Information

MEG from simulated neurons

MEG from simulated neurons



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.

Given MEG is an approximation of neuronal spiking, are there any models that associate the spiking of simulated neurons to an MEG signal? Ideally, this would be a model with some amount of behaviour (mapping to an output, such as a motor action), but purely biological models are fine too.


Techniques that measure brain activity at a distance such as EEG and MEG measure spiking activity very poorly and indirectly. Instead, these techniques primarily measure synaptic currents.

Synaptic currents of course originate in some way from spiking activity, however:

  1. Synaptic currents may not occur in the vicinity of the spiking cells

  2. The amplitude of synaptic currents depends on the current state of the cells impinged upon

  3. Synaptic currents of different types in the same area can cancel each other out

  4. Phase, polarity, and geometry can lead to a complex inverse problem at the sensor that has an infinite number of solutions.

However, by making some rather strong assumptions, one can attempt to solve this inverse problem. Higher frequency EEG/MEG activity is more closely associated with spiking activity, and therefore most appropriate for attempting to localize spiking rather than other types of activity. See for example David and Friston 2003.

I'd also add that although the current agreement seems to be that EEG/MEG primarily come from synaptic currents, because synaptic currents do originate ultimately from some sort of spiking, EEG/MEG signals can correlate well with spiking activity measured with depth electrodes (for example Whittingstall & Logothetis, 2009). It's just likely that the correlation is indirect, mediated through synaptic potentials.


Buzsáki, G., Anastassiou, C. A., & Koch, C. (2012). The origin of extracellular fields and currents-EEG, ECoG, LFP and spikes. Nature reviews neuroscience, 13(6), 407.

David, O., & Friston, K. J. (2003). A neural mass model for MEG/EEG:: coupling and neuronal dynamics. NeuroImage, 20(3), 1743-1755.

Whittingstall, K., & Logothetis, N. K. (2009). Frequency-band coupling in surface EEG reflects spiking activity in monkey visual cortex. Neuron, 64(2), 281-289.


MEG Scanners Are Mega Powerful

To revist this article, visit My Profile, then View saved stories.

To revist this article, visit My Profile, then View saved stories.

Magneto-encephalography, or MEG, scanners are proving to be one of the most powerful tools in the hands of scientists using the machines to observe important details about epilepsy, brain tumors, emotions, pain perception and more.

While the technology has existed for decades, improved computing power and hardware have recently increased interest in the scanners. There are an estimated 100 MEG scanners around the world -- at a potential cost of $2 million each -- and their numbers are growing.

Click here for photos of brain scanning machines.

"The MEG is just exploding," said Robert Knowlton, director of the University of Alabama at Birmingham's Magnetic Source Imaging Laboratory. "It's been around a long time, but it's only now gotten into enough researcher's hands with the technology (being) mature enough."

While other types of brain scans detail the geography of the brain or detect blood flow, the MEG scanners track the magnetic signals that neurons throw off as they communicate. "You can look at how the networks of the brain are talking to each other in real time," said Greg Simpson, director of the Dynamic Neuroimaging Laboratory at the University of California at San Francisco.

The machines are designed to "measure the magnetic field pattern around the entire head and deduce from those patterns where the current flows are occurring within the brain," said Eugene Hirschkoff, vice president of engineering at 4-D Neuroimaging, a MEG scanner manufacturer.

This allows scientists to study magnetic changes in the brain and figure out which areas are busy doing things each millisecond. By contrast, functional magnetic resonance imaging, or fMRI, technology measures the movement of blood within the brain. The scans reveal which brain areas are active and need oxygen from the blood, Simpson said.

But it takes a while for oxygen-filled blood to move in the brain. "If a brain area was active for a 10th of a second, the blood-flow response to that area would take a second or two to start," said Simpson.

That may not sound like a long time, but this is the lickety-split world of the brain. "If I had you read a sentence during an fMRI scan, weɽ see the visual cortex light up, the language cortex light up, and other things that would light up," Knowlton said. "But theyɽ all be lit up, and you wouldn't know which one was first (to become active), which one was most important."

"With the MEG, the sequence becomes more clear," Knowlton said.

One thing they can't do is analyze the physical parts of the brain, so MEGs become even more powerful when combined with other technologies.

When combined with fMRI scans, in particular, "you get the best of everything," Knowlton said.

In a recent experiment, scientists tried to figure out what you might call the Grey's Anatomy effect. When characters on the show get lonely, they act out -- drinking too much at the local bar, baking up a storm in the kitchen, having sex with a guy who ends up with a permanent erection.

Self-medication? Maybe. But a team of scientists prefers to describe the tendency toward bad behavior as cognitive disruption brought on by social exclusion -- a process they think they can see with the help of the MEG.

W. Keith Campbell, a psychology professor at the University of Georgia, and his colleagues recruited 30 female students who underwent what they thought were ordinary personality tests while being monitored by an MEG scanner.

Half the women were told that their test results suggested they would have trouble maintaining relationships and "end up alone in later life." The other women, a control group, got only neutral feedback to their answers, with no indications their lives were destined to go to pot.


Neural correlation of successful cognitive behaviour therapy for spider phobia: a magnetoencephalography study

Cognitive behavioural therapy (CBT) can be an effective treatment for spider phobia, but the underlying neural correlates of therapeutic change are yet to be specified. The present study used magnetoencephalography (MEG) to study responses within the first half second, to phobogenic stimuli in a group of individuals with spider phobia prior to treatment (n=12) and then in nine of them following successful CBT (where they could touch and manage live large common house spiders) at least 9 months later. We also compared responses to a group of age-matched healthy control participants (n=11). Participants viewed static photographs of real spiders, other fear-inducing images (e.g. snakes, sharks) and neutral stimuli (e.g. kittens). Beamforming methods were used to localise sources of significant power changes in response to stimuli. Prior to treatment, participants with spider phobia showed a significant maximum response in the right frontal pole when viewing images of real spiders specifically. No significant frontal response was observed for either control participants or participants with spider phobia post-treatment. In addition, participants' subjective ratings of spider stimuli significantly predicted peak responses in right frontal regions. The implications for understanding brain-based effects of cognitive therapies are discussed.

Keywords: Brain activation Fear Magnetoencephalography Neuroimaging Psychological therapy Simple phobia.


MEG from simulated neurons - Psychology

Basic Principles of Magnetoencephalography

(click on any figure to enlarge)

Magnetoencephalography (MEG) is a non-invasive neurophysiological technique that measures the magnetic fields generated by neuronal activity of the brain (Figure 1). The spatial distributions of the magnetic fields are analyzed to localize the sources of the activity within the brain, and the locations of the sources are superimposed on anatomical images, such as MRI, to provide information about both the structure and function of the brain (Figure 2).

The principle features of MEG are:

MEG is a direct measure of brain function, unlike functional measures such as fMRI, PET and SPECT that are secondary measures of brain function reflecting brain metabolism.

MEG has a very high temporal resolution. Events with time scales on the order of milliseconds can be resolved, again differentiating MEG from fMRI, PET and SPECT, which have much longer time scales.

MEG has excellent spatial resolution sources can be localized with millimeter precision.

MEG is completely non-invasive. It does not require the injection of isotopes or exposure to X-rays or magnetic fields. Children or infants can be studied and repeated tests are possible.

MEG is complementary to other modalities, the information provided by each modality adds to the full picture.

Magnetic fields are found whenever there is a current flow, whether in a wire or a neuronal element, as illustrated in the upper left panel of Figure 3.

The magnetic field passes unaffected through brain tissue and the skull, so it can be recorded outside the head (upper middle). The magnetic field is extremely small, but can be detected by sophisticated sensors that are based on superconductivity (upper right).

By analyzing the spatial distributions of magnetic fields (lower left), it is possible, by using a model such as a single equivalent current dipole (lower middle), to estimate the intracranial localization of the generator source and superimpose it on an MRI (lower right).

The sequence of recording and analysis steps is shown, starting from 148 MEG time series waveforms containing an interictal spike (lower left panel of Figure 4). The segment of data containing the spike has been “clipped out” and shown on a display of the sensors around the head projected on a flat surface (upper left). The amplitude of the spike at its peak is shown as a topographic map (middle). The spike shows a phase reversal over the left temporal region. A calculation of the spike location using the single equivalent current dipole model reveals a localization of the spike source in the left temporal lobe (right).

The spatial and temporal properties of MEG are illustrated in the left panel of Figure 5. Only MEG has extremely high temporal and spatial resolution, as represented in the lower left section of the graph. Other functional modalities, except invasive EEG (iEEG), have either poor temporal or spatial resolution. Clearly iEEG has the distinct disadvantage of being invasive.

The main drawback of MEG is shown on the right panel of Figure 5. The MEG signals of interest are extremely small, several orders of magnitude smaller than other signals in a typical environment that can obscure the signal. Thus, specialized shielding is required to eliminate the magnetic interference found in a typical urban clinical environment.

MEG is a unique and effective diagnostic tool for evaluating brain function in a variety of surgical planning applications. The fundamental advantages of MEG are listed in Figure 6.

Figure 1. Basic principles of MEG.

Figure 2. MEG combines functional information from magnetic field recordings with structural information from MRI.

Figure 3. Electrical activity in neurons produces magnetic fields that can be recorded outside the skull and used to calculate the locations of the activity within the brain.

Figure 4. The sequence of steps to localize sources of neuronal activity from time-domain recordings to MRI overlay.

Figure 5. MEG has the advantages of very high temporal and spatial resolution, however it requires highly sensitive instrumentation and sophisticated methods for eliminating environmental magnetic interference.


Mirror neurons and embodied simulation in the development of archetypes and self-agency

In this paper I explore the role of mirror neurons and motor intentionality in the development of self-agency. I suggest that this will also give us a firmer basis for an emergent view of archetypes, as key components in the development trajectory of self-agency, from its foundation in bodily action to its mature expression in mentalization and a conscious awareness of intentionality. I offer some ideas about the implications of these issues of self-agency for our clinical work with patients whose developmental trajectory of self-agency has been partially inhibited, so that their communications have a coercive effect. I discuss the possibility that this kind of clinical phenomenon may relate to Gallese and Lakoff's hypothesis that abstract thought and imagination are forms of simulated action, and that the same sensory-motor circuits that control action also control imagination, concept formation and understanding, but with a crucial development, that of an inhibition of the connections between secondary pre-motor cortical areas and the primary motor cortex. I shall speculate that in the earlier developmental stages of self-agency, the separation of secondary from primary motor areas is not complete, so that imagination and thought are not entirely independent of physical action.


Simulating human sleep spindle MEG and EEG from ion channel and circuit level dynamics

Background: Although they form a unitary phenomenon, the relationship between extracranial M/EEG and transmembrane ion flows is understood only as a general principle rather than as a well-articulated and quantified causal chain.

Method: We present an integrated multiscale model, consisting of a neural simulation of thalamus and cortex during stage N2 sleep and a biophysical model projecting cortical current densities to M/EEG fields. Sleep spindles were generated through the interactions of local and distant network connections and intrinsic currents within thalamocortical circuits. 32,652 cortical neurons were mapped onto the cortical surface reconstructed from subjects' MRI, interconnected based on geodesic distances, and scaled-up to current dipole densities based on laminar recordings in humans. MRIs were used to generate a quasi-static electromagnetic model enabling simulated cortical activity to be projected to the M/EEG sensors.

Results: The simulated M/EEG spindles were similar in amplitude and topography to empirical examples in the same subjects. Simulated spindles with more core-dominant activity were more MEG weighted.

Comparison with existing methods: Previous models lacked either spindle-generating thalamic neural dynamics or whole head biophysical modeling the framework presented here is the first to simultaneously capture these disparate scales.

Conclusions: This multiscale model provides a platform for the principled quantitative integration of existing information relevant to the generation of sleep spindles, and allows the implications of future findings to be explored. It provides a proof of principle for a methodological framework allowing large-scale integrative brain oscillations to be understood in terms of their underlying channels and synapses.

Keywords: Cortex EEG Forward model Human MEG Sleep Spindle Thalamus.


Why Google DeepMind Is Putting AI on the Psychologist’s Couch

Artificial intelligence can now carry out many of the same cognitive tasks humans can, but we still don’t really understand how AIs think. Google DeepMind plans to train long-standing tests of human cognitive skills on machine minds to learn how they work.

A long-standing problem in AI research has been the fact that deep neural networks are “black boxes.” You can’t tell how these algorithms work just by looking at their code. They teach themselves by training on data and there’s no simple flow diagram a human can follow. The way these networks reach decisions is encoded in the weights of thousands of simulated neurons.

But they’re not the only inscrutable thinking machines. Simply poking around in the human brains they are modeled on yields few clues as to how people reason either, and so over the years, cognitive psychologists have developed tests designed to probe our mental faculties.

Now DeepMind has built a virtual 3D laboratory called Psychlab that will let machines take these tests too, and they’ve open-sourced it so any AI researcher can put their algorithms through their paces.

Psychlab is built on the company’s DeepMind Lab platform, designed for testing virtual agents in 3D environments. It recreates the set-up a human taking part in an experiment would see by providing the subject with a first-person view of a virtual computer monitor that displays a variety of classic cognitive tests.

These include measures of the ability to search for objects in a scene, detect change, remember a growing list of items, track moving objects, or recall stimulus-response pairings. Typically a human would use a mouse to respond to on-screen tasks, but virtual agents use the direction of their gaze.

By mimicking the environment a human would see, the researchers say, humans and AIs can effectively take the same tests. That should not only make it easier to draw direct comparisons between the two, but also allow results to be connected with the existing academic literature in cognitive psychology.

Being able to draw on the accumulated wisdom of 150 years of psychology research could be hugely useful in understanding how the latest AI systems work, the researchers say in a paper published on the arXiv to coincide with the release of the tool.

In recent years there’s been an increasing focus on deep-reinforcement learning AI systems that can carry out complicated tasks in simulated 3D environments. The complex nature of these environments and the variety of strategies these systems can employ to solve problems makes it hard to tease out what combination of cognitive abilities underlies their performance.

But by subjecting a state-of-the-art deep reinforcement learning agent called UNREAL to a variety of tests in Psychlab, they were able to uncover details about how its perceptual system worked, and even use this to improve its performance.

It turns out UNREAL has considerably worse acuity, or keenness of vision, than humans, which means it learns faster when presented with larger objects. Key to human acuity is a dense cluster of photoreceptors at the center of the retina called the fovea, which gives us particularly sharp vision at the center of our visual field.

By adding a simple model of the fovea to UNREAL, the researchers were able to improve the agent’s performance not just on the Psychlab experiments, but also on other standard DeepMind Lab tasks.

DeepMind aren’t the only ones applying cognitive tests to AI. University of Michigan researchers have been subjecting reinforcement learning agents to maze navigation tasks that have long been used to test memory and learning in rats.

The mazes were built in the 3D world of Minecraft, and the AI were set increasingly complex tasks and given different rewards to find out what cognitive skills were important for successfully navigating the experiment. They found that being able to retrieve memories based on the context in which they were stored was key for solving their tests.

As AI continues to improve and develop higher-order cognitive skills such as reasoning, emotional intelligence, and planning, more sophisticated psychological tests could become a crucial way for us to understand the ways their mental processes differ from ours, which they almost certainly will.

That could be an important tool to ensure everyone gets along in a future where humans and AI have to coexist.


Computational model reveals how the brain manages short-term memories

Credit: CC0 Public Domain

If you've ever forgotten something mere seconds after it was at the forefront of your mind—the name of a dish you were about to order at a restaurant, for instance—then you know how important working memory is. This type of short-term recall is how people retain information for a matter of seconds or minutes to solve a problem or carry out a task, like the next step in a series of instructions. But, although it's critical in our day-to-day lives, exactly how the brain manages working memory has been a mystery.

Now, Salk scientists have developed a new computational model showing how the brain maintains information short-term using specific types of neurons. Their findings, published in Nature Neuroscience on December 7, 2020, could help shed light on why working memory is impaired in a broad range of neuropsychiatric disorders, including schizophrenia, as well as in normal aging.

"Most research on working memory focuses on the excitatory neurons in the cortex, which are numerous and broadly connected, rather than the inhibitory neurons, which are locally connected and more diverse," says Terrence Sejnowski, head of Salk's Computational Neurobiology Laboratory and senior author of the new work. "However, a recurrent neural network model that we taught to perform a working memory task surprised us by using inhibitory neurons to make correct decisions after a delay."

In the new paper, Sejnowski and Robert Kim, a Salk and UC San Diego MD/Ph.D. student, developed a computer model of the prefrontal cortex, an area of the brain known to manage working memory. The researchers used learning algorithms to teach their model to carry out a test typically used to gauge working memory in primates—the animals must determine whether a pattern of colored squares on a screen matches one that was seen several seconds earlier.

Sejnowski and Kim analyzed how their model was able to perform this task with high accuracy, and then compared it to existing data on the patterns of brain activity seen in monkeys carrying out the task. In both tests, the real and simulated neurons involved in working memory operated on a slower timescale than other neurons.

Kim and Sejnowski found that good working memory required both that long-timescale neurons be prevalent, and that connections between inhibitory neurons—which suppress brain activity—be strong. When they altered the strength of connections between these inhibitory neurons in their model, the researchers could change how well the model performed on the working memory test as well as the timescale of the pertinent neurons.

The new observations point toward the importance of inhibitory neurons, and could inspire future research on the role of these cells in working memory, the researchers say. They also could inform studies on why some people with neuropsychiatric disorders, including schizophrenia and autism, struggle with working memory.

"Working memory impairment is common in neuropsychiatric disorders, including schizophrenia and autism spectrum disorders," says Kim. "If we can elucidate the mechanism of working memory, that's a step toward understanding how working memory deficits arise in these disorders."


3. Results

3.1. Questionnaires and psychological tests

Table 1 displays the mean values for the questionnaire and test data. The two groups did not differ regarding CPM/SPM and K-ABC scores. However, Asperger's children scored significantly lower on social competence scales and significantly higher on all social impairment scales of the CBCL. Due to experimenter error, the CBCL score was missing for one child in the Asperger's group.

3.2. MEG

3.2.1. Sensor data

The FFT spectra averaged over all sensors for the two groups separately are displayed in Fig. 2A for the repetitive and non-repetitive biological movement condition. In contrast to recent EEG studies, we found elevation of high-Mu power while observing biological movement. This power increase in the 10� Hz range was conspicuous and topographically distinct especially in the non-repetitive condition and particularly in the data recorded at frontal and central sensors in typically developed children, (see Fig. 2B ). In contrast, in both groups there was only a little change in high-Mu power in response to repetitive stimulation. Overall the group difference in power was larger for the non-repetitive than for the repetitive condition (see Fig. 2A ).

(A) Frequency distribution over all 148 MEG sensors for repetitive and non-repetitive movements separately for control (top right) and Asperger's children (bottom right). (B) Topographical illustration of the 10� Hz frequency range.

3.2.2. Source localization of group differences

Group comparisons of MEG source activity revealed that for the non-repetitive movements, the Asperger's syndrome and the control groups differed significantly in the pattern of high-Mu-modulation while no differences were found between the two groups in response to repetitive movements. Therefore, the source analysis focused on the non-repetitive movements condition. Within-group dependent t-statistics on source activity in the control group revealed enhancement in high-Mu power while observing biological movements relative to baseline in the right premotor cortex (Brodmann area, BA 6, Talairach coordinates: X: 39, Y: 𢄩, Z: 70) and a concurrent reduction in high-Mu power in right middle prefrontal cortex (BA 8, X: 32, Y: 28, Z: 38 see Fig. 3A ). In contrast, Asperger's children lacked the premotor enhancement and showed bilaterally increased activity in the prefrontal cortex (BA 8, right: X: 24, Y: 45, Z: 48, left: X: 𢄩, Y: 50, Z: 48 see Fig. 3B ). In addition, in both groups, sources in the occipital cortex showed decreased activity in the high-alpha power. Between-group comparisons using independent t-statistics showed that for non-repetitive movements the right premotor (BA 6, X: 39, Y: 𢄨, Z: 66) source of Mu was stronger in the control than in the Asperger group, whereas the right prefrontal (BA 8, X: 28, Y: 30, Z: 45) source was weaker in the control than in the Asperger group. Two additional regions with differential high-alpha activity were found in both groups: the right middle occipital–temporal gyrus (MOTG, BA 19/37, X: 58, Y: �, Z: 6) and the right inferior parietal lobe (IPL, BA 40, X: 49, Y: �, Z: 36) where non-repetitive stimulation elicited more activity than repetitive stimulation. There were no group differences in occipital sources of alpha ( Fig. 3C ).

Sources of significant differences (t-values) between the non-repetitive condition compared to baseline in control and Asperger's children and for control versus Asperger's children during the non-repetitive condition. (A) Increased 10� Hz activity is localized in right premotor areas (BA 6). Decreased 10� Hz activity is localized in occipital regions and right middle frontal gyrus (BA 8). (B) Decreased 10� Hz activity is localized in occipital regions and increased activity in prefrontal areas (BA 8). (C) Increased 10� Hz activity for the control relative to the Asperger's children is localized in right premotor areas (BA 6). Decreased 10� Hz activity for the control relative to the Asperger's children is localized in right middle/superior frontal gyrus (BA 8).

3.2.3. Correlation with social skill scores

The independently conducted, not regionally constrained, correlations between the sources of the 10� Hz MEG activity and the scores of social competence and social skill impairments across groups corroborated and extended the group comparisons: although the correlations included the entire source space, the social competence and social skills impairment scales correlated significantly only with sources overlapping or in close vicinity to the regions identified previously in the between-group comparisons, namely the right premotor, right middle frontal gyrus, right IPL, and the right MOTG. These correlations indicated that the more socially competent an individual was, the stronger was the source of high-Mu activity during observation of non-repetitive movements in right premotor regions (BA 6), right IPL (BA 40), and right MOTG (BA 19/37, all positive correlations Fig. 4 , Fig. 5 ) and the less high-Mu activity occurred in right prefrontal regions (BA 8, negative correlation Fig. 4 , right). The same or at least largely overlapping regions (see Table 2 ) were correlated with social skill impairments but, obviously, in the reverse direction: the more an individual tended to have social problems, the less high-Mu activity was generated in right premotor regions (negative correlation) while more high-Mu activity was generated in right prefrontal regions (positive correlation).

Left: Mu-activity in premotor cortex (BA 6, X: 42, Y: 𢄥, Z: 65) is strongly correlated with social competence scale. The scatterplot in the lower panel indicates an r = .74 with p < .01 for this region (average over voxels). Right: An inverse correlation was obtained for prefrontal cortex (BA 8, X: 30, Y: 39, Z: 39). The scatterplot in the lower panel indicates an r = −.66 with p < .01 for this region (averaged over voxels). Note the close similarity of the regions identified in the correlation analysis and those identified in the group comparison ( Fig. 2C ) and the almost perfect separation of the groups in the scatterplots.

Left: High-alpha activity in MOTG (BA 19/37, X: 52, Y: �, Z: 3) is strongly correlated with the social competence scale. The scatterplot in the lower panel indicates an r = .73 with p < .01 for this region (average over voxels). Right: Correlation of high-alpha activity in the IPL (BA 40, X: 65, Y: �, Z: 36) with social competence. The scatterplot in the lower panel indicates an r = −.70 with p < .01 for this region (averaged over voxels).

Table 2

Talairach coordinates for 10� Hz sources correlating with social competence and social problems and respective correlation coefficients as well as the correlations between these sources and measures of intellectual abilities (CPM/SPM). Middle occipital-temporal gyrus (MOTG), inferior parietal lobe (IPL).

Social competence Social problems
X, Y, ZrCPM/SPM rX, Y, ZrCPM/SPM r
Premotor right (BA6)42, 𢄥, 65.74 *** −.1641, �, 69−.66 ** −.16
Prefrontal right (BA8)30, 39, 39−.66 ** −.2628, 39, 46.69 ** −.32
IPL right (BA 40)65, �, 36.70 ** −.1746, �, 34−.68 ** −.003
MOTG right (BA 19/37)52, �, 3.73 *** −.0739, �, 6−.63 ** −.004

BA 6 and BA 8 correlated selectively with socially relevant scores (CBCL scales, in particular social competence and social problems) while measures of intellectual abilities did not show any relationship with sources of Mu-activity in these regions. Please see Table 2 for further details.

Permutation tests revealed that it was unlikely that only the group differences between the two groups solely accounted for the correlations. For example, for the relationship between social competence and the premotor area 92.24% of the correlation distribution, were below the here yielded correlation of 0.74. The mean of this distribution was 0.57 which reflects the strength of correlation that could be expected if only the group differences accounted for the relationship. As can be seen in Fig. 4 there is data of one Asperger's child influencing the dampening the correlations due to high scores in social competence. In order to investigate this influence further, the same calculations as above were conducted excluding this one person. It occurs then that 99.72% of the correlation distribution are below the one found here. This suggests that there is a relationship between social competence and premotor activity which is not solely accounted for by group differences.


Simulating human sleep spindle MEG and EEG from ion channel and circuit level dynamics

Background: Although they form a unitary phenomenon, the relationship between extracranial M/EEG and transmembrane ion flows is understood only as a general principle rather than as a well-articulated and quantified causal chain.

Method: We present an integrated multiscale model, consisting of a neural simulation of thalamus and cortex during stage N2 sleep and a biophysical model projecting cortical current densities to M/EEG fields. Sleep spindles were generated through the interactions of local and distant network connections and intrinsic currents within thalamocortical circuits. 32,652 cortical neurons were mapped onto the cortical surface reconstructed from subjects' MRI, interconnected based on geodesic distances, and scaled-up to current dipole densities based on laminar recordings in humans. MRIs were used to generate a quasi-static electromagnetic model enabling simulated cortical activity to be projected to the M/EEG sensors.

Results: The simulated M/EEG spindles were similar in amplitude and topography to empirical examples in the same subjects. Simulated spindles with more core-dominant activity were more MEG weighted.

Comparison with existing methods: Previous models lacked either spindle-generating thalamic neural dynamics or whole head biophysical modeling the framework presented here is the first to simultaneously capture these disparate scales.

Conclusions: This multiscale model provides a platform for the principled quantitative integration of existing information relevant to the generation of sleep spindles, and allows the implications of future findings to be explored. It provides a proof of principle for a methodological framework allowing large-scale integrative brain oscillations to be understood in terms of their underlying channels and synapses.

Keywords: Cortex EEG Forward model Human MEG Sleep Spindle Thalamus.


Mirror neurons and embodied simulation in the development of archetypes and self-agency

In this paper I explore the role of mirror neurons and motor intentionality in the development of self-agency. I suggest that this will also give us a firmer basis for an emergent view of archetypes, as key components in the development trajectory of self-agency, from its foundation in bodily action to its mature expression in mentalization and a conscious awareness of intentionality. I offer some ideas about the implications of these issues of self-agency for our clinical work with patients whose developmental trajectory of self-agency has been partially inhibited, so that their communications have a coercive effect. I discuss the possibility that this kind of clinical phenomenon may relate to Gallese and Lakoff's hypothesis that abstract thought and imagination are forms of simulated action, and that the same sensory-motor circuits that control action also control imagination, concept formation and understanding, but with a crucial development, that of an inhibition of the connections between secondary pre-motor cortical areas and the primary motor cortex. I shall speculate that in the earlier developmental stages of self-agency, the separation of secondary from primary motor areas is not complete, so that imagination and thought are not entirely independent of physical action.


MEG from simulated neurons - Psychology

Basic Principles of Magnetoencephalography

(click on any figure to enlarge)

Magnetoencephalography (MEG) is a non-invasive neurophysiological technique that measures the magnetic fields generated by neuronal activity of the brain (Figure 1). The spatial distributions of the magnetic fields are analyzed to localize the sources of the activity within the brain, and the locations of the sources are superimposed on anatomical images, such as MRI, to provide information about both the structure and function of the brain (Figure 2).

The principle features of MEG are:

MEG is a direct measure of brain function, unlike functional measures such as fMRI, PET and SPECT that are secondary measures of brain function reflecting brain metabolism.

MEG has a very high temporal resolution. Events with time scales on the order of milliseconds can be resolved, again differentiating MEG from fMRI, PET and SPECT, which have much longer time scales.

MEG has excellent spatial resolution sources can be localized with millimeter precision.

MEG is completely non-invasive. It does not require the injection of isotopes or exposure to X-rays or magnetic fields. Children or infants can be studied and repeated tests are possible.

MEG is complementary to other modalities, the information provided by each modality adds to the full picture.

Magnetic fields are found whenever there is a current flow, whether in a wire or a neuronal element, as illustrated in the upper left panel of Figure 3.

The magnetic field passes unaffected through brain tissue and the skull, so it can be recorded outside the head (upper middle). The magnetic field is extremely small, but can be detected by sophisticated sensors that are based on superconductivity (upper right).

By analyzing the spatial distributions of magnetic fields (lower left), it is possible, by using a model such as a single equivalent current dipole (lower middle), to estimate the intracranial localization of the generator source and superimpose it on an MRI (lower right).

The sequence of recording and analysis steps is shown, starting from 148 MEG time series waveforms containing an interictal spike (lower left panel of Figure 4). The segment of data containing the spike has been “clipped out” and shown on a display of the sensors around the head projected on a flat surface (upper left). The amplitude of the spike at its peak is shown as a topographic map (middle). The spike shows a phase reversal over the left temporal region. A calculation of the spike location using the single equivalent current dipole model reveals a localization of the spike source in the left temporal lobe (right).

The spatial and temporal properties of MEG are illustrated in the left panel of Figure 5. Only MEG has extremely high temporal and spatial resolution, as represented in the lower left section of the graph. Other functional modalities, except invasive EEG (iEEG), have either poor temporal or spatial resolution. Clearly iEEG has the distinct disadvantage of being invasive.

The main drawback of MEG is shown on the right panel of Figure 5. The MEG signals of interest are extremely small, several orders of magnitude smaller than other signals in a typical environment that can obscure the signal. Thus, specialized shielding is required to eliminate the magnetic interference found in a typical urban clinical environment.

MEG is a unique and effective diagnostic tool for evaluating brain function in a variety of surgical planning applications. The fundamental advantages of MEG are listed in Figure 6.

Figure 1. Basic principles of MEG.

Figure 2. MEG combines functional information from magnetic field recordings with structural information from MRI.

Figure 3. Electrical activity in neurons produces magnetic fields that can be recorded outside the skull and used to calculate the locations of the activity within the brain.

Figure 4. The sequence of steps to localize sources of neuronal activity from time-domain recordings to MRI overlay.

Figure 5. MEG has the advantages of very high temporal and spatial resolution, however it requires highly sensitive instrumentation and sophisticated methods for eliminating environmental magnetic interference.


Neural correlation of successful cognitive behaviour therapy for spider phobia: a magnetoencephalography study

Cognitive behavioural therapy (CBT) can be an effective treatment for spider phobia, but the underlying neural correlates of therapeutic change are yet to be specified. The present study used magnetoencephalography (MEG) to study responses within the first half second, to phobogenic stimuli in a group of individuals with spider phobia prior to treatment (n=12) and then in nine of them following successful CBT (where they could touch and manage live large common house spiders) at least 9 months later. We also compared responses to a group of age-matched healthy control participants (n=11). Participants viewed static photographs of real spiders, other fear-inducing images (e.g. snakes, sharks) and neutral stimuli (e.g. kittens). Beamforming methods were used to localise sources of significant power changes in response to stimuli. Prior to treatment, participants with spider phobia showed a significant maximum response in the right frontal pole when viewing images of real spiders specifically. No significant frontal response was observed for either control participants or participants with spider phobia post-treatment. In addition, participants' subjective ratings of spider stimuli significantly predicted peak responses in right frontal regions. The implications for understanding brain-based effects of cognitive therapies are discussed.

Keywords: Brain activation Fear Magnetoencephalography Neuroimaging Psychological therapy Simple phobia.


Simulating human sleep spindle MEG and EEG from ion channel and circuit level dynamics

Background: Although they form a unitary phenomenon, the relationship between extracranial M/EEG and transmembrane ion flows is understood only as a general principle rather than as a well-articulated and quantified causal chain.

Method: We present an integrated multiscale model, consisting of a neural simulation of thalamus and cortex during stage N2 sleep and a biophysical model projecting cortical current densities to M/EEG fields. Sleep spindles were generated through the interactions of local and distant network connections and intrinsic currents within thalamocortical circuits. 32,652 cortical neurons were mapped onto the cortical surface reconstructed from subjects' MRI, interconnected based on geodesic distances, and scaled-up to current dipole densities based on laminar recordings in humans. MRIs were used to generate a quasi-static electromagnetic model enabling simulated cortical activity to be projected to the M/EEG sensors.

Results: The simulated M/EEG spindles were similar in amplitude and topography to empirical examples in the same subjects. Simulated spindles with more core-dominant activity were more MEG weighted.

Comparison with existing methods: Previous models lacked either spindle-generating thalamic neural dynamics or whole head biophysical modeling the framework presented here is the first to simultaneously capture these disparate scales.

Conclusions: This multiscale model provides a platform for the principled quantitative integration of existing information relevant to the generation of sleep spindles, and allows the implications of future findings to be explored. It provides a proof of principle for a methodological framework allowing large-scale integrative brain oscillations to be understood in terms of their underlying channels and synapses.

Keywords: Cortex EEG Forward model Human MEG Sleep Spindle Thalamus.


MEG Scanners Are Mega Powerful

To revist this article, visit My Profile, then View saved stories.

To revist this article, visit My Profile, then View saved stories.

Magneto-encephalography, or MEG, scanners are proving to be one of the most powerful tools in the hands of scientists using the machines to observe important details about epilepsy, brain tumors, emotions, pain perception and more.

While the technology has existed for decades, improved computing power and hardware have recently increased interest in the scanners. There are an estimated 100 MEG scanners around the world -- at a potential cost of $2 million each -- and their numbers are growing.

Click here for photos of brain scanning machines.

"The MEG is just exploding," said Robert Knowlton, director of the University of Alabama at Birmingham's Magnetic Source Imaging Laboratory. "It's been around a long time, but it's only now gotten into enough researcher's hands with the technology (being) mature enough."

While other types of brain scans detail the geography of the brain or detect blood flow, the MEG scanners track the magnetic signals that neurons throw off as they communicate. "You can look at how the networks of the brain are talking to each other in real time," said Greg Simpson, director of the Dynamic Neuroimaging Laboratory at the University of California at San Francisco.

The machines are designed to "measure the magnetic field pattern around the entire head and deduce from those patterns where the current flows are occurring within the brain," said Eugene Hirschkoff, vice president of engineering at 4-D Neuroimaging, a MEG scanner manufacturer.

This allows scientists to study magnetic changes in the brain and figure out which areas are busy doing things each millisecond. By contrast, functional magnetic resonance imaging, or fMRI, technology measures the movement of blood within the brain. The scans reveal which brain areas are active and need oxygen from the blood, Simpson said.

But it takes a while for oxygen-filled blood to move in the brain. "If a brain area was active for a 10th of a second, the blood-flow response to that area would take a second or two to start," said Simpson.

That may not sound like a long time, but this is the lickety-split world of the brain. "If I had you read a sentence during an fMRI scan, weɽ see the visual cortex light up, the language cortex light up, and other things that would light up," Knowlton said. "But theyɽ all be lit up, and you wouldn't know which one was first (to become active), which one was most important."

"With the MEG, the sequence becomes more clear," Knowlton said.

One thing they can't do is analyze the physical parts of the brain, so MEGs become even more powerful when combined with other technologies.

When combined with fMRI scans, in particular, "you get the best of everything," Knowlton said.

In a recent experiment, scientists tried to figure out what you might call the Grey's Anatomy effect. When characters on the show get lonely, they act out -- drinking too much at the local bar, baking up a storm in the kitchen, having sex with a guy who ends up with a permanent erection.

Self-medication? Maybe. But a team of scientists prefers to describe the tendency toward bad behavior as cognitive disruption brought on by social exclusion -- a process they think they can see with the help of the MEG.

W. Keith Campbell, a psychology professor at the University of Georgia, and his colleagues recruited 30 female students who underwent what they thought were ordinary personality tests while being monitored by an MEG scanner.

Half the women were told that their test results suggested they would have trouble maintaining relationships and "end up alone in later life." The other women, a control group, got only neutral feedback to their answers, with no indications their lives were destined to go to pot.


Why Google DeepMind Is Putting AI on the Psychologist’s Couch

Artificial intelligence can now carry out many of the same cognitive tasks humans can, but we still don’t really understand how AIs think. Google DeepMind plans to train long-standing tests of human cognitive skills on machine minds to learn how they work.

A long-standing problem in AI research has been the fact that deep neural networks are “black boxes.” You can’t tell how these algorithms work just by looking at their code. They teach themselves by training on data and there’s no simple flow diagram a human can follow. The way these networks reach decisions is encoded in the weights of thousands of simulated neurons.

But they’re not the only inscrutable thinking machines. Simply poking around in the human brains they are modeled on yields few clues as to how people reason either, and so over the years, cognitive psychologists have developed tests designed to probe our mental faculties.

Now DeepMind has built a virtual 3D laboratory called Psychlab that will let machines take these tests too, and they’ve open-sourced it so any AI researcher can put their algorithms through their paces.

Psychlab is built on the company’s DeepMind Lab platform, designed for testing virtual agents in 3D environments. It recreates the set-up a human taking part in an experiment would see by providing the subject with a first-person view of a virtual computer monitor that displays a variety of classic cognitive tests.

These include measures of the ability to search for objects in a scene, detect change, remember a growing list of items, track moving objects, or recall stimulus-response pairings. Typically a human would use a mouse to respond to on-screen tasks, but virtual agents use the direction of their gaze.

By mimicking the environment a human would see, the researchers say, humans and AIs can effectively take the same tests. That should not only make it easier to draw direct comparisons between the two, but also allow results to be connected with the existing academic literature in cognitive psychology.

Being able to draw on the accumulated wisdom of 150 years of psychology research could be hugely useful in understanding how the latest AI systems work, the researchers say in a paper published on the arXiv to coincide with the release of the tool.

In recent years there’s been an increasing focus on deep-reinforcement learning AI systems that can carry out complicated tasks in simulated 3D environments. The complex nature of these environments and the variety of strategies these systems can employ to solve problems makes it hard to tease out what combination of cognitive abilities underlies their performance.

But by subjecting a state-of-the-art deep reinforcement learning agent called UNREAL to a variety of tests in Psychlab, they were able to uncover details about how its perceptual system worked, and even use this to improve its performance.

It turns out UNREAL has considerably worse acuity, or keenness of vision, than humans, which means it learns faster when presented with larger objects. Key to human acuity is a dense cluster of photoreceptors at the center of the retina called the fovea, which gives us particularly sharp vision at the center of our visual field.

By adding a simple model of the fovea to UNREAL, the researchers were able to improve the agent’s performance not just on the Psychlab experiments, but also on other standard DeepMind Lab tasks.

DeepMind aren’t the only ones applying cognitive tests to AI. University of Michigan researchers have been subjecting reinforcement learning agents to maze navigation tasks that have long been used to test memory and learning in rats.

The mazes were built in the 3D world of Minecraft, and the AI were set increasingly complex tasks and given different rewards to find out what cognitive skills were important for successfully navigating the experiment. They found that being able to retrieve memories based on the context in which they were stored was key for solving their tests.

As AI continues to improve and develop higher-order cognitive skills such as reasoning, emotional intelligence, and planning, more sophisticated psychological tests could become a crucial way for us to understand the ways their mental processes differ from ours, which they almost certainly will.

That could be an important tool to ensure everyone gets along in a future where humans and AI have to coexist.


Computational model reveals how the brain manages short-term memories

Credit: CC0 Public Domain

If you've ever forgotten something mere seconds after it was at the forefront of your mind—the name of a dish you were about to order at a restaurant, for instance—then you know how important working memory is. This type of short-term recall is how people retain information for a matter of seconds or minutes to solve a problem or carry out a task, like the next step in a series of instructions. But, although it's critical in our day-to-day lives, exactly how the brain manages working memory has been a mystery.

Now, Salk scientists have developed a new computational model showing how the brain maintains information short-term using specific types of neurons. Their findings, published in Nature Neuroscience on December 7, 2020, could help shed light on why working memory is impaired in a broad range of neuropsychiatric disorders, including schizophrenia, as well as in normal aging.

"Most research on working memory focuses on the excitatory neurons in the cortex, which are numerous and broadly connected, rather than the inhibitory neurons, which are locally connected and more diverse," says Terrence Sejnowski, head of Salk's Computational Neurobiology Laboratory and senior author of the new work. "However, a recurrent neural network model that we taught to perform a working memory task surprised us by using inhibitory neurons to make correct decisions after a delay."

In the new paper, Sejnowski and Robert Kim, a Salk and UC San Diego MD/Ph.D. student, developed a computer model of the prefrontal cortex, an area of the brain known to manage working memory. The researchers used learning algorithms to teach their model to carry out a test typically used to gauge working memory in primates—the animals must determine whether a pattern of colored squares on a screen matches one that was seen several seconds earlier.

Sejnowski and Kim analyzed how their model was able to perform this task with high accuracy, and then compared it to existing data on the patterns of brain activity seen in monkeys carrying out the task. In both tests, the real and simulated neurons involved in working memory operated on a slower timescale than other neurons.

Kim and Sejnowski found that good working memory required both that long-timescale neurons be prevalent, and that connections between inhibitory neurons—which suppress brain activity—be strong. When they altered the strength of connections between these inhibitory neurons in their model, the researchers could change how well the model performed on the working memory test as well as the timescale of the pertinent neurons.

The new observations point toward the importance of inhibitory neurons, and could inspire future research on the role of these cells in working memory, the researchers say. They also could inform studies on why some people with neuropsychiatric disorders, including schizophrenia and autism, struggle with working memory.

"Working memory impairment is common in neuropsychiatric disorders, including schizophrenia and autism spectrum disorders," says Kim. "If we can elucidate the mechanism of working memory, that's a step toward understanding how working memory deficits arise in these disorders."


3. Results

3.1. Questionnaires and psychological tests

Table 1 displays the mean values for the questionnaire and test data. The two groups did not differ regarding CPM/SPM and K-ABC scores. However, Asperger's children scored significantly lower on social competence scales and significantly higher on all social impairment scales of the CBCL. Due to experimenter error, the CBCL score was missing for one child in the Asperger's group.

3.2. MEG

3.2.1. Sensor data

The FFT spectra averaged over all sensors for the two groups separately are displayed in Fig. 2A for the repetitive and non-repetitive biological movement condition. In contrast to recent EEG studies, we found elevation of high-Mu power while observing biological movement. This power increase in the 10� Hz range was conspicuous and topographically distinct especially in the non-repetitive condition and particularly in the data recorded at frontal and central sensors in typically developed children, (see Fig. 2B ). In contrast, in both groups there was only a little change in high-Mu power in response to repetitive stimulation. Overall the group difference in power was larger for the non-repetitive than for the repetitive condition (see Fig. 2A ).

(A) Frequency distribution over all 148 MEG sensors for repetitive and non-repetitive movements separately for control (top right) and Asperger's children (bottom right). (B) Topographical illustration of the 10� Hz frequency range.

3.2.2. Source localization of group differences

Group comparisons of MEG source activity revealed that for the non-repetitive movements, the Asperger's syndrome and the control groups differed significantly in the pattern of high-Mu-modulation while no differences were found between the two groups in response to repetitive movements. Therefore, the source analysis focused on the non-repetitive movements condition. Within-group dependent t-statistics on source activity in the control group revealed enhancement in high-Mu power while observing biological movements relative to baseline in the right premotor cortex (Brodmann area, BA 6, Talairach coordinates: X: 39, Y: 𢄩, Z: 70) and a concurrent reduction in high-Mu power in right middle prefrontal cortex (BA 8, X: 32, Y: 28, Z: 38 see Fig. 3A ). In contrast, Asperger's children lacked the premotor enhancement and showed bilaterally increased activity in the prefrontal cortex (BA 8, right: X: 24, Y: 45, Z: 48, left: X: 𢄩, Y: 50, Z: 48 see Fig. 3B ). In addition, in both groups, sources in the occipital cortex showed decreased activity in the high-alpha power. Between-group comparisons using independent t-statistics showed that for non-repetitive movements the right premotor (BA 6, X: 39, Y: 𢄨, Z: 66) source of Mu was stronger in the control than in the Asperger group, whereas the right prefrontal (BA 8, X: 28, Y: 30, Z: 45) source was weaker in the control than in the Asperger group. Two additional regions with differential high-alpha activity were found in both groups: the right middle occipital–temporal gyrus (MOTG, BA 19/37, X: 58, Y: �, Z: 6) and the right inferior parietal lobe (IPL, BA 40, X: 49, Y: �, Z: 36) where non-repetitive stimulation elicited more activity than repetitive stimulation. There were no group differences in occipital sources of alpha ( Fig. 3C ).

Sources of significant differences (t-values) between the non-repetitive condition compared to baseline in control and Asperger's children and for control versus Asperger's children during the non-repetitive condition. (A) Increased 10� Hz activity is localized in right premotor areas (BA 6). Decreased 10� Hz activity is localized in occipital regions and right middle frontal gyrus (BA 8). (B) Decreased 10� Hz activity is localized in occipital regions and increased activity in prefrontal areas (BA 8). (C) Increased 10� Hz activity for the control relative to the Asperger's children is localized in right premotor areas (BA 6). Decreased 10� Hz activity for the control relative to the Asperger's children is localized in right middle/superior frontal gyrus (BA 8).

3.2.3. Correlation with social skill scores

The independently conducted, not regionally constrained, correlations between the sources of the 10� Hz MEG activity and the scores of social competence and social skill impairments across groups corroborated and extended the group comparisons: although the correlations included the entire source space, the social competence and social skills impairment scales correlated significantly only with sources overlapping or in close vicinity to the regions identified previously in the between-group comparisons, namely the right premotor, right middle frontal gyrus, right IPL, and the right MOTG. These correlations indicated that the more socially competent an individual was, the stronger was the source of high-Mu activity during observation of non-repetitive movements in right premotor regions (BA 6), right IPL (BA 40), and right MOTG (BA 19/37, all positive correlations Fig. 4 , Fig. 5 ) and the less high-Mu activity occurred in right prefrontal regions (BA 8, negative correlation Fig. 4 , right). The same or at least largely overlapping regions (see Table 2 ) were correlated with social skill impairments but, obviously, in the reverse direction: the more an individual tended to have social problems, the less high-Mu activity was generated in right premotor regions (negative correlation) while more high-Mu activity was generated in right prefrontal regions (positive correlation).

Left: Mu-activity in premotor cortex (BA 6, X: 42, Y: 𢄥, Z: 65) is strongly correlated with social competence scale. The scatterplot in the lower panel indicates an r = .74 with p < .01 for this region (average over voxels). Right: An inverse correlation was obtained for prefrontal cortex (BA 8, X: 30, Y: 39, Z: 39). The scatterplot in the lower panel indicates an r = −.66 with p < .01 for this region (averaged over voxels). Note the close similarity of the regions identified in the correlation analysis and those identified in the group comparison ( Fig. 2C ) and the almost perfect separation of the groups in the scatterplots.

Left: High-alpha activity in MOTG (BA 19/37, X: 52, Y: �, Z: 3) is strongly correlated with the social competence scale. The scatterplot in the lower panel indicates an r = .73 with p < .01 for this region (average over voxels). Right: Correlation of high-alpha activity in the IPL (BA 40, X: 65, Y: �, Z: 36) with social competence. The scatterplot in the lower panel indicates an r = −.70 with p < .01 for this region (averaged over voxels).

Table 2

Talairach coordinates for 10� Hz sources correlating with social competence and social problems and respective correlation coefficients as well as the correlations between these sources and measures of intellectual abilities (CPM/SPM). Middle occipital-temporal gyrus (MOTG), inferior parietal lobe (IPL).

Social competence Social problems
X, Y, ZrCPM/SPM rX, Y, ZrCPM/SPM r
Premotor right (BA6)42, 𢄥, 65.74 *** −.1641, �, 69−.66 ** −.16
Prefrontal right (BA8)30, 39, 39−.66 ** −.2628, 39, 46.69 ** −.32
IPL right (BA 40)65, �, 36.70 ** −.1746, �, 34−.68 ** −.003
MOTG right (BA 19/37)52, �, 3.73 *** −.0739, �, 6−.63 ** −.004

BA 6 and BA 8 correlated selectively with socially relevant scores (CBCL scales, in particular social competence and social problems) while measures of intellectual abilities did not show any relationship with sources of Mu-activity in these regions. Please see Table 2 for further details.

Permutation tests revealed that it was unlikely that only the group differences between the two groups solely accounted for the correlations. For example, for the relationship between social competence and the premotor area 92.24% of the correlation distribution, were below the here yielded correlation of 0.74. The mean of this distribution was 0.57 which reflects the strength of correlation that could be expected if only the group differences accounted for the relationship. As can be seen in Fig. 4 there is data of one Asperger's child influencing the dampening the correlations due to high scores in social competence. In order to investigate this influence further, the same calculations as above were conducted excluding this one person. It occurs then that 99.72% of the correlation distribution are below the one found here. This suggests that there is a relationship between social competence and premotor activity which is not solely accounted for by group differences.


Simulating human sleep spindle MEG and EEG from ion channel and circuit level dynamics

Background: Although they form a unitary phenomenon, the relationship between extracranial M/EEG and transmembrane ion flows is understood only as a general principle rather than as a well-articulated and quantified causal chain.

Method: We present an integrated multiscale model, consisting of a neural simulation of thalamus and cortex during stage N2 sleep and a biophysical model projecting cortical current densities to M/EEG fields. Sleep spindles were generated through the interactions of local and distant network connections and intrinsic currents within thalamocortical circuits. 32,652 cortical neurons were mapped onto the cortical surface reconstructed from subjects' MRI, interconnected based on geodesic distances, and scaled-up to current dipole densities based on laminar recordings in humans. MRIs were used to generate a quasi-static electromagnetic model enabling simulated cortical activity to be projected to the M/EEG sensors.

Results: The simulated M/EEG spindles were similar in amplitude and topography to empirical examples in the same subjects. Simulated spindles with more core-dominant activity were more MEG weighted.

Comparison with existing methods: Previous models lacked either spindle-generating thalamic neural dynamics or whole head biophysical modeling the framework presented here is the first to simultaneously capture these disparate scales.

Conclusions: This multiscale model provides a platform for the principled quantitative integration of existing information relevant to the generation of sleep spindles, and allows the implications of future findings to be explored. It provides a proof of principle for a methodological framework allowing large-scale integrative brain oscillations to be understood in terms of their underlying channels and synapses.

Keywords: Cortex EEG Forward model Human MEG Sleep Spindle Thalamus.


Watch the video: ΝΩΤΙΑΙΟΣ ΜΥΕΛΟΣ (August 2022).