysiologic recordings (see Wang, 2010 , for a review). The requirements for measuring oscillations and spiking activity are different. The presentation of an exogenous stimulus typically results in an increase of spiking activity in those brain areas that are functionally sensitive to such inputs. Neural oscillations, on the other hand, can be observed in local field potential recordings in the absence of any external stimulation. Exogenous stimulation however typically modulates oscillatory activity, resulting either in a reset of their phase and/or a change (increase or decrease) in the magnitude of these oscillations ( Howard and Poeppel, 2012 ).

Cortical oscillations are proposed to shape spike-timing dynamics and to impose phases of high and low neuronal excitability ( Britvina and Eggermont, 2007 Schroeder and Lakatos, 2009a, b Panzeri et al., 2010 ). The assumption that it is oscillations that cause spiking to be temporally clustered derives from the observation that spiking tends to occur in specific phases (i.e., the trough) of oscillatory activity ( Womelsdorf et al., 2007 ). It is also assumed that spiking and oscillations do not reflect the same aspect of information processing. Whereas spiking reflects axonal activity, oscillations are said to reflect mostly dendritic synaptic activity ( Wang, 2010 ). While both measures are relevant to address how sensory information is encoded in the brain, we believe that the ability of neural oscillations to temporally organize spiking activity supports the functional relevance of neural oscillations to solve the discretization problem and to permit the integration of complex sensory signals across time.

Neuronal oscillations are ubiquitous in the brain, but they vary in strength and frequency depending on their location and the exact nature of their neuronal generators ( Mantini et al., 2007 Hyafil et al., 2012 ). The notion that neural oscillations shape the way the brain processes sensory information is supported by a wealth of electrophysiologic findings in humans and animals. On the one hand, stimuli that occur in the ideal excitability phase of slow oscillations (< 12 Hz) are processed faster and with a higher accuracy ( Lakatos et al., 2008 Busch et al., 2009 Henry and Obleser, 2012 Ng et al., 2012 Wyart et al., 2012 ). On the other hand, gamma-band 40-Hz activity (low gamma band) can be observed at rest in both monkey ( Fukushima et al., 2012 ) and human auditory cortex. In humans, it can be measured using EEG, MEG, and with a more precise localization with concurrent EEG and functional magnetic resonance imaging ( Morillon et al., 2010 ) and intracranial electroencephalographic recordings (stereotactic EEG (sEEG), Electro-corticography (EcoG)) in patients. Neural oscillations in this range are endogenous in the sense that one can observe a spontaneous spike clustering at approximately 40 Hz even in the absence of external stimulation. This gamma activity is thought to be generated by a “ping-pong” interaction between pyramidal cells and inhibitory interneurons ( Borgers et al., 2005, 2008 ), or even just among interneurons that are located in superficial cortical layers ( Tiesinga and Sejnowski, 2009 ). Exogenous inputs usually increase gamma-band activity in sensory areas, presumably clustering spiking activity that is propagated to higher hierarchic processing stages ( Arnal et al., 2011 Arnal and Giraud, 2012 Bastos et al., 2012 ). By analogy with the proposal of Elhilali et al. (2004) that slow responses gate faster ones, it is interesting to envisage this periodic modulation of spiking by oscillatory activity as an endogenous mechanism to optimize the extraction of relevant sensory input in time. Such integration could occur under the patterning of slower oscillations in the delta-theta range.


Résumé

Le terme d'«oscillations cérປrales ou neuronales» se rapporte à l'activité électrique rythmique et/ou répétitive générພ spontanément et en réponse aux stimuli par le tissu neuronal dans le système nerveux central. L'importance des oscillations cérປrales dans les processus cognitivo-sensoriels devient de plus en plus manifeste. Les oscillations liພs aux événements sont clairement modifiພs dans de nombreux types de pathologies neurologiques, en particulier dans le dຜlin cognitif. Cet article analyse les méthodes comme les spectres et les oscillations évoquພs/liພs à un événement, les analyses de cohérence et le blocage de phase. II donne des exemples d'application de concepts et de méthodes essentiels dans le trouble bipolaire, servant de base pour des notions fondamentales sur des biomarqueurs neurophysiologiques dans le dຜlin cognitif. Le message clé est le suivant: au cours du développement des stratégies diagnostiques et pharmacothérapeutiques, les donnພs neurophysiologiques devraient être analysພs dans un cadre utilisant de multiples méthodes et bandes de fréquence.


Methods

Participants

Forty-two participants were recruited from the student population of a German university. Participants were between 19 and 29 years old (on average 22.5 years, SD = 2.8), 36 participants were female. Participants received course credit for participation and, in addition, they were paid € 17. The study took about 2 hours. Written informed consent was obtained for all participants. For our sample size, α = 0.05 and a medium to large effect of r = 0.37, the statistical power to discover an effect if it exists in the population is (1-β) = 0.80.

Task and Measures

The n-back task consisted of three blocks with 62 trials each. On each trial, participants saw a fixation-cross, followed by a capital letter taken from a pool of eleven letters. Each trial took 2.5 seconds (see Fig. 1). Participants were instructed to press a button with their right index finger when the letter was a target, or a second button with their left index finger if the letter was a non-target. In the first block (0-back), the target was the letter “X”. In the second block (1-back), the letter from the last trial served as target. Therefore, participants had to constantly update and memorize the target. In the third and final block (2-back), the letter presented two trials ago served as target. In addition to the 1-back condition, participants had to memorize und constantly update a second target along with the information which of the two letters was the target. Half of the trials per block were targets. The same randomized sequence of targets and non-targets and the same letters were used for all participants. Performance on the n-back task was coded trial-by-trial. In less than 1% of the trials, participants did not respond within the 2-second time span therefore, these trials were excluded from the analyses regarding reaction time and hit rate as corresponding data were not available. Otherwise, responses were coded as correct if the correct button was pressed for either a target or a non-target and as non-correct if the wrong button was pressed for either a target or a non-target.

For the assessment of need for cognition, the 18-item short scale was used 25 . The construct need for cognition can be located in the Big Five personality framework, pertaining to the aspect Intellect of the domain openness to experience 32,47 . Items were presented in German language. Participants indicated on a seven point Likert scale, ranging from “do not agree at all” to “fully agree”, the extent to which each of the statements pertained to them. Half of the items are reversed scored and were recoded before aggregating across all items. In the present study, the scale had an internal consistency of α = 0.88 (descriptive statistics were M = 84.1, SD = 14.8). For Fig. 2c, a median split was performed on need for cognition, resulting in one sub-group with low (N = 22 M = 72.7 SD = 10.3 range [39 81]) and one sub-group with high (N = 20 M = 96.6 SD = 6.7 range [84 107]) values on need for cognition.

Working memory capacity was assessed by three tasks from the Wechsler Adult Intelligence Scale 48 . The tasks were manually administered by an examiner in a one-to-one situation. Examiners were blind with regard to all other study variables (e.g., need for cognition, theta power). The first task demanded the repetition of an increasing number of digits (2 to 18) the second task was a backwards-repetition of an increasing number of digits (2 to 16). Both tasks had two items on each difficulty level. The third task consisted of the sorting and repetition of an increasing number of digits and letters (2 to 8) with three items per difficulty level thereby, digits had to be given first, in ascending order, followed by alphabetically ordered letters. Within each task and each difficulty level, a termination criterion was defined specifically, if all items within a difficulty level were answered incorrectly, the task was aborted. Within each task, performance was computed as the sum of all correctly answered items. Working memory performance was estimated as the aggregate of z-standardized values in the three tasks.

All methods were carried out in accordance with the approved guidelines of the Julius Maximilians University Würzburg and all experimental protocols were approved by its ethic committee.

EEG Recording and Quantification

While participants performed the n-back task, EEG (analog bandpass: 0.1–80 Hz, sampling rate: 250 Hz) was recorded from 31 scalp sites according to the 10–20 system, using Ag/AgCl electrodes and a BrainAmpDC amplifier (Brain Products GmbH, Gilching, Germany). Impedances were kept below 10 kΩ and electrodes were referenced to the vertex (Cz). For detection of blinks and eye-movements the vertical electrooculogram (EOG) was recorded. Data were processed offline, using Brain Vision Analyzer 2.0 software (Brain Products GmbH, Gilching, Germany). First, data were filtered, using a 0.15 Hz high-pass and a 40 Hz low-pass filter (24 dB/Octave) and, additionally, a 50 Hz notch filter. Subsequently, the EEG was segmented into epochs of 3200 ms (−800 to 2600 ms, relative to the presentation of the stimulus). Afterwards, data were corrected for ocular artifacts using an Independent Component Analysis based correction method implemented in the Brain Vision Analyzer. Larger artifacts were automatically detected by a computer algorithm implemented in Brain Vision Analyzer 2.0 software and discarded if applicable, both prior and after the ocular correction. For this purpose the following exclusion criteria were applied: (1) maximal voltage difference >250 μV within 1000 ms (prior to the ocular correction, to enhance data quality for the independent component analysis without excluding trials according to eye-related activity) (2) maximal voltage difference >100 μV within 1000 ms across the epoch and maximal voltage step of 20 μV/msec (after the correction for ocular artifacts). At least 32 artifact-free trials (on average 56 trials) were available per participant and condition. Subsequently, data were re-referenced to an averaged reference across all electrodes (excluding vertical EOG).

Event related potentials were computed by segmenting the data to −500 to 2000 ms relative to the presentation of the stimulus, averaging across all trials and correcting for baseline activity (−150 to −50 ms).

For time-frequency analyses, segmented data were convolved using a family of complex Morlet wavelets from 1 to 20 Hz in linear steps of 1 Hz. The complex Morlet wavelets are defined as Gaussian-windowed complex sine functions: , with and , the latter resulting in total energy of 1 for all frequency levels 49 . Constant ratio (fof) was set 6.7 50,51 . For each frequency layer, power values were baseline corrected by subtracting the mean activity in the time window −150 to −50 ms before stimulus onset from each data point. (Please see the supplemental material for detailed information on the rational for choosing a baseline period rather close to stimulus onset and for results for an alternative baseline −450 to −350 ms before stimulus onset.) Segments were averaged for each participant and condition of interest. For analyses of FMθ activity, mean power values 4 to 7 Hz at Fz, F3, FCz, F4 and Cz were extracted from 650 to 1900 ms following the presentation of the stimulus and aggregated across these electrodes.

Heart period measurement

We used three disposable Ag/AgCl electrodes (Covidien Kendall ECG Electrodes H98LG) placed according to a modified Einthoven II lead to measure heart period, defined as the time between two adjacent R-peaks 52 . The ground electrode was placed below the left collarbone, the negative electrode below the right collarbone and the positive electrode on the left side below the rib cage. The signal was digitized using a Brain Vision BrainAmp ExG amplifier (Brain Products GmbH, Gilching, Germany) with a sampling rate of 250 Hz and the BrainVision Recorder 1.20 software (Brain Products GmbH).

To analyze the event-related heart period response we first detected the peaks of the R-waves using QRSTool 53 . Following an automatic detection, the detected beats were checked and if necessary corrected manually. Next, the times of the R-peaks were exported and event-related inter-beat-intervals were extracted using a custom-built Matlab script. For every stimulus in the n-back task we extracted the inter-beat-interval surrounding the stimulus as well as the two following intervals. Heart period was defined as the interval in milliseconds between sequential R-waves and was estimated for the interval following the stimulus.

High-frequency heart period variability (HF-HPV) was estimated using the software ARTiiFACT 54 . Based on the timing of the R-peaks exported from QRSTool we computed separated inter-beat-interval series for every condition of the n-back task per participant (mean length of the series across all participants and conditions: 154.65 seconds). These inter-beat-interval series were then submitted to a frequency domain analysis of HF-HPV in ARTiiFACT. Except for the width of the Hanning window (153 seconds), standard settings of the Fast Fourier Transformation (FFT) were used (4 Hz spline interpolation, 50% overlap of resampled and detrended data). The absolute values (in ms 2 ) of the HF-HPV component (0.15–0.40 Hz) were used for statistical analysis 55 . Since those values were not normally distributed, we transformed them using the natural logarithm before running the statistical analysis.

Statistical analyses

Latent growth curve analysis 23 with maximum likelihood function was performed using AMOS 23.0 with bootstrapping to account for the medium sample size. Latent growth curve analysis is a special case of multilevel modeling and captures the pattern of change across multiple sampling points. For each individual, a linear function described by slope and intercept was estimated. As depicted in Fig. 4a, the intercept was specified by a latent variable with regression weights fixed to 1 between the latent intercept and the three manifest variables indicating theta power in the three conditions. The slope is specified by a latent variable indicated by regression weights fixed to −2, 1 and 0 for theta power in the 0-back, 1-back and 2-back condition, respectively. Therefore, the intercept specifies predicted theta power in the 2-back condition.

Need for cognition was modeled as latent variable indicated by three manifest variables the latter were computed by randomly assigning each of the 18 items (see above) to one of three parcels and subsequently averaging across these items. Working memory was modeled as a latent variable with the z-standardized values of each of the three working memory tasks as indicators.

Factor score weights were used to predict the estimated slope for each subject and each condition for subsequent analyses. We used mixed models for binomial data with a LOGIT link function for the prediction of trial-by-trial performance in the n-back task, using the GENLINMIXED procedure in SPSS 23.0.


Role in neural syntax

Brain oscillations can segregate and group neuronal activity to decompose and package neuronal information for communication between brain areas. Because all neuronal oscillations are based on inhibition, they can parse and concatenate neuronal messages, a prerequisite for any coding mechanism. Gamma waves combine neurons into assemblies, which can be conceived as a neuronal “letter.” The hierarchical nature of cross-frequency-coupled rhythms can serve as a mechanism for combining neuronal letters into neuronal words and words into sentences, so that unbounded combinatorial information can be generated from spike patterns. Neuronal oscillators are pulsatile and can readily entrain one another, facilitating the effectiveness of message exchange between brain areas.


Submitted

Kragel, J. E., Worrell, G. A., Sperling, M. R., Gross, R. E., Lega, B. C., Jobst, B. C., et al. Distinct cortical systems reinstate content and context information during memory search. Submitted.
(pdf, supplemental, code, Ephys data)

Sakon, J. J. and Kahana, M. J. Hippocampal ripples signal contextually-mediated episodic recall. bioRxiv.
(pdf, code, Spectrograms)

Adamovich-Zeitlin, R., Wanda, P. A., Solomon, E., Phan, T., Lega, B., Jobst, B. C., et al. (2020). Biomarkers of memory variability in traumatic brain injury. Brain Communications, 3(1).
(code, Ephys data)

Meisler, S. L., Kahana, M. J., and Ezzyat, Y. (2020). Does data cleaning improve brain state classification? Journal of Neuroscience Methods.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2020). Neural measures of subsequent memory reflect endogenous variability in cognitive function. Journal of Experimental Psychology: Learning, Memory, and Cognition.
(pdf, code, Ephys data)

Solomon, E. A., Lega, B. C., Sperling, M. R., and Kahana, M. J. (2019). Hippocampal theta codes for distances in semantic and temporal spaces. PNAS, 116(48), 24343-24352.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2019). Dynamics of brain activity reveal a unitary recognition signal. Journal of Experimental Psychology: Learning, Memory, and Cognition, 45(3), 440�.
(pdf, code, Ephys data)

Weidemann, C. T., Kragel, J. E., Lega, B. C., Worrell, G. A., Sperling, M. R., Sharan, A. D., et al. (2019). Neural activity reveals interactions between episodic and semantic memory systems during retrieval. Journal of Experimental Psychology: General, 148(1), 1󈝸.
(pdf, code, Ephys data)

Ezzyat, Y., Kragel, J. E., Burke, J. F., Levy, D. F., Lyalenko, A., Wanda, P., et al. (2017). Direct brain stimulation modulates encoding states and memory performance in humans. Current Biology, 27(9), 1251�.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2016). Assessing recognition memory using confidence ratings and response times. Royal Society Open Science, 3, 150670.
(pdf, code, Ephys data)

Long, N. M. and Kahana, M. J. (2015). Successful memory formation is driven by contextual encoding in the core memory network. NeuroImage, 119, 332-337.
(pdf, code, Ephys data)

Merkow, M. B., Burke, J. F., and Kahana, M. J. (2015). The human hippocampus contributes to both the recollection and familiarity components of recognition memory. Proceedings of the National Academy of Sciences, 112(46), 14378�.
(pdf, Ephys data)

Ramayya, A. G., Pedisich, I., and Kahana, M. J. (2015). Expectation modulates neural representations of valence throughout the human brain. NeuroImage, 115, 214-223.
(pdf, code, Ephys data)

Geller, A. S., Burke, J. F., Sperling, M. R., Sharan, A. D., Litt, B., Baltuch, G. H., et al. (2014). Eye closure causes widespread low-frequency power increase and focal gamma attenuation in the human electrocorticogram. Clinical Neurophysiology, 125(9), 1764-73.
(pdf, Ephys data)

Long, N. M., Burke, J. F., and Kahana, M. J. (2014). Subsequent memory effect in intracranial and scalp EEG. NeuroImage, 84, 488�.
(pdf, Ephys data)

Ramayya, A. G., Misra, A., Baltuch, G. H., and Kahana, M. J. (2014). Microstimulation of the human substantia nigra following feedback alters reinforcement learning. Journal of Neuroscience, 34(20), 6887�.
(pdf, Ephys data)

Miller, J. F., Neufang, M., Solway, A., Brandt, A., Trippel, M., Mader, I., et al. (2013). Neural activity in human hippocampal formation reveals the spatial context of retrieved memories. Science, 342(6162), 1111-1114.
(pdf, supplemental, Ephys data)

Morton, N. W., Kahana, M. J., Rosenberg, E. A., Sperling, M. R., Sharan, A. D., and Polyn, S. M. (2013). Category-specific neural oscillations predict recall organization during memory search. Cerebral Cortex, 23(10), 2407�.
(pdf, Ephys data)

van Vugt, M. K., Sekuler, R., Wilson, H. R., and Kahana, M. J. (2013). An electrophysiological signature of summed similarity in visual working memory. Journal of Experimental Psychology: General, 142(2), 412�.
(pdf, Ephys data)

van der Meij, R., Kahana, M. J., and Maris, E. (2012). Phase-amplitude coupling in human ECoG is spatially distributed and phase diverse. Journal of Neuroscience, 32(1), 111-123.
(pdf, Ephys data)

Maris, E., van Vugt, M. K., and Kahana, M. J. (2011). Spatially distributed patterns of oscillatory coupling between high-frequency amplitudes and low-frequency phases in human ieeg. NeuroImage, 54(2), 836-850.
(pdf, Ephys data)

van Vugt, M. K., Schulze-Bonhage, A., Litt, B., Brandt, A., and Kahana, M. J. (2010). Hippocampal gamma oscillations increase with working memory load. Journal of Neuroscience, 30(7), 2694�.
(pdf, Ephys data)

Jacobs, J. and Kahana, M. J. (2009). Neural representations of individual stimuli in humans revealed by gamma-band electrocorticographic activity. Journal of Neuroscience, 29(33), 10203�.
(pdf, Ephys data)

van Vugt, M. K., Schulze-Bonhage, A., Sekuler, R., Litt, B., Brandt, A., Baltuch, G., et al. (2009). Intracranial electroencephalography reveals two distinct similarity effects during item recognition. Brain Research, 1299, 33󈞘.
(pdf, Ephys data)

Weidemann, C. T., Mollison, M. V., and Kahana, M. J. (2009). Electrophysiological correlates of high-level perception during spatial navigation. Psychonomic Bulletin & Review, 16(2), 313�.
(pdf, Ephys data)

Jacobs, J., Hwang, G., Curran, T., and Kahana, M. J. (2006). EEG oscillations and recognition memory: Theta correlates of memory retrieval and decision making. NeuroImage, 15(2), 978󈟃.
(pdf, Ephys data)

Rizzuto, D., Madsen, J. R., Bromfield, E. B., Schulze-Bonhage, A., and Kahana, M. J. (2006). Human neocortical oscillations exhibit theta phase differences between encoding and retrieval. NeuroImage, 31(3), 1352�.
(pdf, Ephys data)

Ekstrom, A. D., Caplan, J., Ho, E., Shattuck, K., Fried, I., and Kahana, M. (2005). Human hippocampal theta activity during virtual navigation. Hippocampus, 15, 881�.
(pdf, Ephys data)

Howard, M. W., Rizzuto, D. S., Caplan, J. C., Madsen, J. R., Lisman, J., Aschenbrenner-Scheibe, R., et al. (2003). Gamma oscillations correlate with working memory load in humans. Cerebral Cortex, 13, 1369�.
(pdf, Ephys data)

Rizzuto, D., Madsen, J. R., Bromfield, E. B., Schulze-Bonhage, A., Seelig, D., Aschenbrenner-Scheibe, R., et al. (2003). Reset of human neocortical oscillations during a working memory task. Proceedings of the National Academy of Sciences, USA, 100(13), 7931�.
(pdf, Ephys data)

Sederberg, P. B., Kahana, M. J., Howard, M. W., Donner, E. J., and Madsen, J. R. (2003). Theta and gamma oscillations during encoding predict subsequent recall. Journal of Neuroscience, 23(34), 10809�.
(pdf, Ephys data)

Raghavachari, S., Kahana, M. J., Rizzuto, D. S., Caplan, J. B., Kirschen, M. P., Bourgeois, B., et al. (2001). Gating of human theta oscillations by a working memory task. Journal of Neuroscience, 21(9), 3175-3183.
(pdf, Ephys data)


Contents

Gamma waves can be detected by electroencephalography or magnetoencephalography. One of the earliest reports of gamma wave activity was recorded from the visual cortex of awake monkeys. [7] Subsequently, significant research activity has concentrated on gamma activity in visual cortex. [8] [9] [10] [11]

Gamma activity has also been detected and studied across premotor, parietal, temporal, and frontal cortical regions [12] Gamma waves constitute a common class of oscillatory activity in neurons belonging to the cortico-basal ganglia-thalamo-cortical loop. [13] Typically, this activity is understood to reflect feedforward connections between distinct brain regions, in contrast to alpha wave feedback across the same regions. [14] Gamma oscillations have also been shown to correlate with the firing of single neurons, mostly inhibitory neurons, during all states of the wake-sleep cycle. [15] Gamma wave activity is most prominent during alert, attentive wakefulness. [13] However, the mechanisms and substrates by which gamma activity may help to generate different states of consciousness remain unknown.

Controversy Edit

Some researchers contest the validity or meaningfulness of gamma wave activity detected by scalp EEG, because the frequency band of gamma waves overlaps with the electromyographic frequency band. Thus, gamma signal recordings could be contaminated by muscle activity. [16] Studies utilizing local muscle paralysis techniques have confirmed that EEG recordings do contain EMG signal, [17] [18] and these signals can be traced to local motor dynamics such as saccade rate [19] or other motor actions involving the head. Advances in signal processing and separation, such as the application of independent component analysis or other techniques based on spatial filtering, have been proposed to reduce the presence of EMG artifacts. [16]

Conscious perception Edit

Gamma waves may participate in the formation of coherent, unified perception, also known as the problem of combination in the binding problem, due to their apparent synchronization of neural firing rates across distinct brain regions. [20] [21] [22] 40-Hz gamma waves were first suggested to participate in visual consciousness in 1988, [23] .e.g, two neurons oscillate synchronously (though they are not directly connected) when a single external object stimulates their respective receptive fields. Subsequent experiments by many others demonstrated this phenomenon in a wide range of visual cognition. In particular, Francis Crick and Christof Koch in 1990 [24] argued that there is a significant relation between the binding problem and the problem of visual consciousness and, as a result, that synchronous 40 Hz oscillations may be causally implicated in visual awareness as well as in visual binding. Later the same authors expressed skepticism over the idea that 40-Hz oscillations are a sufficient condition for visual awareness. [25]

A number of experiments conducted by Rodolfo Llinás supports a hypothesis that the basis for consciousness in awake states and dreaming is 40-Hz oscillations throughout the cortical mantle in the form of thalamocortical iterative recurrent activity. In two papers entitled "Coherent 40-Hz oscillation characterizes dream state in humans” (Rodolfo Llinás and Urs Ribary, Proc Natl Acad Sci USA 90:2078-2081, 1993) and "Of dreaming and wakefulness” (Llinas & Pare, 1991), Llinás proposes that the conjunction into a single cognitive event could come about by the concurrent summation of specific and nonspecific 40-Hz activity along the radial dendritic axis of given cortical elements, and that the resonance is modulated by the brainstem and is given content by sensory input in the awake state and intrinsic activity during dreaming. According to Llinás’ hypothesis, known as the thalamocortical dialogue hypothesis for consciousness, the 40-Hz oscillation seen in wakefulness and in dreaming is proposed to be a correlate of cognition, resultant from coherent 40-Hz resonance between thalamocortical-specific and nonspecific loops. In Llinás & Ribary (1993), the authors propose that the specific loops give the content of cognition, and that a nonspecific loop gives the temporal binding required for the unity of cognitive experience.

A lead article by Andreas K. Engel et al. in the journal Consciousness and Cognition (1999) that argues for temporal synchrony as the basis for consciousness, defines the gamma wave hypothesis thus: [26]

The hypothesis is that synchronization of neuronal discharges can serve for the integration of distributed neurons into cell assemblies and that this process may underlie the selection of perceptually and behaviorally relevant information.

Attention Edit

The suggested mechanism is that gamma waves relate to neural consciousness via the mechanism for conscious attention:

The proposed answer lies in a wave that, originating in the thalamus, sweeps the brain from front to back, 40 times per second, drawing different neuronal circuits into synch with the precept [sic], and thereby bringing the precept [sic] into the attentional foreground. If the thalamus is damaged even a little bit, this wave stops, conscious awarenesses do not form, and the patient slips into profound coma. [21]

Thus the claim is that when all these neuronal clusters oscillate together during these transient periods of synchronized firing, they help bring up memories and associations from the visual percept to other notions. This brings a distributed matrix of cognitive processes together to generate a coherent, concerted cognitive act, such as perception. This has led to theories that gamma waves are associated with solving the binding problem. [20]

Gamma waves are observed as neural synchrony from visual cues in both conscious and subliminal stimuli. [27] [28] [29] [30] This research also sheds light on how neural synchrony may explain stochastic resonance in the nervous system. [31]

Mood disorders Edit

Altered gamma wave activity is associated with mood disorders such as major depression or bipolar disorder and may be a potential biomarker to differentiate between unipolar and bipolar disorders. For example, human subjects with high depression scores exhibit differential gamma signaling when performing emotional, spatial, or arithmetic tasks. Increased gamma signaling is also observed in brain regions that participate in the default mode network, which is normally suppressed during tasks requiring significant attention. Rodent models of depression-like behaviors also exhibit deficient gamma rhythms. [32]

Schizophrenia Edit

Decreased gamma-wave activity is observed in schizophrenia. Specifically, the amplitude of gamma oscillations is reduced, as is the synchrony of different brain regions involved in tasks such as visual oddball and Gestalt perception. People with schizophrenia perform worse on these behavioral tasks, which relate to perception and continuous recognition memory. [33] The neurobiological basis of gamma dysfunction in schizophrenia is thought to lie with GABAergic interneurons involved in known brain wave rhythm-generating networks. [34] Antipsychotic treatment, which diminishes some behavioral symptoms of schizophrenia, does not restore gamma synchrony to normal levels. [33]

Epilepsy Edit

Gamma oscillations are observed in the majority of seizures [5] and may contribute to their onset in epilepsy. Visual stimuli such as large, high-contrast gratings that are known to trigger seizures in photosensitive epilepsy also drive gamma oscillations in visual cortex. [35] During a focal seizure event, maximal gamma rhythm synchrony of interneurons is always observed in the seizure onset zone, and synchrony propagates from the onset zone over the whole epileptogenic zone. [36]

Alzheimer's disease Edit

Enhanced gamma band power and lagged gamma responses have been observed in patients with Alzheimer's disease (AD). [4] [37] Interestingly, the tg APP-PS1 mouse model of AD exhibits decreased gamma oscillation power in the lateral entorhinal cortex, which transmits various sensory inputs to the hippocampus and thus participates in memory processes analogous to those affected by human AD. [38] Decreased hippocampal slow gamma power has also been observed in the 3xTg mouse model of AD. [39]

Gamma stimulation may have therapeutic potential for AD and other neurodegenerative diseases. Optogenetic stimulation of fast-spiking interneurons in the gamma wave frequency range was first demonstrated in mice in 2009. [40] Entrainment or synchronization of hippocampal gamma oscillations and spiking to 40 Hz via non-invasive stimuli in the gamma frequency band, such as flashing lights or pulses of sound, [3] reduces amyloid beta load and activates microglia in the well-established 5XFAD mouse model of AD. [41] Subsequent human clinical trials of gamma band stimulation have shown mild cognitive improvements in AD patients who have been exposed to light, sound, or tactile stimuli in the 40 Hz range. [1] However, the precise molecular and cellular mechanisms by which gamma band stimulation ameliorates AD pathology is unknown.

Fragile X syndrome Edit

Hypersensitivity and memory deficits in Fragile X syndrome may be linked to gamma rhythm abnormalities in sensory cortex and hippocampus. For example, decreased synchrony of gamma oscillations has been observed in auditory cortex of FXS patients. The FMR1 knockout rat model of FXS exhibits an increased ratio of slow (

High-amplitude gamma wave synchrony can be self-induced via meditation. Long-term practitioners of meditation such as Tibetan Buddhist monks exhibit both increased gamma-band activity at baseline as well as significant increases in gamma synchrony during meditation, as determined by scalp EEG. [2] fMRI on the same monks revealed greater activation of right insular cortex and caudate nucleus during meditation. [42] The neurobiological mechanisms of gamma synchrony induction are thus highly plastic. [43] This evidence may support the hypothesis that one's sense of consciousness, stress management ability, and focus, often said to be enhanced after meditation, are all underpinned by gamma activity. At the 2005 annual meeting of the Society for Neuroscience, the current Dalai Lama commented that if neuroscience could propose a way to induce the psychological and biological benefits of meditation without intensive practice, he "would be an enthusiastic volunteer." [44]


Neuronal firing patterns outweigh circuitry oscillations in parkinsonian motor control

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Find articles by Pan, M. in: JCI | PubMed | Google Scholar | />

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Find articles by Lai, W. in: JCI | PubMed | Google Scholar | />

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Published October 31, 2016 - More info

Neuronal oscillations at beta frequencies (20–50 Hz) in the cortico-basal ganglia circuits have long been the leading theory for bradykinesia, the slow movements that are cardinal symptoms in Parkinson’s disease (PD). The beta oscillation theory helped to drive a frequency-based design in the development of deep brain stimulation therapy for PD. However, in contrast to this theory, here we have found that bradykinesia can be completely dissociated from beta oscillations in rodent models. Instead, we observed that bradykinesia is causatively regulated by the burst-firing pattern of the subthalamic nucleus (STN) in a feed-forward, or efferent-only, mechanism. Furthermore, STN burst-firing and beta oscillations are two independent mechanisms that are regulated by different NMDA receptors in STN. Our results shift the understanding of bradykinesia pathophysiology from an interactive oscillatory theory toward a feed-forward mechanism that is coded by firing patterns. This distinct mechanism may improve understanding of the fundamental concepts of motor control and enable more selective targeting of bradykinesia-specific mechanisms to improve PD therapy.

A cardinal feature of Parkinson’s disease (PD) is slow movements, known as bradykinesia. The neuronal activities related to bradykinesia are two electrophysiological landmarks in PD: oscillations, the pathological augmentation of cerebral field activities in beta frequencies (20–50 Hz) between cortex and subthalamic nucleus (STN) ( 1 – 9 ), versus codes, the excessive burst-firing patterns in STN ( 10 – 13 ). The leading hypothesis has long been that beta oscillations underlie bradykinesia, supported by the fact that beta power correlates with bradykinesia severity ( 6 – 9 ) and injecting beta electric activities into cortex ( 14 , 15 ) and STN ( 13 , 16 ) worsens motor performances. The oscillatory theory has deeply impacted PD therapy development and has served as important conceptual basis for deep brain stimulation (DBS) ( 16 – 18 ).

However, we recently found that excessive STN bursts, the abnormal codes in PD, can also lead to bradykinesia ( 10 , 11 ). The generation of STN bursts requires T-type calcium channels (CaTs), which are the intrinsic ion channels in STN serving as burst initiator ( 19 ). The cortex regulates STN bursts via NMDAergic cortico-subthalamic transmission ( 12 ), which also generates beta oscillations ( 12 ). The new understandings of the online modulatory mechanism of both STN bursts and beta oscillations open the window to approaching the fundamental question, what is the mechanism directly responsible for bradykinesia: the frequency-dependent oscillations or STN bursting codes? We therefore applied online modulations by selectively manipulating STN bursts and beta oscillations in 6-hydroxydopamine (6-OHDA) hemiparkinsonian rat models ( 10 – 13 , 20 ) and investigated their effects on bradykinesia (Figure 1, Supplemental Figure 1, and Supplemental Video 1 supplemental material available online with this article doi:10.1172/JCI88170DS1).

Behavioral and real-time neuronal abnormalities in 6-OHDA–lesioned parkinsonian rat model. (A) Scheme of the experimental design. Rats received surgical placements of drug infusion cannula coupled with a stimulating electrode, and recording electrodes in STN and cortex. Implanted rats received part or all of the following evaluations, including locomotor behaviors (open-field free movements and rotarod forced movements), single-unit recordings, and LFPs, before and after pharmacological and/or electrical manipulations. inj., injection Str, striatum GP, globus pallidus SNc, substantia nigra pars compacta. (B and C) Locomotor behaviors. 6-OHDA rats developed motor deficits, especially slow movements (bradykinesia), in both (B) free-moving and (C) forced-moving paradigms (n = 11 in both paradigms). (D) STN firing patterns. 6-OHDA rats developed excessive burst firings in STN, while the intra-burst profiles remained unchanged (n =10). (EG) Oscillatory profiles. (E and F) In situ synchronization of oscillatory activities presented as LFPs. STN and cortical power in beta frequencies (20–50 Hz) were pathologically increased in both resting and moving conditions in 6-OHDA rats (n = 11). (G) Long-range cortico-subthalamic oscillations presented by time-coherence plot. 6-OHDA rats developed robust oscillations in beta frequencies (n = 11). Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM **P < 0.01.

Modulating intrinsic burst-firing properties of STN dissociates beta oscillations from bradykinesia. Beta oscillations and STN bursts, the two bradykinesia-generating candidates, either can work synchronously in a cascade, or one of them is an epiphenomenon. We first manipulated STN bursts while observing beta oscillations and bradykinesia. Taking advantage of our previous studies ( 10 , 11 ), we suppressed or facilitated STN bursts by manipulating CaTs. Applying CaT blockers (NiCl2 and mibefradil) into STN suppressed STN bursts and also remedied bradykinesia in 6-OHDA hemiparkinsonian rats (Figure 2, A–F, Supplemental Figure 2, and Supplemental Video 2). However, the oscillatory profiles remained unchanged, including in situ beta synchronization of STN and cortex (Figure 2, G and H) and cortico-subthalamic oscillations (Figure 2, I and J) in both resting and moving conditions. These results clearly demonstrated that STN bursts do not cause bradykinesia via beta oscillations, and oscillation and bradykinesia could be dissociated in PD models. Consistently, application of constant hyperpolarizing current (HC) into STN can increase CaT availability ( 10 ) and burst discharges (Figure 3, A and B), which sufficiently recapitulated bradykinesia in normal rats without generating beta oscillations (Figure 3, C–J). Instead, HC further suppressed regional beta power in STN (Figure 3G), suggesting that HC augments automaticity of individual STN neurons, and therefore weakens synchronization between nearby neurons and suppresses beta power. These results provided direct evidence that STN bursts, not beta oscillations, are the immediate mechanism of bradykinesia. STN bursts are either the downstream player in the bradykinesia-generating cascade or an isolated mechanism independent of beta oscillations.

Suppression of burst-generating capacity in STN rescues bradykinesia but not beta oscillations in 6-OHDA–lesioned hemiparkinsonian rats. (A) Subthalamic infusion of NiCl2 (Ni 2+ ), a T-type calcium channel blocker, in 6-OHDA rats. (B) Sample sweeps of single-unit recordings and quantitative burst analysis. Ni 2+ suppressed burst firings without changing the intra-burst profiles in 6-OHDA rats (n = 25). (CF) Behavioral assessments. (C) Typical traces showing Ni 2+ effects in free-moving activities. Ni 2+ ameliorated (D) motion difficulties and (E) asymmetries (n = 9). Note that moving velocity was rescued in both (D) free-moving and (F) forced-moving (n =8) paradigms. (GJ) Oscillatory profiles. (G and H) In situ synchronization. Ni 2+ had no effect on STN or cortical powers in beta frequency (20-50 Hz) in both rest and moving conditions. Quantitative analysis in these rats showed no change in beta power (bar plots, n = 11). (I) Long-range cortico-subthalamic oscillations. Dark gray section of the bar above indicates Ni 2+ infusion. (J) Quantitative analysis of coherence shows that Ni 2+ did not change the pathological state of interlocking power (right panel) or frequency (left panel) in beta ranges. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P < 0.05, **P < 0.01.

Augmentation of burst-generating propensity in STN recapitulates bradykinesia but not beta oscillations in normal rats. (A) Scheme showing subthalamic application of constant HCs in normal rats. (B) HC increased burst rates in STN (n = 15) and left intra-burst profiles unchanged. (CF) Behavioral assessments. (C) Typical free-moving traces showing HC effects. HC transformed normal rats into hemiparkinsonian states and recapitulated (D) motion difficulties and (E) asymmetries (n = 11). The capacity of fast movements was also compromised in (F) the forced-moving paradigm (n = 6). (GJ) Oscillatory profiles. (G) HC further suppressed LFPs in STN, instead of reinforcing beta power mimicking the parkinsonian state (n = 13). (H) HC also remotely suppressed cortical beta power. (I and J) The cortico-subthalamic oscillations remained unsynchronized. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P <0.05, **P < 0.01, ***P < 0.001.

Bursting codes and beta oscillations are mediated by different NMDA receptor subtypes in STN. We then examined whether beta oscillations are the upstream regulator of STN bursts in the bradykinesia-generating cascade. Beta oscillations depend on NMDAergic cortico-subthalamic transmission ( 12 ) we thus selectively inhibited NMDA receptor (NMDAR) containing the GluN2A subunit, which has the fastest kinetics in the beta range ( 21 ). Subthalamic application of CPP, a GluN2A antagonist, markedly suppressed oscillatory profiles in 6-OHDA rats but had no effect on either bradykinesia or STN bursts (Figure 4, A–D, and Supplemental Figure 3, A–D). The results directly showed that beta oscillations are not involved in the genesis of STN bursts. Therefore, beta oscillations are not the upstream regulator in the bradykinesia-generating cascade. We then examined the roles of the GluN2B and GluN2D subunits. Blockers (Ro 25-6891 [RO] and PPDA) of the GluN2B/D receptors specifically suppressed STN bursts and ameliorated bradykinesia, but oscillatory profiles remained unchanged (Figure 4, Supplemental Figures 3 and 4, and Supplemental Video 3). These results demonstrated that bursting codes and beta oscillations are two parallel mechanisms regulated by different NMDARs in STN. In striking contrast to the oscillatory hypothesis, beta oscillations are not even involved in the bradykinesia-generating cascade and STN bursts play an independent role in bradykinesia.

Differential contributions of NMDAR subtypes in bradykinesia and electrophysiological profiles of 6-OHDA rats. (AD) CPP, a selective GluN2A blocker, had no effect on (A) locomotor behaviors (n = 18) or (B) STN burst firings (n =29), but dramatically reduced both (C) in situ synchronization and (D) cortico-subthalamic oscillations (n = 9). (EH) RO, a selective GluN2B/D blocker, preferentially inhibiting GluN2B subunit, rescued (E) parkinsonian motor deficits (n = 13), suppressed (F) pathological bursts in STN (n =38), but had no effect on (G) in situ synchronization or (H) cortico-subthalamic oscillations (n = 11). Also refer to Supplemental Figure 3 for additional behavioral and single-unit profiles. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P < 0.05, **P < 0.01.

Bradykinesia is regulated by STN bursts via a feed-forward mechanism. We next investigated the mechanisms by which STN bursts could lead to bradykinesia. Beyond beta oscillations, STN bursts may still involve other forms of frequency-dependent mechanisms. Without the phase synchronicity of nearby neurons detected as beta oscillations, individual STN neurons may still require regular NMDAergic inputs to generate meaningful bursts and thus bradykinesia. To evaluate whether transsynaptic regularity modulates bradykinesia, we optogenetically activated cortico-subthalamic axonal terminals by illuminating STN in Thy1-ChR2 transgenic mice ( 12 , 13 ) (Figure 5A) with either fixed-frequency (10 Hz) stimulation or frequency-independent shuffles (Figure 5B). The two stimulation protocols were of the same stimulation loads (10 pulses/s) and generated robust and similar motor deficits in normal mice (Figure 5, C–I, Supplemental Figure 5, and Supplemental Video 4). The results indicated that motor inhibition is not only dissociated from low-frequency oscillations of grouped neuronal activities, but also independent of the regularity of action potentials transmitted in the cortico-subthalamic axons.

Bradykinesia is independent of the regularity of cortico-subthalamic transmissions. (A) Schematic illustration of fiber optic cannula implanted in STN for the stimulation of cortico-subthalamic axonal terminals in Thy1-ChR2 mice. (B) Illustration of two illumination protocols: fixed 10 Hz stimulation and randomized frequency shuffles with the same stimulation loads (10 pulses/s). (CI) Sample traces of locomotor behaviors and corresponding statistic results, showing that fixed-frequency and randomized (Rand) stimulation recapitulated bradykinesia with similar severity, quantified by (E–G) motion difficulties and (H and I) asymmetries (n = 4). Also refer to Supplemental Figure 5 for thermodynamic controls, which followed the same protocols with non-activating yellow light (589 nm) laser. Statistical analyses were performed using 1-way ANOVA with post-hoc Bonferroni correction. Data are presented as mean ± SEM *P < 0.05, **P < 0.01, ***P < 0.001.

The timing of shuffled illumination was completely artificial and unpredictable, and therefore minimized the opportunity for the circuitry to adapt from the feedback interaction. Our results strongly suggested that once STN bursts are generated, the circuitry passes this information via a feed-forward mechanism to the downstream nuclei and no longer requires continuous NMDAergic monitoring from cortex to STN. To test this hypothesis, we gave HC to generate STN bursts while simultaneously blocking NMDAergic cortico-subthalamic transmission (Figure 6A). HC sufficiently induced STN bursts and bradykinesia in normal animals, but the bradykinesia-generating effect of HC could not be rescued by interrupting NMDAergic cortico-subthalamic transmission (Figure 6, B–D, and Supplemental Figure 6A). The same principles also apply to 6-OHDA rats, which had excessive STN bursts (Figure 1D and Supplemental Figure 7A) ( 10 – 12 ). Blocking NMDAergic transmission in STN suppressed bursts and rescued bradykinesia. However, HC sufficiently restored STN bursts and eliminated the therapeutic effects of NMDAR blockers (Figure 6, E–H, and Supplemental Figure 7). Once more, animal behaviors in both normal and 6-OHDA rats only aligned with STN firing patterns. Oscillatory profiles either remained unchanged or were contrary to the profiles that would be predicted by the current oscillatory theory (Supplemental Figures 6 and 7). These results clearly indicated that once STN bursts are generated, bradykinesia no longer requires ongoing cortical regulation, regardless of its synchronicity (Figures 2–4), regularity (Figure 5), or continuity (Figure 6). These characteristics reveal a delicate feed-forward role of STN bursts in bradykinesia, and are compatible with the fast-acting and quick-responsive nature of motor execution. They are also consistent with the race model of basal ganglia circuity in the STN–substantia nigra pars reticulata (SNr) axis ( 22 ), which is immediately downstream from the STN. The interruption of planned actions in the STN/SNr axis has a critical gate of timing, and intervention beyond this time point fails to stop motor execution ( 22 ). In contrast to the feed-forward mechanism shown in bradykinesia, beta oscillations have a feedback nature that requires continuous reciprocal interactions in the circuitry. Inhibition of GluN2A transmission at the cortico-subthalamic terminals in STN sufficiently disrupted beta power in the “upstream” cortex (Figure 4C). Consistently, HC suppressed beta power in STN locally and also remotely in the cortex (Figure 3, G and H).

Bradykinesia is independent of the continuity of NMDAergic cortico-subthalamic transmissions. (A) Schematic illustration of HC application in STN of a normal rat, with or without simultaneous microinfusion of AP5, a nonselective NMDAR blocker. (BD) Sample traces and quantitative analysis of locomotor activities. Bradykinesia can be recapitulated in normal rats by subthalamic HC application, but additional NMDAergic cortico-subthalamic interruption cannot reverse bradykinesia (n = 12). (EH) Similar settings in 6-OHDA rats, showing that the therapeutic effect of NMDAergic interruption by AP5 can be abolished by additional HC application in STN (n = 15). Statistical analyses were performed using 1-way ANOVA with post-hoc Bonferroni correction. Data are presented as mean ± SEM *P < 0.05, **P < 0.01, ***P < 0.001.

Taken together, our results show that STN bursts control bradykinesia via a feed-forward mechanism. Nevertheless, it should be noted that we focused on the fastest cortico-subthalamic “hyperdirect” pathway in this study. It is evident that the slower indirect pathway eventually gets involved and tunes the motor behaviors in the later steps ( 20 ). It should also be noted that this study targeted bradykinesia, a cardinal involuntary movement in PD. Although not related to bradykinesia, beta oscillations seemed to reflect the volitional aspect of motor decision states (moving versus resting states Figures 2 and 4), regardless of the motor performances with or without NMDAR/CaT modulations (see Discussion).

We discovered that bradykinesia is regulated by STN bursting codes in a feed-forward mechanism and can be completely dissociated from beta oscillations. STN bursts and beta oscillations are two parallel mechanisms controlled by different NMDARs in STN. In this study, we quantified bradykinesia (slow movements) by velocity measurement, which has been reliably used in other studies in the 6-OHDA model ( 20 , 23 – 26 ). Nevertheless, there are other behavioral tests linked to bradykinesia and other motor deficits in PD ( 27 , 28 ), and these deserve further investigation.

Differential distributions of NMDARs and CaTs in STN and their potential impacts. The burst-generating cascade requires the collaboration between GluN2B/D NMDARs and CaTs in STN, while beta oscillations depend on GluN2A NMDARs. Consistent with the clear-cut dissociation in electrophysiology, we found differential distributions of NMDAR subtypes in the STN neurons of 6-OHDA rats (Supplemental Figure 8). The oscillation-contributing GluN2A subunits were diffusely expressed in STN soma (Supplemental Figure 8, A–C), while the burst-generating GluN2B/D subunits had punctal expression extended to STN neurites (Supplemental Figure 8, D–I). Also, both GluN2B/D and CaTs (e.g., CaV3.3, the predominant CaT subtype in STN) ( 29 ) had the characteristic punctal pattern and distribution (Supplemental Figure 9), supporting their collaborative role in burst generation. It is interesting that NMDAR subtypes and CaTs are differentially segregated according to their burst- or oscillation-generating roles. Beta oscillations have been linked to the tightly time-locked STN firings in response to NMDAergic cortico-subthalamic transmission ( 12 ). The fast kinetics of GluN2A NMDARs and their prominent somatic expression could provide better temporal precision and larger positive currents near the axon hillock and thus facilitate the time-locked STN responses ( 12 ). This kinetic profile may also partly explain why the oscillations fall into beta frequencies, which are the frequency range of GluN2A kinetics ( 12 , 21 ). Bursting cells such as STN neurons and Purkinje cells (PCs) were shown to have their CaT currents initiated in dendrites ( 30 , 31 ). In theory, dendrites have less capacitance than soma and permit wider voltage fluctuations locally to unleash inactivated CaTs. Therefore, activation of GluN2B/D NMDARs in dendrites may initiate CaT-dependent bursts in STN. Although the causal relationship remains to be established, our data suggest that receptor distributions may contribute to the dissociation between bursts and oscillations.

Potential interactions between oscillation-based motor preparation and firing pattern–based motor execution in PD. Bradykinesia is characterized by slow movements beyond volitional control and therefore serves as a prototypical disease model of motor execution. Other than bradykinesia and STN bursts, the volitional motor controls, including motor decision or preparation, heavily modulate beta oscillations in PD ( 4 , 7 , 32 ). This concept is supported by the observations in our free-moving paradigm, which revealed the typical shift of oscillatory frequencies (Figure 2J and Figure 4H) and the reduction of beta powers (Figure 2, G and H, and Figure 4G) when the rats decided to move, regardless of whether their motor performances changed due to CaT/NMDAR manipulations. STN has two parallel mechanisms, oscillations and firing patterns, to modulate movements. The interplay of the two mechanisms could explain how oscillation-based motor preparation tunes the firing pattern–based motor execution. Cortical oscillatory activities can be transmitted to STN via a GluN2A-mediated mechanism, which may oscillate the somatic membrane potentials in STN and disturb the precise timing of firing-pattern switches for motor execution. Motor preparation evidently desynchronizes cortical oscillatory activities ( 4 , 7 , 33 , 34 ), which may suppress the above-mentioned processes and result in better motor execution. Similar mechanisms are well documented in the thalamus, the homolog of STN in developmental biology ( 35 ). Sleep induces thalamic oscillations and interferes with sensory information relays ( 36 – 38 ). Based on the fact that STN receives first-order command directly from the cortex via cortico-subthalamic pathway, it may be one of the fastest and the key mechanisms to explain how volitional motor preparation talks to the automatic/involuntary motor execution, and deserves further investigation by preparation-triggered protocols other than the free-moving paradigms in this study. By revisiting the oscillatory theory with the results in this study, parkinsonian motor control may be divided into two steps: the feedback, interactive oscillations for volitional motor decisions and preparations and fast-responsive, feed-forward neuronal codes for involuntary motor execution, which results in bradykinesia. We did not investigate the interactions between beta oscillations and motor decisions in this study. However, PD patients have significant problems in decision making ( 39 , 40 ), and this study provides the mechanism of beta oscillation that may help in further investigation of this issue.

The roles of neuronal codes versus oscillations in regulating normal motor behaviors. This study focused on the online circuitry mechanism of bradykinesia in PD. Notably, STN bursts and beta oscillations, which are pathologically augmented in PD, also exist in normal motor circuitry. Physiological amounts of STN bursts and beta oscillations are both present in normal conditions ( 12 , 13 ). Reduction of cortical oscillatory activities in beta frequencies (beta desynchronization) is also observed in normal motor preparations ( 33 , 34 , 41 – 43 ). Our study showed that the feed-forward, burst-coded mechanism also regulated inhibitory motor execution in rodents with intact basal ganglia circuits (Figures 3, 5, and 6, and Supplemental Figure 6) and is independent of oscillatory profiles. Naive rats also had cortical beta desynchronization in moving scenarios (Figure 3H), regardless of whether the motor performances were being modulated by HC. The patterns and distributions of NMDARs and CaTs of STN in naive rodents were also similar to those in 6-OHDA rat models (Supplemental Figures 10 and 11). Beyond PD pathophysiology, these results may also apply to physiological states and improve our understandings of the fundamental principles of motor control physiology. Feedback circuitry oscillations may contribute to the volitional aspects of motor commands, while the feed-forward, firing pattern–coded neurotransmissions regulate motor execution (schematic summary, Supplemental Figure 12.

Therapeutic potential based on the new mechanism of bradykinesia. Our results shows that bradykinesia requires the collaboration between GluN2B/D NMDARs and CaTs in STN. Amantadine ( 44 ) and zonisamide ( 45 ), a weak NMDAR and a CaT blocker, respectively, already show modest clinical benefits in PD patients. However, potent and nonspecific NMDAR or CaT blockers are not ideal therapeutic options due to their cognitive side effects. In contrast, GluN2D ( 46 ) and CaV3.3 ( 29 ) have low expression levels in the neocortex but are enriched in PD STN (Supplemental Figure 8 and 9). Regardless of the changes in dopaminergic system and direct-indirect pathways ( 20 , 47 ), targeting of the neuron-modulatory consequences via GluN2D and CaV3.3 may provide better therapeutic options. The standard dopaminergic therapy in PD is notorious for its motor complications ( 48 ). In contrast, NMDAR and CaT blockers did not induce the paradoxical rotations and head tilts (Figures 2 and 4) typically provoked by dopaminergic agents ( 12 ). Amantadine is the best-known anti-dyskinetic therapy in PD ( 44 ). In fact, intervention in the cortico-subthalamic pathway is the key mechanism of DBS ( 12 , 13 ), and the therapeutic effect of DBS is better than the traditional therapy in terms of motor complications ( 49 ). Moreover, systemic dopaminergic therapy contributes to major cognitive and impulse control problems in PD ( 40 , 48 ). Therefore, GluN2D and CaV3.3 could be new bradykinesia-specific targets for PD motor therapy, and may be superior to the standard dopaminergic therapy in terms of its motor and cognitive complications.

Animals. Male adult Wistar rats were entered into the study at

8 weeks of age and 250–350 g. 6-OHDA–lesioned (Sigma-Aldrich) hemiparkinsonian rats were used in all the PD experiments in this study (Supplemental Figure 1 see also Supplemental Methods). For optogenetics experiments, we used male adult Thy1-ChR2-EYFP line 18 transgenic mice (catalog 007612 The Jackson Laboratory), which express channelrhodopsin-2 in cortical neuron layer V ( 13 ) and have been validated as an ideal animal model for selective stimulation of cortico-subthalamic axons ( 12 , 13 ). Mice were entered into the study at

5 weeks of age and weights greater than 20 g. The animals were housed in a vivarium with controlled 12-hour dark/light cycles.

NMDAR and CaT modulators. We used NMDAR blockers with different subunit specificities. (D)-AP5 (2 mM, Tocris) is a non-selective NMDAR blocker. (R)-CPP (200 μM, Tocris) is a selective NMDAR antagonist targeting the GluN2A subunit. Ro 25-6891 (RO 1 mM, Tocris) and PPDA (500 μM, Tocris) inhibit GluN2B/D subunit selectively. To inhibit CaTs, we selected NiCl2 (6 mM, Sigma-Aldrich) and mibefradil (500 μM, Tocris). PPDA was dissolved in DMSO to 50 mM first and then diluted with saline to achieve a final concentration of 500 μM. All the other drugs were dissolved in artificial CSF (aCSF). The pH of all solutions was adjusted to 7.4.

Behavioral recordings and in vivo electrophysiology. We used the open-field test to evaluate the free-moving locomotor behaviors in 6-OHDA and control rodents, and the rotarod test for forced-moving behaviors (see Supplemental Methods for detailed paradigms). In valid 6-OHDA or normal control rats, we implanted microwire deep electrodes for single-unit and local field potential (LFP) recordings, as well as applying HCs. Epidural screw electrodes were also implanted for cortical LFPs. An STN cannula was inserted ipsilateral to 6-OHDA lesioning for real-time NMDAR or CaT modulations (see Supplemental Methods for all surgical procedures). We performed simultaneous behavioral and electrophysiological recordings, before and after online electric and/or pharmacological manipulations. Single-unit firings and LFPs were prefiltered and analyzed separately. For details, see Supplemental Methods.

Optic stimulation and simultaneous behavioral recordings. We activated the cortico-subthalamic axons optogenetically by implanting an optic fiber unilaterally into STN in Thy1-ChR2 mice (Figure 5A). We applied two different protocols: frequency-dependent (10 Hz) and frequency-independent (randomized) illumination (Figure 5B). Free-moving behaviors were accessed under baseline, light-off, and light-on states, with one of the stimulation protocols applied first by random process. For details, see Supplemental Methods.

Analysis of single-unit recordings and LFPs. Signals recorded for single-unit settings were post-processed with spike-sorting software (SciWorks 8.0, DataWave Technologies) and quality-controlled algorithm ( 12 ). Burst patterns were detected in each qualified single unit as described previously ( 10 – 12 ). The LFP data were post-processed with MATLAB 7.4 (MathWorks). Regional power spectrum represented in situ synchronizations, while coherence analysis referred to long-ranged synchronization. For details, see Supplemental Methods.

Immunohistochemistry. Three-month-old adult C57BL/6J mice and 4-month-old Wistar rats with or without 6-OHDA lesioning were used for immunohistochemistry study. Mice and rats were anesthetized with isoflurane and then sacrificed by an overdose of urethane (2 g/kg i.p., Sigma-Aldrich) and transcardially perfused with 4% paraformaldehyde in PBS. The brain was then removed and immersed in the 4% paraformaldehyde overnight and moved to PBS for 3 days. The brain was sliced coronally at the thickness of 30 μm by vibrotome. The sections were washed with PBS, followed by the suppression in 10% normal donkey serum in 0.1% Triton. The sections were subsequently incubated with respective primary antibodies overnight at 4°C and then secondary fluorescent antibodies (all from Invitrogen). Primary antibodies included GluN2A (Neuromab, Cat. No. 75-288), GluN2B (Neuromab, 75-097), GluN2D (Bioss, Bs-1072R), MAP2 (Abcam, Ab5392), CaV3.1 (Alomone Labs, ACC-021), CaV3.2 (Alomone Labs, ACC-025) and CaV3.3 (Alomone Labs, ACC-009). Images were taken using confocal laser scanning microscope (Leica TCS SP2 two-photon microscope). See also Supplemental Methods and Supplemental Figure 13 for more details.

Statistics. The statistics were managed with SPSS 13.0 and plotted with Excel 2013 (Microsoft). Nonparametric Wilcoxon signed-rank test was used to analyze paired data, including animal behaviors, single-unit recordings, and LFP analyses (Figures 1–4 and Supplemental Figures 2–4). For those data with 3 or more conditions, we applied 1-way ANOVA with post-hoc Bonferroni correction (Figures 5 and 6 and Supplemental Figures 5–7). In all statistical methods, a P value less than 0.05 was considered significant.

Study approval. The study was approved by the IACUC of National Taiwan University College of Medicine and College of Public Health.

MKP designed the study and generated the behavioral and electrophysiological results in rats with CHT, YMW, WCL, and TRW. SHK designed and generated the histology/pathology data in rodents and human subjects. MKP and WSL designed the optogenetic experiments in mice and analyzed the results with JCP and CYC. MKP and JYL designed MATLAB codes. CCK led the team and coordinated the study. MKP and CCK interpreted the results and wrote the article with input and comments from the other authors.

We thank F.-C. Lin (Southport Tech. Co.) for the technical support. This research is supported by the Ministry of Science and Technology in Taiwan (MOST-103-2320-B-002-026-MY3 and MOST-104-2321-B-002-067 to CCK MOST-104-2314-B-002-076-MY3 to MKP), the National Health Research Institute in Taiwan (NHRI-EX105-10503NI to CCK) National Taiwan University Hospital (104-N2870 and MG380 to MKP) and the Yin-Lin branch of the hospital (NTUHYL104.N007 to MKP and YMW).

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2016126(12):4516–4526. doi:10.1172/JCI88170.


Neural oscillations and activity patterns - Psychology

Our lab uses psychophysics and computational modeling in concert with tools from cognitive neuroscience to measure and manipulate the human brain. We seek to understand the neural basis of visual consciousness, attention, perceptual decision making, metacognition, and working memory, and have a particular interest in the role of neural oscillations in these domains.

We are housed within the Department of Psychology at the University of California, Santa Cruz.

In "Pre-Stimulus Alpha-Band Phase Gates Afferent Visual Cortex Responses" led by lab Ph.D. student Wei Dou, we find that the phase of ongoing alpha-band oscillations modulated the afferent cortical response to a visual stimulus, as captured by EEG responses taken during the time window of the C1 event-related potential component. We take this as evidence for a thalamic inhibition model of alpha, consistent with animal models of alpha that posit generators of phasic inhibition in the LGN.

In "Spontaneous neural oscillations influence behavior and sensory representations by suppressing neuronal excitability" collaborator Luca Iemi and colleagues analyzed electrocorticography data recorded across many brain areas to ask how oscillatory amplitude modulates broad-band gamma activity in response to visual and auditory stimuli. We report that alpha and beta power inversely correlates with spontaneous and stimulus-evoked broad-band gamma activity - an effect that mediated reaction times in a discrimination task. Interestingly, a pattern classifier trained to discriminate different stimuli found that ongoing alpha power modulated the magnitude of the responses, leading to higher classifier confidence, but no change in classifier accuracy.

Pre- s timulus a lpha- b and p hase g ates a fferent v isual c ortex r esponses

W Dou, A Morrow, L Iemi, J Samaha biorXiv

Spontaneous neural oscillations influence behavior and sensory representations by suppressing neuronal excitability

L Iemi, L Gwilliams, J Samaha, R Auksztulewicz, YM Cycowicz, JR King, V V Nikulin, T Thesen, W Doyle, O Devinsky, CE Schroeder, L Melloni, S Haegens bior Xiv


4. Discussion

In this work we introduced a simple oscillatory memory model of short-term memory, examined some of its properties, and compared its behavior to that of human subjects on a running memory span task. Our model's dynamics are intrinsically oscillatory due to the use of rapidly varying threshold values, and recall of an item is dependent on the time elapsed since it was observed due to the use of rapidly decaying weights. Unlike with many past neurocomputational models of memory, we assessed recall by initializing the model's activity to a random state rather than by initializing it to a noisy or partial stored memory pattern, or by biasing the network's dynamics by applying an external input pattern that represents a noisy or partial stored pattern. We found that when moderate decay rates were used, this approach resulted in a short-term memory capacity of two to three items, a value that is comparable to what has been observed in experimental studies by others (Baddeley, 2000 Cowan, 2001 Cowan et al., 2005) and that matches the memory capacity that we observed in a group of human subjects performing a similar running memory task. The model also showed a prominent recency effect, as would be expected given the use of weight decay and as is also seen in human subjects.

It should also be noted here that this model is intended to simulate short-term memory processing only, and so it is not intended to address any processes by which semantic or other long-term memory information is accessed to aid storage or recall. Indeed, it has been well established that short-term memory capacity is higher for familiar items for which long-term representations exist, compared to novel stimuli (Hulme, Maughan, & Brown, 1991 see Cowan, 2001, for review). This benefit is likely due to the fact that for novel stimuli, representations must be created before retention can successfully occur. As the model made no assumptions regarding the relationship between short-term and long-term memory, the choice of letters as the retained stimuli seemed a reasonable assumption. Still, we recognize that more complex models of short-term memory could benefit from expansion by making predictions regarding the important role of long-term memory for temporarily retained information.

Also, although the computational model is oscillatory in nature (though not periodic), it is not intended to make any predictions regarding frequency-specific responses actually produced in the brain during short-term memory retention. There is a rich and expanding research literature showing that EEG activity in the theta band plays a key role in the maintenance of temporarily retained information (Jensen & Tesche, 2002) and that hippocampal structures play a significant role in this frequency component (Kahana, Deelig, & Madsen, 2001 Rizzuto et al., 2003). There has also been recent research to show that higher-frequency oscillatory activity (i.e., gamma response) increases approximately linearly with increased memory load (Howard et al., 2003), and that greater gamma activity during encoding predicts greater likelihood of later recall (Sederberg, Kahana, Howard, Donner, & Madsen, 2003). The results of our model are promising in suggesting that oscillatory neural models can show similar capacity limitations as with humans, but they do not allow us to make predictions regarding frequency-specific contributions to EEG, especially as the model oscillations recorded are in terms of the extent to which a specific distributed memory pattern is present (quantity sλ) and not in terms of amount of network activity. While the model does retain information about stimulus order—indeed, stimulus order effects emerge in the model as a result of gradual decay—it does not address issues of temporal sequencing. As recent single-unit recording evidence suggests that ordered sequences of activation are observed in rats (Foster & Wilson, 2006), this may be an interesting area of future expansion of the model.

Our study adds to a rapidly growing literature on computational models of short-term memory by examining the role of weight decay on simple oscillatory memories. Many past models of short-term memory have employed lateral inhibition between representational units to establish competition between activated entities, and thus capacity limitations. For example, Haarmann and Usher (2001) present a model of semantic short-term memory that functions in this fashion. Our approach differs in not explicitly building in such lateral inhibition (although inhibitory weights do occur during pattern storage), with competition between memory patterns arising in the dynamics due to the interference occurring between the nonorthogonal memory patterns. Other recent models of short-term memory, inspired by specific neuroanatomical structures, have used separate modules for memory representation, maintenance, and selective gating. For example, Frank, Loughry, and O'Reilly (2001) and O'Reilly and Frank (2006) incorporate modules representing prefrontal cortex and basal ganglia, with the latter modulating which sensory stimuli are kept active. Our approach does not use a complex architecture or gating mechanisms, and thus shows that some basic behavioral properties of human short-term memory (limited memory capacity, recency effect, and shifts in position-specific stimulus recall) can be captured by a surprisingly simple neurocomputational mechanism. Still other recent short-term memory models have been based on modulation of persistent neuronal firing by rhythmic changes to membrane potential at theta frequencies (Koene & Hasselmo, 2007, 2008). Our approach is quite different in that storage is based primarily on synaptic connectivity, and memory capacity limitations arise mainly due to synaptic decay and pattern interference. However, it is an interesting question as to whether the dynamic thresholds in our model might correspond to the changes in effective thresholds brought about by the modulating theta activity in these more physiologically realistic implementations.

Perhaps the most interesting finding with the model is that by adjusting just the weight decay rate, one can produce shifts in the model's memory capacity and position-specific recall rates, as is demonstrated in Figure 4. This represents a prediction of the model that by adjusting the decay rate, one could reasonably match the shifts exhibited by human subjects who were instructed to recall different-length stimuli sequences. This would be especially remarkable given the simplicity of our model and that it requires adjustment of only a single parameter. This prediction relates to long-standing issues in the cognitive science literature concerning the nature of forgetting. For example, one view of forgetting is that short-term memory is subject to decay (Brown, 1958), while an alternative view is that forgetting is due to interference between competing elements that are simultaneously vying for attention (Waugh & Norman, 1965). Our model incorporates both interference and decay as mechanisms for forgetting and shows that the latter can partially mitigate effects from interference, consistent with evidence in past behavioral studies (Altmann & Gray, 2002).

In general, it is difficult to map processes in neurocomputational models to cognitive processes, but sometimes there are analogs that are worth considering. The observation that adjustments to decay rate control not only the total short-term memory capacity (see Figure 5) but also position-specific stimulus recall rates (see Figure 4) raises the issue of whether altering decay rate might be a useful mechanism permitting a cognitive system to control short-term memory characteristics. Specifically, our model is consistent with the hypothesis that dynamic adjustments to activity decay rate may be an important aspect of the human attention mechanisms that control forgetting (Altmann & Gray, 2002).

It is already well established that attention is a cognitive property that can be manipulated based on the needs of the task at hand (Broadbent, 1982 Downing & Pinker, 1985 Eriksen & St. James, 1986) and that attentional scope can be adjusted during visual search and memory recall between being more focused or more diffuse (Engle, 2002 Kane & Engle, 2002). Based on our modeling results, we hypothesize that altering the decay rate could serve as a means by which attentional mechanisms could act to manipulate attentional scope. More focused attention is simulated in the model by a higher decay rate, so that attention is directed more intently on a smaller number of items. In this way, decay is used as a means for combating proactive interference, with higher decay rates leading to more effective retention of recent information, but also at the expense of that which was presented before it.

For the running memory span task used here involving rapid presentation of stimuli, human subjects attempt to hold presented stimuli in a limited-capacity memory without the use of rehearsal (Bunting et al., 2006). Assuming that maintaining such stimuli depends on attentional resources, then changing instructions requiring subjects to retain varying numbers of stimuli (i.e., not just six as we did in our behavioral experiments) would be expected to have a great effect. Specifically, if attention is drawn sufficiently thin so that activation maintenance is small across all retained stimuli (a low decay rate in our model), then with longer stimulus sequences (e.g., a task requiring human subjects to recall 12 stimuli) few or none of the stimuli would be expected to retain activation levels above some cognitive threshold required for successful recall due to interference, although no doubt some attenuated recency effect will still be present. This is both a surprising and informative prediction from the model, and it suggests that overloading subjects' attentional resources (i.e., drawing attention sufficiently thin) has a detrimental effect on retention. Future behavioral testing with a varying-length recall task could therefore either refute or strongly support the model we have presented here.

To our knowledge, the work reported here is the first comparison of simple oscillatory model properties to human behavioral data. While the results are encouraging, they leave open a number of issues that should be examined in future work. Perhaps the most pressing issue concerns the generality of these results. It will be important to determine whether similar correspondences between model properties and human behavior can be produced as readily with other stimulus sets, while varying the number of stimuli that subjects are instructed to retain and recall, and across other short-term memory tasks. For example, it would be useful to compare the model's results against human data using unfamiliar or abstract symbols or nonwords where subject performance is based on free recall. Similarly, with the model, it would be useful to examine how long-term memory influences the results of recall. Another future issue, not examined in our work, would be to characterize model behavior in the absence of interference using orthogonal stimulus patterns, allowing the effects of weight decay to be studied in isolation.


Contents

Gamma waves can be detected by electroencephalography or magnetoencephalography. One of the earliest reports of gamma wave activity was recorded from the visual cortex of awake monkeys. [7] Subsequently, significant research activity has concentrated on gamma activity in visual cortex. [8] [9] [10] [11]

Gamma activity has also been detected and studied across premotor, parietal, temporal, and frontal cortical regions [12] Gamma waves constitute a common class of oscillatory activity in neurons belonging to the cortico-basal ganglia-thalamo-cortical loop. [13] Typically, this activity is understood to reflect feedforward connections between distinct brain regions, in contrast to alpha wave feedback across the same regions. [14] Gamma oscillations have also been shown to correlate with the firing of single neurons, mostly inhibitory neurons, during all states of the wake-sleep cycle. [15] Gamma wave activity is most prominent during alert, attentive wakefulness. [13] However, the mechanisms and substrates by which gamma activity may help to generate different states of consciousness remain unknown.

Controversy Edit

Some researchers contest the validity or meaningfulness of gamma wave activity detected by scalp EEG, because the frequency band of gamma waves overlaps with the electromyographic frequency band. Thus, gamma signal recordings could be contaminated by muscle activity. [16] Studies utilizing local muscle paralysis techniques have confirmed that EEG recordings do contain EMG signal, [17] [18] and these signals can be traced to local motor dynamics such as saccade rate [19] or other motor actions involving the head. Advances in signal processing and separation, such as the application of independent component analysis or other techniques based on spatial filtering, have been proposed to reduce the presence of EMG artifacts. [16]

Conscious perception Edit

Gamma waves may participate in the formation of coherent, unified perception, also known as the problem of combination in the binding problem, due to their apparent synchronization of neural firing rates across distinct brain regions. [20] [21] [22] 40-Hz gamma waves were first suggested to participate in visual consciousness in 1988, [23] .e.g, two neurons oscillate synchronously (though they are not directly connected) when a single external object stimulates their respective receptive fields. Subsequent experiments by many others demonstrated this phenomenon in a wide range of visual cognition. In particular, Francis Crick and Christof Koch in 1990 [24] argued that there is a significant relation between the binding problem and the problem of visual consciousness and, as a result, that synchronous 40 Hz oscillations may be causally implicated in visual awareness as well as in visual binding. Later the same authors expressed skepticism over the idea that 40-Hz oscillations are a sufficient condition for visual awareness. [25]

A number of experiments conducted by Rodolfo Llinás supports a hypothesis that the basis for consciousness in awake states and dreaming is 40-Hz oscillations throughout the cortical mantle in the form of thalamocortical iterative recurrent activity. In two papers entitled "Coherent 40-Hz oscillation characterizes dream state in humans” (Rodolfo Llinás and Urs Ribary, Proc Natl Acad Sci USA 90:2078-2081, 1993) and "Of dreaming and wakefulness” (Llinas & Pare, 1991), Llinás proposes that the conjunction into a single cognitive event could come about by the concurrent summation of specific and nonspecific 40-Hz activity along the radial dendritic axis of given cortical elements, and that the resonance is modulated by the brainstem and is given content by sensory input in the awake state and intrinsic activity during dreaming. According to Llinás’ hypothesis, known as the thalamocortical dialogue hypothesis for consciousness, the 40-Hz oscillation seen in wakefulness and in dreaming is proposed to be a correlate of cognition, resultant from coherent 40-Hz resonance between thalamocortical-specific and nonspecific loops. In Llinás & Ribary (1993), the authors propose that the specific loops give the content of cognition, and that a nonspecific loop gives the temporal binding required for the unity of cognitive experience.

A lead article by Andreas K. Engel et al. in the journal Consciousness and Cognition (1999) that argues for temporal synchrony as the basis for consciousness, defines the gamma wave hypothesis thus: [26]

The hypothesis is that synchronization of neuronal discharges can serve for the integration of distributed neurons into cell assemblies and that this process may underlie the selection of perceptually and behaviorally relevant information.

Attention Edit

The suggested mechanism is that gamma waves relate to neural consciousness via the mechanism for conscious attention:

The proposed answer lies in a wave that, originating in the thalamus, sweeps the brain from front to back, 40 times per second, drawing different neuronal circuits into synch with the precept [sic], and thereby bringing the precept [sic] into the attentional foreground. If the thalamus is damaged even a little bit, this wave stops, conscious awarenesses do not form, and the patient slips into profound coma. [21]

Thus the claim is that when all these neuronal clusters oscillate together during these transient periods of synchronized firing, they help bring up memories and associations from the visual percept to other notions. This brings a distributed matrix of cognitive processes together to generate a coherent, concerted cognitive act, such as perception. This has led to theories that gamma waves are associated with solving the binding problem. [20]

Gamma waves are observed as neural synchrony from visual cues in both conscious and subliminal stimuli. [27] [28] [29] [30] This research also sheds light on how neural synchrony may explain stochastic resonance in the nervous system. [31]

Mood disorders Edit

Altered gamma wave activity is associated with mood disorders such as major depression or bipolar disorder and may be a potential biomarker to differentiate between unipolar and bipolar disorders. For example, human subjects with high depression scores exhibit differential gamma signaling when performing emotional, spatial, or arithmetic tasks. Increased gamma signaling is also observed in brain regions that participate in the default mode network, which is normally suppressed during tasks requiring significant attention. Rodent models of depression-like behaviors also exhibit deficient gamma rhythms. [32]

Schizophrenia Edit

Decreased gamma-wave activity is observed in schizophrenia. Specifically, the amplitude of gamma oscillations is reduced, as is the synchrony of different brain regions involved in tasks such as visual oddball and Gestalt perception. People with schizophrenia perform worse on these behavioral tasks, which relate to perception and continuous recognition memory. [33] The neurobiological basis of gamma dysfunction in schizophrenia is thought to lie with GABAergic interneurons involved in known brain wave rhythm-generating networks. [34] Antipsychotic treatment, which diminishes some behavioral symptoms of schizophrenia, does not restore gamma synchrony to normal levels. [33]

Epilepsy Edit

Gamma oscillations are observed in the majority of seizures [5] and may contribute to their onset in epilepsy. Visual stimuli such as large, high-contrast gratings that are known to trigger seizures in photosensitive epilepsy also drive gamma oscillations in visual cortex. [35] During a focal seizure event, maximal gamma rhythm synchrony of interneurons is always observed in the seizure onset zone, and synchrony propagates from the onset zone over the whole epileptogenic zone. [36]

Alzheimer's disease Edit

Enhanced gamma band power and lagged gamma responses have been observed in patients with Alzheimer's disease (AD). [4] [37] Interestingly, the tg APP-PS1 mouse model of AD exhibits decreased gamma oscillation power in the lateral entorhinal cortex, which transmits various sensory inputs to the hippocampus and thus participates in memory processes analogous to those affected by human AD. [38] Decreased hippocampal slow gamma power has also been observed in the 3xTg mouse model of AD. [39]

Gamma stimulation may have therapeutic potential for AD and other neurodegenerative diseases. Optogenetic stimulation of fast-spiking interneurons in the gamma wave frequency range was first demonstrated in mice in 2009. [40] Entrainment or synchronization of hippocampal gamma oscillations and spiking to 40 Hz via non-invasive stimuli in the gamma frequency band, such as flashing lights or pulses of sound, [3] reduces amyloid beta load and activates microglia in the well-established 5XFAD mouse model of AD. [41] Subsequent human clinical trials of gamma band stimulation have shown mild cognitive improvements in AD patients who have been exposed to light, sound, or tactile stimuli in the 40 Hz range. [1] However, the precise molecular and cellular mechanisms by which gamma band stimulation ameliorates AD pathology is unknown.

Fragile X syndrome Edit

Hypersensitivity and memory deficits in Fragile X syndrome may be linked to gamma rhythm abnormalities in sensory cortex and hippocampus. For example, decreased synchrony of gamma oscillations has been observed in auditory cortex of FXS patients. The FMR1 knockout rat model of FXS exhibits an increased ratio of slow (

High-amplitude gamma wave synchrony can be self-induced via meditation. Long-term practitioners of meditation such as Tibetan Buddhist monks exhibit both increased gamma-band activity at baseline as well as significant increases in gamma synchrony during meditation, as determined by scalp EEG. [2] fMRI on the same monks revealed greater activation of right insular cortex and caudate nucleus during meditation. [42] The neurobiological mechanisms of gamma synchrony induction are thus highly plastic. [43] This evidence may support the hypothesis that one's sense of consciousness, stress management ability, and focus, often said to be enhanced after meditation, are all underpinned by gamma activity. At the 2005 annual meeting of the Society for Neuroscience, the current Dalai Lama commented that if neuroscience could propose a way to induce the psychological and biological benefits of meditation without intensive practice, he "would be an enthusiastic volunteer." [44]


Methods

Participants

Forty-two participants were recruited from the student population of a German university. Participants were between 19 and 29 years old (on average 22.5 years, SD = 2.8), 36 participants were female. Participants received course credit for participation and, in addition, they were paid € 17. The study took about 2 hours. Written informed consent was obtained for all participants. For our sample size, α = 0.05 and a medium to large effect of r = 0.37, the statistical power to discover an effect if it exists in the population is (1-β) = 0.80.

Task and Measures

The n-back task consisted of three blocks with 62 trials each. On each trial, participants saw a fixation-cross, followed by a capital letter taken from a pool of eleven letters. Each trial took 2.5 seconds (see Fig. 1). Participants were instructed to press a button with their right index finger when the letter was a target, or a second button with their left index finger if the letter was a non-target. In the first block (0-back), the target was the letter “X”. In the second block (1-back), the letter from the last trial served as target. Therefore, participants had to constantly update and memorize the target. In the third and final block (2-back), the letter presented two trials ago served as target. In addition to the 1-back condition, participants had to memorize und constantly update a second target along with the information which of the two letters was the target. Half of the trials per block were targets. The same randomized sequence of targets and non-targets and the same letters were used for all participants. Performance on the n-back task was coded trial-by-trial. In less than 1% of the trials, participants did not respond within the 2-second time span therefore, these trials were excluded from the analyses regarding reaction time and hit rate as corresponding data were not available. Otherwise, responses were coded as correct if the correct button was pressed for either a target or a non-target and as non-correct if the wrong button was pressed for either a target or a non-target.

For the assessment of need for cognition, the 18-item short scale was used 25 . The construct need for cognition can be located in the Big Five personality framework, pertaining to the aspect Intellect of the domain openness to experience 32,47 . Items were presented in German language. Participants indicated on a seven point Likert scale, ranging from “do not agree at all” to “fully agree”, the extent to which each of the statements pertained to them. Half of the items are reversed scored and were recoded before aggregating across all items. In the present study, the scale had an internal consistency of α = 0.88 (descriptive statistics were M = 84.1, SD = 14.8). For Fig. 2c, a median split was performed on need for cognition, resulting in one sub-group with low (N = 22 M = 72.7 SD = 10.3 range [39 81]) and one sub-group with high (N = 20 M = 96.6 SD = 6.7 range [84 107]) values on need for cognition.

Working memory capacity was assessed by three tasks from the Wechsler Adult Intelligence Scale 48 . The tasks were manually administered by an examiner in a one-to-one situation. Examiners were blind with regard to all other study variables (e.g., need for cognition, theta power). The first task demanded the repetition of an increasing number of digits (2 to 18) the second task was a backwards-repetition of an increasing number of digits (2 to 16). Both tasks had two items on each difficulty level. The third task consisted of the sorting and repetition of an increasing number of digits and letters (2 to 8) with three items per difficulty level thereby, digits had to be given first, in ascending order, followed by alphabetically ordered letters. Within each task and each difficulty level, a termination criterion was defined specifically, if all items within a difficulty level were answered incorrectly, the task was aborted. Within each task, performance was computed as the sum of all correctly answered items. Working memory performance was estimated as the aggregate of z-standardized values in the three tasks.

All methods were carried out in accordance with the approved guidelines of the Julius Maximilians University Würzburg and all experimental protocols were approved by its ethic committee.

EEG Recording and Quantification

While participants performed the n-back task, EEG (analog bandpass: 0.1–80 Hz, sampling rate: 250 Hz) was recorded from 31 scalp sites according to the 10–20 system, using Ag/AgCl electrodes and a BrainAmpDC amplifier (Brain Products GmbH, Gilching, Germany). Impedances were kept below 10 kΩ and electrodes were referenced to the vertex (Cz). For detection of blinks and eye-movements the vertical electrooculogram (EOG) was recorded. Data were processed offline, using Brain Vision Analyzer 2.0 software (Brain Products GmbH, Gilching, Germany). First, data were filtered, using a 0.15 Hz high-pass and a 40 Hz low-pass filter (24 dB/Octave) and, additionally, a 50 Hz notch filter. Subsequently, the EEG was segmented into epochs of 3200 ms (−800 to 2600 ms, relative to the presentation of the stimulus). Afterwards, data were corrected for ocular artifacts using an Independent Component Analysis based correction method implemented in the Brain Vision Analyzer. Larger artifacts were automatically detected by a computer algorithm implemented in Brain Vision Analyzer 2.0 software and discarded if applicable, both prior and after the ocular correction. For this purpose the following exclusion criteria were applied: (1) maximal voltage difference >250 μV within 1000 ms (prior to the ocular correction, to enhance data quality for the independent component analysis without excluding trials according to eye-related activity) (2) maximal voltage difference >100 μV within 1000 ms across the epoch and maximal voltage step of 20 μV/msec (after the correction for ocular artifacts). At least 32 artifact-free trials (on average 56 trials) were available per participant and condition. Subsequently, data were re-referenced to an averaged reference across all electrodes (excluding vertical EOG).

Event related potentials were computed by segmenting the data to −500 to 2000 ms relative to the presentation of the stimulus, averaging across all trials and correcting for baseline activity (−150 to −50 ms).

For time-frequency analyses, segmented data were convolved using a family of complex Morlet wavelets from 1 to 20 Hz in linear steps of 1 Hz. The complex Morlet wavelets are defined as Gaussian-windowed complex sine functions: , with and , the latter resulting in total energy of 1 for all frequency levels 49 . Constant ratio (fof) was set 6.7 50,51 . For each frequency layer, power values were baseline corrected by subtracting the mean activity in the time window −150 to −50 ms before stimulus onset from each data point. (Please see the supplemental material for detailed information on the rational for choosing a baseline period rather close to stimulus onset and for results for an alternative baseline −450 to −350 ms before stimulus onset.) Segments were averaged for each participant and condition of interest. For analyses of FMθ activity, mean power values 4 to 7 Hz at Fz, F3, FCz, F4 and Cz were extracted from 650 to 1900 ms following the presentation of the stimulus and aggregated across these electrodes.

Heart period measurement

We used three disposable Ag/AgCl electrodes (Covidien Kendall ECG Electrodes H98LG) placed according to a modified Einthoven II lead to measure heart period, defined as the time between two adjacent R-peaks 52 . The ground electrode was placed below the left collarbone, the negative electrode below the right collarbone and the positive electrode on the left side below the rib cage. The signal was digitized using a Brain Vision BrainAmp ExG amplifier (Brain Products GmbH, Gilching, Germany) with a sampling rate of 250 Hz and the BrainVision Recorder 1.20 software (Brain Products GmbH).

To analyze the event-related heart period response we first detected the peaks of the R-waves using QRSTool 53 . Following an automatic detection, the detected beats were checked and if necessary corrected manually. Next, the times of the R-peaks were exported and event-related inter-beat-intervals were extracted using a custom-built Matlab script. For every stimulus in the n-back task we extracted the inter-beat-interval surrounding the stimulus as well as the two following intervals. Heart period was defined as the interval in milliseconds between sequential R-waves and was estimated for the interval following the stimulus.

High-frequency heart period variability (HF-HPV) was estimated using the software ARTiiFACT 54 . Based on the timing of the R-peaks exported from QRSTool we computed separated inter-beat-interval series for every condition of the n-back task per participant (mean length of the series across all participants and conditions: 154.65 seconds). These inter-beat-interval series were then submitted to a frequency domain analysis of HF-HPV in ARTiiFACT. Except for the width of the Hanning window (153 seconds), standard settings of the Fast Fourier Transformation (FFT) were used (4 Hz spline interpolation, 50% overlap of resampled and detrended data). The absolute values (in ms 2 ) of the HF-HPV component (0.15–0.40 Hz) were used for statistical analysis 55 . Since those values were not normally distributed, we transformed them using the natural logarithm before running the statistical analysis.

Statistical analyses

Latent growth curve analysis 23 with maximum likelihood function was performed using AMOS 23.0 with bootstrapping to account for the medium sample size. Latent growth curve analysis is a special case of multilevel modeling and captures the pattern of change across multiple sampling points. For each individual, a linear function described by slope and intercept was estimated. As depicted in Fig. 4a, the intercept was specified by a latent variable with regression weights fixed to 1 between the latent intercept and the three manifest variables indicating theta power in the three conditions. The slope is specified by a latent variable indicated by regression weights fixed to −2, 1 and 0 for theta power in the 0-back, 1-back and 2-back condition, respectively. Therefore, the intercept specifies predicted theta power in the 2-back condition.

Need for cognition was modeled as latent variable indicated by three manifest variables the latter were computed by randomly assigning each of the 18 items (see above) to one of three parcels and subsequently averaging across these items. Working memory was modeled as a latent variable with the z-standardized values of each of the three working memory tasks as indicators.

Factor score weights were used to predict the estimated slope for each subject and each condition for subsequent analyses. We used mixed models for binomial data with a LOGIT link function for the prediction of trial-by-trial performance in the n-back task, using the GENLINMIXED procedure in SPSS 23.0.


Neural oscillations and activity patterns - Psychology

Our lab uses psychophysics and computational modeling in concert with tools from cognitive neuroscience to measure and manipulate the human brain. We seek to understand the neural basis of visual consciousness, attention, perceptual decision making, metacognition, and working memory, and have a particular interest in the role of neural oscillations in these domains.

We are housed within the Department of Psychology at the University of California, Santa Cruz.

In "Pre-Stimulus Alpha-Band Phase Gates Afferent Visual Cortex Responses" led by lab Ph.D. student Wei Dou, we find that the phase of ongoing alpha-band oscillations modulated the afferent cortical response to a visual stimulus, as captured by EEG responses taken during the time window of the C1 event-related potential component. We take this as evidence for a thalamic inhibition model of alpha, consistent with animal models of alpha that posit generators of phasic inhibition in the LGN.

In "Spontaneous neural oscillations influence behavior and sensory representations by suppressing neuronal excitability" collaborator Luca Iemi and colleagues analyzed electrocorticography data recorded across many brain areas to ask how oscillatory amplitude modulates broad-band gamma activity in response to visual and auditory stimuli. We report that alpha and beta power inversely correlates with spontaneous and stimulus-evoked broad-band gamma activity - an effect that mediated reaction times in a discrimination task. Interestingly, a pattern classifier trained to discriminate different stimuli found that ongoing alpha power modulated the magnitude of the responses, leading to higher classifier confidence, but no change in classifier accuracy.

Pre- s timulus a lpha- b and p hase g ates a fferent v isual c ortex r esponses

W Dou, A Morrow, L Iemi, J Samaha biorXiv

Spontaneous neural oscillations influence behavior and sensory representations by suppressing neuronal excitability

L Iemi, L Gwilliams, J Samaha, R Auksztulewicz, YM Cycowicz, JR King, V V Nikulin, T Thesen, W Doyle, O Devinsky, CE Schroeder, L Melloni, S Haegens bior Xiv


The Human Auditory System

Luc H. Arnal , . Anne-lise Giraud , in Handbook of Clinical Neurology , 2015

Neural oscillations as endogenous temporal constraints

Neural oscillations correspond to synchronous activity of neuronal assemblies that are both intrinsically coupled and coupled by a common input. It was proposed that these oscillations reflect modulations of neuronal excitability that temporally constrain the sampling of sensory information ( Schroeder and Lakatos, 2009a ). The intriguing correspondence between the size of certain speech temporal units and the frequency of oscillations in certain frequency bands ( Fig. 5.1 ) has elicited the intuition that they might play a functional role in sensory sampling (see below). Oscillations are evidenced by means of a spectrotemporal analysis of electrophysiologic recordings (see Wang, 2010 , for a review). The requirements for measuring oscillations and spiking activity are different. The presentation of an exogenous stimulus typically results in an increase of spiking activity in those brain areas that are functionally sensitive to such inputs. Neural oscillations, on the other hand, can be observed in local field potential recordings in the absence of any external stimulation. Exogenous stimulation however typically modulates oscillatory activity, resulting either in a reset of their phase and/or a change (increase or decrease) in the magnitude of these oscillations ( Howard and Poeppel, 2012 ).

Cortical oscillations are proposed to shape spike-timing dynamics and to impose phases of high and low neuronal excitability ( Britvina and Eggermont, 2007 Schroeder and Lakatos, 2009a, b Panzeri et al., 2010 ). The assumption that it is oscillations that cause spiking to be temporally clustered derives from the observation that spiking tends to occur in specific phases (i.e., the trough) of oscillatory activity ( Womelsdorf et al., 2007 ). It is also assumed that spiking and oscillations do not reflect the same aspect of information processing. Whereas spiking reflects axonal activity, oscillations are said to reflect mostly dendritic synaptic activity ( Wang, 2010 ). While both measures are relevant to address how sensory information is encoded in the brain, we believe that the ability of neural oscillations to temporally organize spiking activity supports the functional relevance of neural oscillations to solve the discretization problem and to permit the integration of complex sensory signals across time.

Neuronal oscillations are ubiquitous in the brain, but they vary in strength and frequency depending on their location and the exact nature of their neuronal generators ( Mantini et al., 2007 Hyafil et al., 2012 ). The notion that neural oscillations shape the way the brain processes sensory information is supported by a wealth of electrophysiologic findings in humans and animals. On the one hand, stimuli that occur in the ideal excitability phase of slow oscillations (< 12 Hz) are processed faster and with a higher accuracy ( Lakatos et al., 2008 Busch et al., 2009 Henry and Obleser, 2012 Ng et al., 2012 Wyart et al., 2012 ). On the other hand, gamma-band 40-Hz activity (low gamma band) can be observed at rest in both monkey ( Fukushima et al., 2012 ) and human auditory cortex. In humans, it can be measured using EEG, MEG, and with a more precise localization with concurrent EEG and functional magnetic resonance imaging ( Morillon et al., 2010 ) and intracranial electroencephalographic recordings (stereotactic EEG (sEEG), Electro-corticography (EcoG)) in patients. Neural oscillations in this range are endogenous in the sense that one can observe a spontaneous spike clustering at approximately 40 Hz even in the absence of external stimulation. This gamma activity is thought to be generated by a “ping-pong” interaction between pyramidal cells and inhibitory interneurons ( Borgers et al., 2005, 2008 ), or even just among interneurons that are located in superficial cortical layers ( Tiesinga and Sejnowski, 2009 ). Exogenous inputs usually increase gamma-band activity in sensory areas, presumably clustering spiking activity that is propagated to higher hierarchic processing stages ( Arnal et al., 2011 Arnal and Giraud, 2012 Bastos et al., 2012 ). By analogy with the proposal of Elhilali et al. (2004) that slow responses gate faster ones, it is interesting to envisage this periodic modulation of spiking by oscillatory activity as an endogenous mechanism to optimize the extraction of relevant sensory input in time. Such integration could occur under the patterning of slower oscillations in the delta-theta range.


Résumé

Le terme d'«oscillations cérປrales ou neuronales» se rapporte à l'activité électrique rythmique et/ou répétitive générພ spontanément et en réponse aux stimuli par le tissu neuronal dans le système nerveux central. L'importance des oscillations cérປrales dans les processus cognitivo-sensoriels devient de plus en plus manifeste. Les oscillations liພs aux événements sont clairement modifiພs dans de nombreux types de pathologies neurologiques, en particulier dans le dຜlin cognitif. Cet article analyse les méthodes comme les spectres et les oscillations évoquພs/liພs à un événement, les analyses de cohérence et le blocage de phase. II donne des exemples d'application de concepts et de méthodes essentiels dans le trouble bipolaire, servant de base pour des notions fondamentales sur des biomarqueurs neurophysiologiques dans le dຜlin cognitif. Le message clé est le suivant: au cours du développement des stratégies diagnostiques et pharmacothérapeutiques, les donnພs neurophysiologiques devraient être analysພs dans un cadre utilisant de multiples méthodes et bandes de fréquence.


Role in neural syntax

Brain oscillations can segregate and group neuronal activity to decompose and package neuronal information for communication between brain areas. Because all neuronal oscillations are based on inhibition, they can parse and concatenate neuronal messages, a prerequisite for any coding mechanism. Gamma waves combine neurons into assemblies, which can be conceived as a neuronal “letter.” The hierarchical nature of cross-frequency-coupled rhythms can serve as a mechanism for combining neuronal letters into neuronal words and words into sentences, so that unbounded combinatorial information can be generated from spike patterns. Neuronal oscillators are pulsatile and can readily entrain one another, facilitating the effectiveness of message exchange between brain areas.


4. Discussion

In this work we introduced a simple oscillatory memory model of short-term memory, examined some of its properties, and compared its behavior to that of human subjects on a running memory span task. Our model's dynamics are intrinsically oscillatory due to the use of rapidly varying threshold values, and recall of an item is dependent on the time elapsed since it was observed due to the use of rapidly decaying weights. Unlike with many past neurocomputational models of memory, we assessed recall by initializing the model's activity to a random state rather than by initializing it to a noisy or partial stored memory pattern, or by biasing the network's dynamics by applying an external input pattern that represents a noisy or partial stored pattern. We found that when moderate decay rates were used, this approach resulted in a short-term memory capacity of two to three items, a value that is comparable to what has been observed in experimental studies by others (Baddeley, 2000 Cowan, 2001 Cowan et al., 2005) and that matches the memory capacity that we observed in a group of human subjects performing a similar running memory task. The model also showed a prominent recency effect, as would be expected given the use of weight decay and as is also seen in human subjects.

It should also be noted here that this model is intended to simulate short-term memory processing only, and so it is not intended to address any processes by which semantic or other long-term memory information is accessed to aid storage or recall. Indeed, it has been well established that short-term memory capacity is higher for familiar items for which long-term representations exist, compared to novel stimuli (Hulme, Maughan, & Brown, 1991 see Cowan, 2001, for review). This benefit is likely due to the fact that for novel stimuli, representations must be created before retention can successfully occur. As the model made no assumptions regarding the relationship between short-term and long-term memory, the choice of letters as the retained stimuli seemed a reasonable assumption. Still, we recognize that more complex models of short-term memory could benefit from expansion by making predictions regarding the important role of long-term memory for temporarily retained information.

Also, although the computational model is oscillatory in nature (though not periodic), it is not intended to make any predictions regarding frequency-specific responses actually produced in the brain during short-term memory retention. There is a rich and expanding research literature showing that EEG activity in the theta band plays a key role in the maintenance of temporarily retained information (Jensen & Tesche, 2002) and that hippocampal structures play a significant role in this frequency component (Kahana, Deelig, & Madsen, 2001 Rizzuto et al., 2003). There has also been recent research to show that higher-frequency oscillatory activity (i.e., gamma response) increases approximately linearly with increased memory load (Howard et al., 2003), and that greater gamma activity during encoding predicts greater likelihood of later recall (Sederberg, Kahana, Howard, Donner, & Madsen, 2003). The results of our model are promising in suggesting that oscillatory neural models can show similar capacity limitations as with humans, but they do not allow us to make predictions regarding frequency-specific contributions to EEG, especially as the model oscillations recorded are in terms of the extent to which a specific distributed memory pattern is present (quantity sλ) and not in terms of amount of network activity. While the model does retain information about stimulus order—indeed, stimulus order effects emerge in the model as a result of gradual decay—it does not address issues of temporal sequencing. As recent single-unit recording evidence suggests that ordered sequences of activation are observed in rats (Foster & Wilson, 2006), this may be an interesting area of future expansion of the model.

Our study adds to a rapidly growing literature on computational models of short-term memory by examining the role of weight decay on simple oscillatory memories. Many past models of short-term memory have employed lateral inhibition between representational units to establish competition between activated entities, and thus capacity limitations. For example, Haarmann and Usher (2001) present a model of semantic short-term memory that functions in this fashion. Our approach differs in not explicitly building in such lateral inhibition (although inhibitory weights do occur during pattern storage), with competition between memory patterns arising in the dynamics due to the interference occurring between the nonorthogonal memory patterns. Other recent models of short-term memory, inspired by specific neuroanatomical structures, have used separate modules for memory representation, maintenance, and selective gating. For example, Frank, Loughry, and O'Reilly (2001) and O'Reilly and Frank (2006) incorporate modules representing prefrontal cortex and basal ganglia, with the latter modulating which sensory stimuli are kept active. Our approach does not use a complex architecture or gating mechanisms, and thus shows that some basic behavioral properties of human short-term memory (limited memory capacity, recency effect, and shifts in position-specific stimulus recall) can be captured by a surprisingly simple neurocomputational mechanism. Still other recent short-term memory models have been based on modulation of persistent neuronal firing by rhythmic changes to membrane potential at theta frequencies (Koene & Hasselmo, 2007, 2008). Our approach is quite different in that storage is based primarily on synaptic connectivity, and memory capacity limitations arise mainly due to synaptic decay and pattern interference. However, it is an interesting question as to whether the dynamic thresholds in our model might correspond to the changes in effective thresholds brought about by the modulating theta activity in these more physiologically realistic implementations.

Perhaps the most interesting finding with the model is that by adjusting just the weight decay rate, one can produce shifts in the model's memory capacity and position-specific recall rates, as is demonstrated in Figure 4. This represents a prediction of the model that by adjusting the decay rate, one could reasonably match the shifts exhibited by human subjects who were instructed to recall different-length stimuli sequences. This would be especially remarkable given the simplicity of our model and that it requires adjustment of only a single parameter. This prediction relates to long-standing issues in the cognitive science literature concerning the nature of forgetting. For example, one view of forgetting is that short-term memory is subject to decay (Brown, 1958), while an alternative view is that forgetting is due to interference between competing elements that are simultaneously vying for attention (Waugh & Norman, 1965). Our model incorporates both interference and decay as mechanisms for forgetting and shows that the latter can partially mitigate effects from interference, consistent with evidence in past behavioral studies (Altmann & Gray, 2002).

In general, it is difficult to map processes in neurocomputational models to cognitive processes, but sometimes there are analogs that are worth considering. The observation that adjustments to decay rate control not only the total short-term memory capacity (see Figure 5) but also position-specific stimulus recall rates (see Figure 4) raises the issue of whether altering decay rate might be a useful mechanism permitting a cognitive system to control short-term memory characteristics. Specifically, our model is consistent with the hypothesis that dynamic adjustments to activity decay rate may be an important aspect of the human attention mechanisms that control forgetting (Altmann & Gray, 2002).

It is already well established that attention is a cognitive property that can be manipulated based on the needs of the task at hand (Broadbent, 1982 Downing & Pinker, 1985 Eriksen & St. James, 1986) and that attentional scope can be adjusted during visual search and memory recall between being more focused or more diffuse (Engle, 2002 Kane & Engle, 2002). Based on our modeling results, we hypothesize that altering the decay rate could serve as a means by which attentional mechanisms could act to manipulate attentional scope. More focused attention is simulated in the model by a higher decay rate, so that attention is directed more intently on a smaller number of items. In this way, decay is used as a means for combating proactive interference, with higher decay rates leading to more effective retention of recent information, but also at the expense of that which was presented before it.

For the running memory span task used here involving rapid presentation of stimuli, human subjects attempt to hold presented stimuli in a limited-capacity memory without the use of rehearsal (Bunting et al., 2006). Assuming that maintaining such stimuli depends on attentional resources, then changing instructions requiring subjects to retain varying numbers of stimuli (i.e., not just six as we did in our behavioral experiments) would be expected to have a great effect. Specifically, if attention is drawn sufficiently thin so that activation maintenance is small across all retained stimuli (a low decay rate in our model), then with longer stimulus sequences (e.g., a task requiring human subjects to recall 12 stimuli) few or none of the stimuli would be expected to retain activation levels above some cognitive threshold required for successful recall due to interference, although no doubt some attenuated recency effect will still be present. This is both a surprising and informative prediction from the model, and it suggests that overloading subjects' attentional resources (i.e., drawing attention sufficiently thin) has a detrimental effect on retention. Future behavioral testing with a varying-length recall task could therefore either refute or strongly support the model we have presented here.

To our knowledge, the work reported here is the first comparison of simple oscillatory model properties to human behavioral data. While the results are encouraging, they leave open a number of issues that should be examined in future work. Perhaps the most pressing issue concerns the generality of these results. It will be important to determine whether similar correspondences between model properties and human behavior can be produced as readily with other stimulus sets, while varying the number of stimuli that subjects are instructed to retain and recall, and across other short-term memory tasks. For example, it would be useful to compare the model's results against human data using unfamiliar or abstract symbols or nonwords where subject performance is based on free recall. Similarly, with the model, it would be useful to examine how long-term memory influences the results of recall. Another future issue, not examined in our work, would be to characterize model behavior in the absence of interference using orthogonal stimulus patterns, allowing the effects of weight decay to be studied in isolation.


Neuronal firing patterns outweigh circuitry oscillations in parkinsonian motor control

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Find articles by Pan, M. in: JCI | PubMed | Google Scholar | />

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Find articles by Lai, W. in: JCI | PubMed | Google Scholar | />

1 Department of Medical Research and

2 Department of Neurology, National Taiwan University Hospital, Taipei, Taiwan.

3 Department of Neurology, National Taiwan University Hospital, Yun-Lin Branch, Yun-Lin, Taiwan.

4 Department of Neurology, College of Physicians and Surgeons, and

5 Department of Physiology and Cellular Biophysics, Columbia University, New York, New York, USA.

6 Department of Psychology, National Taiwan University, Taipei, Taiwan.

7 Neurobiology and Cognitive Science Center and

8 Department of Physiology, National Taiwan University, Taipei, Taiwan.

Address correspondence to: Chung-Chin Kuo, Departments of Physiology and Neurology, National Taiwan University College of Medicine, No. 1, Sec. 1, Ren-Ai Road, Taipei 100, Taiwan. Phone: 886.2.23123456 ext. 88236 E-mail: [email protected]

Published October 31, 2016 - More info

Neuronal oscillations at beta frequencies (20–50 Hz) in the cortico-basal ganglia circuits have long been the leading theory for bradykinesia, the slow movements that are cardinal symptoms in Parkinson’s disease (PD). The beta oscillation theory helped to drive a frequency-based design in the development of deep brain stimulation therapy for PD. However, in contrast to this theory, here we have found that bradykinesia can be completely dissociated from beta oscillations in rodent models. Instead, we observed that bradykinesia is causatively regulated by the burst-firing pattern of the subthalamic nucleus (STN) in a feed-forward, or efferent-only, mechanism. Furthermore, STN burst-firing and beta oscillations are two independent mechanisms that are regulated by different NMDA receptors in STN. Our results shift the understanding of bradykinesia pathophysiology from an interactive oscillatory theory toward a feed-forward mechanism that is coded by firing patterns. This distinct mechanism may improve understanding of the fundamental concepts of motor control and enable more selective targeting of bradykinesia-specific mechanisms to improve PD therapy.

A cardinal feature of Parkinson’s disease (PD) is slow movements, known as bradykinesia. The neuronal activities related to bradykinesia are two electrophysiological landmarks in PD: oscillations, the pathological augmentation of cerebral field activities in beta frequencies (20–50 Hz) between cortex and subthalamic nucleus (STN) ( 1 – 9 ), versus codes, the excessive burst-firing patterns in STN ( 10 – 13 ). The leading hypothesis has long been that beta oscillations underlie bradykinesia, supported by the fact that beta power correlates with bradykinesia severity ( 6 – 9 ) and injecting beta electric activities into cortex ( 14 , 15 ) and STN ( 13 , 16 ) worsens motor performances. The oscillatory theory has deeply impacted PD therapy development and has served as important conceptual basis for deep brain stimulation (DBS) ( 16 – 18 ).

However, we recently found that excessive STN bursts, the abnormal codes in PD, can also lead to bradykinesia ( 10 , 11 ). The generation of STN bursts requires T-type calcium channels (CaTs), which are the intrinsic ion channels in STN serving as burst initiator ( 19 ). The cortex regulates STN bursts via NMDAergic cortico-subthalamic transmission ( 12 ), which also generates beta oscillations ( 12 ). The new understandings of the online modulatory mechanism of both STN bursts and beta oscillations open the window to approaching the fundamental question, what is the mechanism directly responsible for bradykinesia: the frequency-dependent oscillations or STN bursting codes? We therefore applied online modulations by selectively manipulating STN bursts and beta oscillations in 6-hydroxydopamine (6-OHDA) hemiparkinsonian rat models ( 10 – 13 , 20 ) and investigated their effects on bradykinesia (Figure 1, Supplemental Figure 1, and Supplemental Video 1 supplemental material available online with this article doi:10.1172/JCI88170DS1).

Behavioral and real-time neuronal abnormalities in 6-OHDA–lesioned parkinsonian rat model. (A) Scheme of the experimental design. Rats received surgical placements of drug infusion cannula coupled with a stimulating electrode, and recording electrodes in STN and cortex. Implanted rats received part or all of the following evaluations, including locomotor behaviors (open-field free movements and rotarod forced movements), single-unit recordings, and LFPs, before and after pharmacological and/or electrical manipulations. inj., injection Str, striatum GP, globus pallidus SNc, substantia nigra pars compacta. (B and C) Locomotor behaviors. 6-OHDA rats developed motor deficits, especially slow movements (bradykinesia), in both (B) free-moving and (C) forced-moving paradigms (n = 11 in both paradigms). (D) STN firing patterns. 6-OHDA rats developed excessive burst firings in STN, while the intra-burst profiles remained unchanged (n =10). (EG) Oscillatory profiles. (E and F) In situ synchronization of oscillatory activities presented as LFPs. STN and cortical power in beta frequencies (20–50 Hz) were pathologically increased in both resting and moving conditions in 6-OHDA rats (n = 11). (G) Long-range cortico-subthalamic oscillations presented by time-coherence plot. 6-OHDA rats developed robust oscillations in beta frequencies (n = 11). Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM **P < 0.01.

Modulating intrinsic burst-firing properties of STN dissociates beta oscillations from bradykinesia. Beta oscillations and STN bursts, the two bradykinesia-generating candidates, either can work synchronously in a cascade, or one of them is an epiphenomenon. We first manipulated STN bursts while observing beta oscillations and bradykinesia. Taking advantage of our previous studies ( 10 , 11 ), we suppressed or facilitated STN bursts by manipulating CaTs. Applying CaT blockers (NiCl2 and mibefradil) into STN suppressed STN bursts and also remedied bradykinesia in 6-OHDA hemiparkinsonian rats (Figure 2, A–F, Supplemental Figure 2, and Supplemental Video 2). However, the oscillatory profiles remained unchanged, including in situ beta synchronization of STN and cortex (Figure 2, G and H) and cortico-subthalamic oscillations (Figure 2, I and J) in both resting and moving conditions. These results clearly demonstrated that STN bursts do not cause bradykinesia via beta oscillations, and oscillation and bradykinesia could be dissociated in PD models. Consistently, application of constant hyperpolarizing current (HC) into STN can increase CaT availability ( 10 ) and burst discharges (Figure 3, A and B), which sufficiently recapitulated bradykinesia in normal rats without generating beta oscillations (Figure 3, C–J). Instead, HC further suppressed regional beta power in STN (Figure 3G), suggesting that HC augments automaticity of individual STN neurons, and therefore weakens synchronization between nearby neurons and suppresses beta power. These results provided direct evidence that STN bursts, not beta oscillations, are the immediate mechanism of bradykinesia. STN bursts are either the downstream player in the bradykinesia-generating cascade or an isolated mechanism independent of beta oscillations.

Suppression of burst-generating capacity in STN rescues bradykinesia but not beta oscillations in 6-OHDA–lesioned hemiparkinsonian rats. (A) Subthalamic infusion of NiCl2 (Ni 2+ ), a T-type calcium channel blocker, in 6-OHDA rats. (B) Sample sweeps of single-unit recordings and quantitative burst analysis. Ni 2+ suppressed burst firings without changing the intra-burst profiles in 6-OHDA rats (n = 25). (CF) Behavioral assessments. (C) Typical traces showing Ni 2+ effects in free-moving activities. Ni 2+ ameliorated (D) motion difficulties and (E) asymmetries (n = 9). Note that moving velocity was rescued in both (D) free-moving and (F) forced-moving (n =8) paradigms. (GJ) Oscillatory profiles. (G and H) In situ synchronization. Ni 2+ had no effect on STN or cortical powers in beta frequency (20-50 Hz) in both rest and moving conditions. Quantitative analysis in these rats showed no change in beta power (bar plots, n = 11). (I) Long-range cortico-subthalamic oscillations. Dark gray section of the bar above indicates Ni 2+ infusion. (J) Quantitative analysis of coherence shows that Ni 2+ did not change the pathological state of interlocking power (right panel) or frequency (left panel) in beta ranges. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P < 0.05, **P < 0.01.

Augmentation of burst-generating propensity in STN recapitulates bradykinesia but not beta oscillations in normal rats. (A) Scheme showing subthalamic application of constant HCs in normal rats. (B) HC increased burst rates in STN (n = 15) and left intra-burst profiles unchanged. (CF) Behavioral assessments. (C) Typical free-moving traces showing HC effects. HC transformed normal rats into hemiparkinsonian states and recapitulated (D) motion difficulties and (E) asymmetries (n = 11). The capacity of fast movements was also compromised in (F) the forced-moving paradigm (n = 6). (GJ) Oscillatory profiles. (G) HC further suppressed LFPs in STN, instead of reinforcing beta power mimicking the parkinsonian state (n = 13). (H) HC also remotely suppressed cortical beta power. (I and J) The cortico-subthalamic oscillations remained unsynchronized. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P <0.05, **P < 0.01, ***P < 0.001.

Bursting codes and beta oscillations are mediated by different NMDA receptor subtypes in STN. We then examined whether beta oscillations are the upstream regulator of STN bursts in the bradykinesia-generating cascade. Beta oscillations depend on NMDAergic cortico-subthalamic transmission ( 12 ) we thus selectively inhibited NMDA receptor (NMDAR) containing the GluN2A subunit, which has the fastest kinetics in the beta range ( 21 ). Subthalamic application of CPP, a GluN2A antagonist, markedly suppressed oscillatory profiles in 6-OHDA rats but had no effect on either bradykinesia or STN bursts (Figure 4, A–D, and Supplemental Figure 3, A–D). The results directly showed that beta oscillations are not involved in the genesis of STN bursts. Therefore, beta oscillations are not the upstream regulator in the bradykinesia-generating cascade. We then examined the roles of the GluN2B and GluN2D subunits. Blockers (Ro 25-6891 [RO] and PPDA) of the GluN2B/D receptors specifically suppressed STN bursts and ameliorated bradykinesia, but oscillatory profiles remained unchanged (Figure 4, Supplemental Figures 3 and 4, and Supplemental Video 3). These results demonstrated that bursting codes and beta oscillations are two parallel mechanisms regulated by different NMDARs in STN. In striking contrast to the oscillatory hypothesis, beta oscillations are not even involved in the bradykinesia-generating cascade and STN bursts play an independent role in bradykinesia.

Differential contributions of NMDAR subtypes in bradykinesia and electrophysiological profiles of 6-OHDA rats. (AD) CPP, a selective GluN2A blocker, had no effect on (A) locomotor behaviors (n = 18) or (B) STN burst firings (n =29), but dramatically reduced both (C) in situ synchronization and (D) cortico-subthalamic oscillations (n = 9). (EH) RO, a selective GluN2B/D blocker, preferentially inhibiting GluN2B subunit, rescued (E) parkinsonian motor deficits (n = 13), suppressed (F) pathological bursts in STN (n =38), but had no effect on (G) in situ synchronization or (H) cortico-subthalamic oscillations (n = 11). Also refer to Supplemental Figure 3 for additional behavioral and single-unit profiles. Statistical analyses were performed using a nonparametric Wilcoxon signed-rank test. Data are presented as mean ± SEM *P < 0.05, **P < 0.01.

Bradykinesia is regulated by STN bursts via a feed-forward mechanism. We next investigated the mechanisms by which STN bursts could lead to bradykinesia. Beyond beta oscillations, STN bursts may still involve other forms of frequency-dependent mechanisms. Without the phase synchronicity of nearby neurons detected as beta oscillations, individual STN neurons may still require regular NMDAergic inputs to generate meaningful bursts and thus bradykinesia. To evaluate whether transsynaptic regularity modulates bradykinesia, we optogenetically activated cortico-subthalamic axonal terminals by illuminating STN in Thy1-ChR2 transgenic mice ( 12 , 13 ) (Figure 5A) with either fixed-frequency (10 Hz) stimulation or frequency-independent shuffles (Figure 5B). The two stimulation protocols were of the same stimulation loads (10 pulses/s) and generated robust and similar motor deficits in normal mice (Figure 5, C–I, Supplemental Figure 5, and Supplemental Video 4). The results indicated that motor inhibition is not only dissociated from low-frequency oscillations of grouped neuronal activities, but also independent of the regularity of action potentials transmitted in the cortico-subthalamic axons.

Bradykinesia is independent of the regularity of cortico-subthalamic transmissions. (A) Schematic illustration of fiber optic cannula implanted in STN for the stimulation of cortico-subthalamic axonal terminals in Thy1-ChR2 mice. (B) Illustration of two illumination protocols: fixed 10 Hz stimulation and randomized frequency shuffles with the same stimulation loads (10 pulses/s). (CI) Sample traces of locomotor behaviors and corresponding statistic results, showing that fixed-frequency and randomized (Rand) stimulation recapitulated bradykinesia with similar severity, quantified by (E–G) motion difficulties and (H and I) asymmetries (n = 4). Also refer to Supplemental Figure 5 for thermodynamic controls, which followed the same protocols with non-activating yellow light (589 nm) laser. Statistical analyses were performed using 1-way ANOVA with post-hoc Bonferroni correction. Data are presented as mean ± SEM *P < 0.05, **P < 0.01, ***P < 0.001.

The timing of shuffled illumination was completely artificial and unpredictable, and therefore minimized the opportunity for the circuitry to adapt from the feedback interaction. Our results strongly suggested that once STN bursts are generated, the circuitry passes this information via a feed-forward mechanism to the downstream nuclei and no longer requires continuous NMDAergic monitoring from cortex to STN. To test this hypothesis, we gave HC to generate STN bursts while simultaneously blocking NMDAergic cortico-subthalamic transmission (Figure 6A). HC sufficiently induced STN bursts and bradykinesia in normal animals, but the bradykinesia-generating effect of HC could not be rescued by interrupting NMDAergic cortico-subthalamic transmission (Figure 6, B–D, and Supplemental Figure 6A). The same principles also apply to 6-OHDA rats, which had excessive STN bursts (Figure 1D and Supplemental Figure 7A) ( 10 – 12 ). Blocking NMDAergic transmission in STN suppressed bursts and rescued bradykinesia. However, HC sufficiently restored STN bursts and eliminated the therapeutic effects of NMDAR blockers (Figure 6, E–H, and Supplemental Figure 7). Once more, animal behaviors in both normal and 6-OHDA rats only aligned with STN firing patterns. Oscillatory profiles either remained unchanged or were contrary to the profiles that would be predicted by the current oscillatory theory (Supplemental Figures 6 and 7). These results clearly indicated that once STN bursts are generated, bradykinesia no longer requires ongoing cortical regulation, regardless of its synchronicity (Figures 2–4), regularity (Figure 5), or continuity (Figure 6). These characteristics reveal a delicate feed-forward role of STN bursts in bradykinesia, and are compatible with the fast-acting and quick-responsive nature of motor execution. They are also consistent with the race model of basal ganglia circuity in the STN–substantia nigra pars reticulata (SNr) axis ( 22 ), which is immediately downstream from the STN. The interruption of planned actions in the STN/SNr axis has a critical gate of timing, and intervention beyond this time point fails to stop motor execution ( 22 ). In contrast to the feed-forward mechanism shown in bradykinesia, beta oscillations have a feedback nature that requires continuous reciprocal interactions in the circuitry. Inhibition of GluN2A transmission at the cortico-subthalamic terminals in STN sufficiently disrupted beta power in the “upstream” cortex (Figure 4C). Consistently, HC suppressed beta power in STN locally and also remotely in the cortex (Figure 3, G and H).

Bradykinesia is independent of the continuity of NMDAergic cortico-subthalamic transmissions. (A) Schematic illustration of HC application in STN of a normal rat, with or without simultaneous microinfusion of AP5, a nonselective NMDAR blocker. (BD) Sample traces and quantitative analysis of locomotor activities. Bradykinesia can be recapitulated in normal rats by subthalamic HC application, but additional NMDAergic cortico-subthalamic interruption cannot reverse bradykinesia (n = 12). (EH) Similar settings in 6-OHDA rats, showing that the therapeutic effect of NMDAergic interruption by AP5 can be abolished by additional HC application in STN (n = 15). Statistical analyses were performed using 1-way ANOVA with post-hoc Bonferroni correction. Data are presented as mean ± SEM *P < 0.05, **P < 0.01, ***P < 0.001.

Taken together, our results show that STN bursts control bradykinesia via a feed-forward mechanism. Nevertheless, it should be noted that we focused on the fastest cortico-subthalamic “hyperdirect” pathway in this study. It is evident that the slower indirect pathway eventually gets involved and tunes the motor behaviors in the later steps ( 20 ). It should also be noted that this study targeted bradykinesia, a cardinal involuntary movement in PD. Although not related to bradykinesia, beta oscillations seemed to reflect the volitional aspect of motor decision states (moving versus resting states Figures 2 and 4), regardless of the motor performances with or without NMDAR/CaT modulations (see Discussion).

We discovered that bradykinesia is regulated by STN bursting codes in a feed-forward mechanism and can be completely dissociated from beta oscillations. STN bursts and beta oscillations are two parallel mechanisms controlled by different NMDARs in STN. In this study, we quantified bradykinesia (slow movements) by velocity measurement, which has been reliably used in other studies in the 6-OHDA model ( 20 , 23 – 26 ). Nevertheless, there are other behavioral tests linked to bradykinesia and other motor deficits in PD ( 27 , 28 ), and these deserve further investigation.

Differential distributions of NMDARs and CaTs in STN and their potential impacts. The burst-generating cascade requires the collaboration between GluN2B/D NMDARs and CaTs in STN, while beta oscillations depend on GluN2A NMDARs. Consistent with the clear-cut dissociation in electrophysiology, we found differential distributions of NMDAR subtypes in the STN neurons of 6-OHDA rats (Supplemental Figure 8). The oscillation-contributing GluN2A subunits were diffusely expressed in STN soma (Supplemental Figure 8, A–C), while the burst-generating GluN2B/D subunits had punctal expression extended to STN neurites (Supplemental Figure 8, D–I). Also, both GluN2B/D and CaTs (e.g., CaV3.3, the predominant CaT subtype in STN) ( 29 ) had the characteristic punctal pattern and distribution (Supplemental Figure 9), supporting their collaborative role in burst generation. It is interesting that NMDAR subtypes and CaTs are differentially segregated according to their burst- or oscillation-generating roles. Beta oscillations have been linked to the tightly time-locked STN firings in response to NMDAergic cortico-subthalamic transmission ( 12 ). The fast kinetics of GluN2A NMDARs and their prominent somatic expression could provide better temporal precision and larger positive currents near the axon hillock and thus facilitate the time-locked STN responses ( 12 ). This kinetic profile may also partly explain why the oscillations fall into beta frequencies, which are the frequency range of GluN2A kinetics ( 12 , 21 ). Bursting cells such as STN neurons and Purkinje cells (PCs) were shown to have their CaT currents initiated in dendrites ( 30 , 31 ). In theory, dendrites have less capacitance than soma and permit wider voltage fluctuations locally to unleash inactivated CaTs. Therefore, activation of GluN2B/D NMDARs in dendrites may initiate CaT-dependent bursts in STN. Although the causal relationship remains to be established, our data suggest that receptor distributions may contribute to the dissociation between bursts and oscillations.

Potential interactions between oscillation-based motor preparation and firing pattern–based motor execution in PD. Bradykinesia is characterized by slow movements beyond volitional control and therefore serves as a prototypical disease model of motor execution. Other than bradykinesia and STN bursts, the volitional motor controls, including motor decision or preparation, heavily modulate beta oscillations in PD ( 4 , 7 , 32 ). This concept is supported by the observations in our free-moving paradigm, which revealed the typical shift of oscillatory frequencies (Figure 2J and Figure 4H) and the reduction of beta powers (Figure 2, G and H, and Figure 4G) when the rats decided to move, regardless of whether their motor performances changed due to CaT/NMDAR manipulations. STN has two parallel mechanisms, oscillations and firing patterns, to modulate movements. The interplay of the two mechanisms could explain how oscillation-based motor preparation tunes the firing pattern–based motor execution. Cortical oscillatory activities can be transmitted to STN via a GluN2A-mediated mechanism, which may oscillate the somatic membrane potentials in STN and disturb the precise timing of firing-pattern switches for motor execution. Motor preparation evidently desynchronizes cortical oscillatory activities ( 4 , 7 , 33 , 34 ), which may suppress the above-mentioned processes and result in better motor execution. Similar mechanisms are well documented in the thalamus, the homolog of STN in developmental biology ( 35 ). Sleep induces thalamic oscillations and interferes with sensory information relays ( 36 – 38 ). Based on the fact that STN receives first-order command directly from the cortex via cortico-subthalamic pathway, it may be one of the fastest and the key mechanisms to explain how volitional motor preparation talks to the automatic/involuntary motor execution, and deserves further investigation by preparation-triggered protocols other than the free-moving paradigms in this study. By revisiting the oscillatory theory with the results in this study, parkinsonian motor control may be divided into two steps: the feedback, interactive oscillations for volitional motor decisions and preparations and fast-responsive, feed-forward neuronal codes for involuntary motor execution, which results in bradykinesia. We did not investigate the interactions between beta oscillations and motor decisions in this study. However, PD patients have significant problems in decision making ( 39 , 40 ), and this study provides the mechanism of beta oscillation that may help in further investigation of this issue.

The roles of neuronal codes versus oscillations in regulating normal motor behaviors. This study focused on the online circuitry mechanism of bradykinesia in PD. Notably, STN bursts and beta oscillations, which are pathologically augmented in PD, also exist in normal motor circuitry. Physiological amounts of STN bursts and beta oscillations are both present in normal conditions ( 12 , 13 ). Reduction of cortical oscillatory activities in beta frequencies (beta desynchronization) is also observed in normal motor preparations ( 33 , 34 , 41 – 43 ). Our study showed that the feed-forward, burst-coded mechanism also regulated inhibitory motor execution in rodents with intact basal ganglia circuits (Figures 3, 5, and 6, and Supplemental Figure 6) and is independent of oscillatory profiles. Naive rats also had cortical beta desynchronization in moving scenarios (Figure 3H), regardless of whether the motor performances were being modulated by HC. The patterns and distributions of NMDARs and CaTs of STN in naive rodents were also similar to those in 6-OHDA rat models (Supplemental Figures 10 and 11). Beyond PD pathophysiology, these results may also apply to physiological states and improve our understandings of the fundamental principles of motor control physiology. Feedback circuitry oscillations may contribute to the volitional aspects of motor commands, while the feed-forward, firing pattern–coded neurotransmissions regulate motor execution (schematic summary, Supplemental Figure 12.

Therapeutic potential based on the new mechanism of bradykinesia. Our results shows that bradykinesia requires the collaboration between GluN2B/D NMDARs and CaTs in STN. Amantadine ( 44 ) and zonisamide ( 45 ), a weak NMDAR and a CaT blocker, respectively, already show modest clinical benefits in PD patients. However, potent and nonspecific NMDAR or CaT blockers are not ideal therapeutic options due to their cognitive side effects. In contrast, GluN2D ( 46 ) and CaV3.3 ( 29 ) have low expression levels in the neocortex but are enriched in PD STN (Supplemental Figure 8 and 9). Regardless of the changes in dopaminergic system and direct-indirect pathways ( 20 , 47 ), targeting of the neuron-modulatory consequences via GluN2D and CaV3.3 may provide better therapeutic options. The standard dopaminergic therapy in PD is notorious for its motor complications ( 48 ). In contrast, NMDAR and CaT blockers did not induce the paradoxical rotations and head tilts (Figures 2 and 4) typically provoked by dopaminergic agents ( 12 ). Amantadine is the best-known anti-dyskinetic therapy in PD ( 44 ). In fact, intervention in the cortico-subthalamic pathway is the key mechanism of DBS ( 12 , 13 ), and the therapeutic effect of DBS is better than the traditional therapy in terms of motor complications ( 49 ). Moreover, systemic dopaminergic therapy contributes to major cognitive and impulse control problems in PD ( 40 , 48 ). Therefore, GluN2D and CaV3.3 could be new bradykinesia-specific targets for PD motor therapy, and may be superior to the standard dopaminergic therapy in terms of its motor and cognitive complications.

Animals. Male adult Wistar rats were entered into the study at

8 weeks of age and 250–350 g. 6-OHDA–lesioned (Sigma-Aldrich) hemiparkinsonian rats were used in all the PD experiments in this study (Supplemental Figure 1 see also Supplemental Methods). For optogenetics experiments, we used male adult Thy1-ChR2-EYFP line 18 transgenic mice (catalog 007612 The Jackson Laboratory), which express channelrhodopsin-2 in cortical neuron layer V ( 13 ) and have been validated as an ideal animal model for selective stimulation of cortico-subthalamic axons ( 12 , 13 ). Mice were entered into the study at

5 weeks of age and weights greater than 20 g. The animals were housed in a vivarium with controlled 12-hour dark/light cycles.

NMDAR and CaT modulators. We used NMDAR blockers with different subunit specificities. (D)-AP5 (2 mM, Tocris) is a non-selective NMDAR blocker. (R)-CPP (200 μM, Tocris) is a selective NMDAR antagonist targeting the GluN2A subunit. Ro 25-6891 (RO 1 mM, Tocris) and PPDA (500 μM, Tocris) inhibit GluN2B/D subunit selectively. To inhibit CaTs, we selected NiCl2 (6 mM, Sigma-Aldrich) and mibefradil (500 μM, Tocris). PPDA was dissolved in DMSO to 50 mM first and then diluted with saline to achieve a final concentration of 500 μM. All the other drugs were dissolved in artificial CSF (aCSF). The pH of all solutions was adjusted to 7.4.

Behavioral recordings and in vivo electrophysiology. We used the open-field test to evaluate the free-moving locomotor behaviors in 6-OHDA and control rodents, and the rotarod test for forced-moving behaviors (see Supplemental Methods for detailed paradigms). In valid 6-OHDA or normal control rats, we implanted microwire deep electrodes for single-unit and local field potential (LFP) recordings, as well as applying HCs. Epidural screw electrodes were also implanted for cortical LFPs. An STN cannula was inserted ipsilateral to 6-OHDA lesioning for real-time NMDAR or CaT modulations (see Supplemental Methods for all surgical procedures). We performed simultaneous behavioral and electrophysiological recordings, before and after online electric and/or pharmacological manipulations. Single-unit firings and LFPs were prefiltered and analyzed separately. For details, see Supplemental Methods.

Optic stimulation and simultaneous behavioral recordings. We activated the cortico-subthalamic axons optogenetically by implanting an optic fiber unilaterally into STN in Thy1-ChR2 mice (Figure 5A). We applied two different protocols: frequency-dependent (10 Hz) and frequency-independent (randomized) illumination (Figure 5B). Free-moving behaviors were accessed under baseline, light-off, and light-on states, with one of the stimulation protocols applied first by random process. For details, see Supplemental Methods.

Analysis of single-unit recordings and LFPs. Signals recorded for single-unit settings were post-processed with spike-sorting software (SciWorks 8.0, DataWave Technologies) and quality-controlled algorithm ( 12 ). Burst patterns were detected in each qualified single unit as described previously ( 10 – 12 ). The LFP data were post-processed with MATLAB 7.4 (MathWorks). Regional power spectrum represented in situ synchronizations, while coherence analysis referred to long-ranged synchronization. For details, see Supplemental Methods.

Immunohistochemistry. Three-month-old adult C57BL/6J mice and 4-month-old Wistar rats with or without 6-OHDA lesioning were used for immunohistochemistry study. Mice and rats were anesthetized with isoflurane and then sacrificed by an overdose of urethane (2 g/kg i.p., Sigma-Aldrich) and transcardially perfused with 4% paraformaldehyde in PBS. The brain was then removed and immersed in the 4% paraformaldehyde overnight and moved to PBS for 3 days. The brain was sliced coronally at the thickness of 30 μm by vibrotome. The sections were washed with PBS, followed by the suppression in 10% normal donkey serum in 0.1% Triton. The sections were subsequently incubated with respective primary antibodies overnight at 4°C and then secondary fluorescent antibodies (all from Invitrogen). Primary antibodies included GluN2A (Neuromab, Cat. No. 75-288), GluN2B (Neuromab, 75-097), GluN2D (Bioss, Bs-1072R), MAP2 (Abcam, Ab5392), CaV3.1 (Alomone Labs, ACC-021), CaV3.2 (Alomone Labs, ACC-025) and CaV3.3 (Alomone Labs, ACC-009). Images were taken using confocal laser scanning microscope (Leica TCS SP2 two-photon microscope). See also Supplemental Methods and Supplemental Figure 13 for more details.

Statistics. The statistics were managed with SPSS 13.0 and plotted with Excel 2013 (Microsoft). Nonparametric Wilcoxon signed-rank test was used to analyze paired data, including animal behaviors, single-unit recordings, and LFP analyses (Figures 1–4 and Supplemental Figures 2–4). For those data with 3 or more conditions, we applied 1-way ANOVA with post-hoc Bonferroni correction (Figures 5 and 6 and Supplemental Figures 5–7). In all statistical methods, a P value less than 0.05 was considered significant.

Study approval. The study was approved by the IACUC of National Taiwan University College of Medicine and College of Public Health.

MKP designed the study and generated the behavioral and electrophysiological results in rats with CHT, YMW, WCL, and TRW. SHK designed and generated the histology/pathology data in rodents and human subjects. MKP and WSL designed the optogenetic experiments in mice and analyzed the results with JCP and CYC. MKP and JYL designed MATLAB codes. CCK led the team and coordinated the study. MKP and CCK interpreted the results and wrote the article with input and comments from the other authors.

We thank F.-C. Lin (Southport Tech. Co.) for the technical support. This research is supported by the Ministry of Science and Technology in Taiwan (MOST-103-2320-B-002-026-MY3 and MOST-104-2321-B-002-067 to CCK MOST-104-2314-B-002-076-MY3 to MKP), the National Health Research Institute in Taiwan (NHRI-EX105-10503NI to CCK) National Taiwan University Hospital (104-N2870 and MG380 to MKP) and the Yin-Lin branch of the hospital (NTUHYL104.N007 to MKP and YMW).

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2016126(12):4516–4526. doi:10.1172/JCI88170.


Submitted

Kragel, J. E., Worrell, G. A., Sperling, M. R., Gross, R. E., Lega, B. C., Jobst, B. C., et al. Distinct cortical systems reinstate content and context information during memory search. Submitted.
(pdf, supplemental, code, Ephys data)

Sakon, J. J. and Kahana, M. J. Hippocampal ripples signal contextually-mediated episodic recall. bioRxiv.
(pdf, code, Spectrograms)

Adamovich-Zeitlin, R., Wanda, P. A., Solomon, E., Phan, T., Lega, B., Jobst, B. C., et al. (2020). Biomarkers of memory variability in traumatic brain injury. Brain Communications, 3(1).
(code, Ephys data)

Meisler, S. L., Kahana, M. J., and Ezzyat, Y. (2020). Does data cleaning improve brain state classification? Journal of Neuroscience Methods.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2020). Neural measures of subsequent memory reflect endogenous variability in cognitive function. Journal of Experimental Psychology: Learning, Memory, and Cognition.
(pdf, code, Ephys data)

Solomon, E. A., Lega, B. C., Sperling, M. R., and Kahana, M. J. (2019). Hippocampal theta codes for distances in semantic and temporal spaces. PNAS, 116(48), 24343-24352.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2019). Dynamics of brain activity reveal a unitary recognition signal. Journal of Experimental Psychology: Learning, Memory, and Cognition, 45(3), 440�.
(pdf, code, Ephys data)

Weidemann, C. T., Kragel, J. E., Lega, B. C., Worrell, G. A., Sperling, M. R., Sharan, A. D., et al. (2019). Neural activity reveals interactions between episodic and semantic memory systems during retrieval. Journal of Experimental Psychology: General, 148(1), 1󈝸.
(pdf, code, Ephys data)

Ezzyat, Y., Kragel, J. E., Burke, J. F., Levy, D. F., Lyalenko, A., Wanda, P., et al. (2017). Direct brain stimulation modulates encoding states and memory performance in humans. Current Biology, 27(9), 1251�.
(pdf, code, Ephys data)

Weidemann, C. T. and Kahana, M. J. (2016). Assessing recognition memory using confidence ratings and response times. Royal Society Open Science, 3, 150670.
(pdf, code, Ephys data)

Long, N. M. and Kahana, M. J. (2015). Successful memory formation is driven by contextual encoding in the core memory network. NeuroImage, 119, 332-337.
(pdf, code, Ephys data)

Merkow, M. B., Burke, J. F., and Kahana, M. J. (2015). The human hippocampus contributes to both the recollection and familiarity components of recognition memory. Proceedings of the National Academy of Sciences, 112(46), 14378�.
(pdf, Ephys data)

Ramayya, A. G., Pedisich, I., and Kahana, M. J. (2015). Expectation modulates neural representations of valence throughout the human brain. NeuroImage, 115, 214-223.
(pdf, code, Ephys data)

Geller, A. S., Burke, J. F., Sperling, M. R., Sharan, A. D., Litt, B., Baltuch, G. H., et al. (2014). Eye closure causes widespread low-frequency power increase and focal gamma attenuation in the human electrocorticogram. Clinical Neurophysiology, 125(9), 1764-73.
(pdf, Ephys data)

Long, N. M., Burke, J. F., and Kahana, M. J. (2014). Subsequent memory effect in intracranial and scalp EEG. NeuroImage, 84, 488�.
(pdf, Ephys data)

Ramayya, A. G., Misra, A., Baltuch, G. H., and Kahana, M. J. (2014). Microstimulation of the human substantia nigra following feedback alters reinforcement learning. Journal of Neuroscience, 34(20), 6887�.
(pdf, Ephys data)

Miller, J. F., Neufang, M., Solway, A., Brandt, A., Trippel, M., Mader, I., et al. (2013). Neural activity in human hippocampal formation reveals the spatial context of retrieved memories. Science, 342(6162), 1111-1114.
(pdf, supplemental, Ephys data)

Morton, N. W., Kahana, M. J., Rosenberg, E. A., Sperling, M. R., Sharan, A. D., and Polyn, S. M. (2013). Category-specific neural oscillations predict recall organization during memory search. Cerebral Cortex, 23(10), 2407�.
(pdf, Ephys data)

van Vugt, M. K., Sekuler, R., Wilson, H. R., and Kahana, M. J. (2013). An electrophysiological signature of summed similarity in visual working memory. Journal of Experimental Psychology: General, 142(2), 412�.
(pdf, Ephys data)

van der Meij, R., Kahana, M. J., and Maris, E. (2012). Phase-amplitude coupling in human ECoG is spatially distributed and phase diverse. Journal of Neuroscience, 32(1), 111-123.
(pdf, Ephys data)

Maris, E., van Vugt, M. K., and Kahana, M. J. (2011). Spatially distributed patterns of oscillatory coupling between high-frequency amplitudes and low-frequency phases in human ieeg. NeuroImage, 54(2), 836-850.
(pdf, Ephys data)

van Vugt, M. K., Schulze-Bonhage, A., Litt, B., Brandt, A., and Kahana, M. J. (2010). Hippocampal gamma oscillations increase with working memory load. Journal of Neuroscience, 30(7), 2694�.
(pdf, Ephys data)

Jacobs, J. and Kahana, M. J. (2009). Neural representations of individual stimuli in humans revealed by gamma-band electrocorticographic activity. Journal of Neuroscience, 29(33), 10203�.
(pdf, Ephys data)

van Vugt, M. K., Schulze-Bonhage, A., Sekuler, R., Litt, B., Brandt, A., Baltuch, G., et al. (2009). Intracranial electroencephalography reveals two distinct similarity effects during item recognition. Brain Research, 1299, 33󈞘.
(pdf, Ephys data)

Weidemann, C. T., Mollison, M. V., and Kahana, M. J. (2009). Electrophysiological correlates of high-level perception during spatial navigation. Psychonomic Bulletin & Review, 16(2), 313�.
(pdf, Ephys data)

Jacobs, J., Hwang, G., Curran, T., and Kahana, M. J. (2006). EEG oscillations and recognition memory: Theta correlates of memory retrieval and decision making. NeuroImage, 15(2), 978󈟃.
(pdf, Ephys data)

Rizzuto, D., Madsen, J. R., Bromfield, E. B., Schulze-Bonhage, A., and Kahana, M. J. (2006). Human neocortical oscillations exhibit theta phase differences between encoding and retrieval. NeuroImage, 31(3), 1352�.
(pdf, Ephys data)

Ekstrom, A. D., Caplan, J., Ho, E., Shattuck, K., Fried, I., and Kahana, M. (2005). Human hippocampal theta activity during virtual navigation. Hippocampus, 15, 881�.
(pdf, Ephys data)

Howard, M. W., Rizzuto, D. S., Caplan, J. C., Madsen, J. R., Lisman, J., Aschenbrenner-Scheibe, R., et al. (2003). Gamma oscillations correlate with working memory load in humans. Cerebral Cortex, 13, 1369�.
(pdf, Ephys data)

Rizzuto, D., Madsen, J. R., Bromfield, E. B., Schulze-Bonhage, A., Seelig, D., Aschenbrenner-Scheibe, R., et al. (2003). Reset of human neocortical oscillations during a working memory task. Proceedings of the National Academy of Sciences, USA, 100(13), 7931�.
(pdf, Ephys data)

Sederberg, P. B., Kahana, M. J., Howard, M. W., Donner, E. J., and Madsen, J. R. (2003). Theta and gamma oscillations during encoding predict subsequent recall. Journal of Neuroscience, 23(34), 10809�.
(pdf, Ephys data)

Raghavachari, S., Kahana, M. J., Rizzuto, D. S., Caplan, J. B., Kirschen, M. P., Bourgeois, B., et al. (2001). Gating of human theta oscillations by a working memory task. Journal of Neuroscience, 21(9), 3175-3183.
(pdf, Ephys data)


Watch the video: Fundamentals of neuronal oscillations and synchrony (June 2022).


Comments:

  1. Kedal

    Wacker, the ideal answer.

  2. Fitzsimon

    Is the entertaining phrase

  3. Gozshura

    I would like to encourage you to look for a site where many articles on the subject you are interested in.

  4. Angell

    You have hit the mark. In it something is and it is good idea. It is ready to support you.



Write a message