Cortical Dynamics

The cerebral cortex is a thin sheet of neurons forming a shell around the rest of the brain, and it is this particular structure that has most disproportionately enlarged in primates and especially in humans. Accordingly, an understanding of the function of this structure may give us great insight into the neural basis of the unique cognitive abilities we possess. Complicating this question are the complex patterns of activity commonly observed in the cortex: many neurons are active in ways that seem unrelated to other neurons, and unrelated to behavior. In fact, a majority of neurons in cortex continue their activity when a subject sits quietly, and even when the subject sleeps. From this, we can infer that it is not the mere presence of activity in cortex that guides behavior, but the particular pattern and dynamics (the way that patterns change over time) of neuronal activity that matters.

What patterns of activity are observed in the cortex, and how are these patterns related to perception and behavior?

How are they controlled by cognitive processes like attention?

Working with Tatiana Engel (now Assistant Professor at Cold Spring Harbor Labs), we asked specifically: How does visual attention control and route perceptual information through the cortex? We recorded the activity of neurons in the visual cortex of non-human primates while they performed an visual attention task. By fitting models to describe the activity patterns we recorded, we observed that attention creates a specific change in the dynamics of the activity of local populations of cortical neurons (Engel*, Steinmetz*, et al., Science, 2016). Specifically, during attention activity is desynchronized, such that neurons spend longer in a high-activity "ON" state, along with other changes. These changes mirror those seen when subjects become more aroused and alert, but for the first time we demonstrated that such desychronization can occur locally rather than globally, and can be controlled by a cognitive process such as attention. This finding suggests that this kind of change in the dynamics of neuronal population activity may be a part of the general mechanisms by which information is routed through the brain.

Top, spiking rasters recorded from area V4, showing results of a model fit to describe the ON/OFF dynamics. Bottom, we found that both firing rate and duration of ON phases increased.

In another study, led by Michael Okun (now Research Fellow at University of Leicester), we asked: how does the activity of individual neurons relate to the dynamics of the whole population? How is this population coupling related to functional properties of neurons, like their modulation during attention? By recording populations of neurons extracellularly in multiple species, we observed a striking and consistent diversity of population coupling: some neurons closely followed the population dynamics, whereas others were independent (Okun et al., Nature 2015). This property of neurons was stable across conditions, and, importantly, was linked to functional properties, including visual receptive field structure and modulation during attention. These results provide a new description of the patterns and dynamics of activity in the cortex, and clarify the way that these patterns relate to perception and cognition.

In ongoing work, we are extending our studies of cortical dynamics to larger scales to ask:

How are cortical dynamics coordinated across different cortical areas?

How are cortical dynamics related to activity in other structures, such as thalamus, hippocampus, striatum, and the colliculus?

We are addressing the first of these questions with widefield calcium imaging, and the second with large-scale electrophysiology using Neuropixels probes.