Modeling, Computation, Nonlinearity, Randomness and Waves Seminar

Populations of neurons in the brain can produce distinctive, rhythmic activity observable in extracellular recordings.  Activity in the ~30-90 Hz frequency band of these recordings has been the subject of much attention and is known as gamma rhythm.  It has been found in the visual cortex, for example, where power in the gamma band is modulated by the orientation and contrast of visual stimuli.  Whether it performs a specific function in the brain or is merely a side-effect of neural computation is not yet settled.  My aim will be to discuss the mechanism that produces gamma rhythm.

As gamma rhythm is thought to be locally produced, I will begin by introducing a spiking network model of a generic local population of neurons in the brain, where biophysical parameters and network structure are chosen to reflect data where possible.  When the network is driven, an emergent spike pattern appears consisting of the briefly coordinated spiking activity of small groups of neurons followed by relative lulls in activity.  The frequency and size of these events are consistent with gamma rhythm.  I will explain the mechanism of these events, their statistics, and how synaptic parameters influence their size and inter-event times.

Next I will show how such rhythms appear in a semi-realistic model of the visual cortex.  The model covers a small patch of monkey V1 and reproduces many of its known phenomena.  Driven now by visual stimuli, a gamma rhythm is again produced in a way that is graded with stimulus features like contrast and orientation.  The more detailed and realistic model benchmarked against different forms of experimental data increases our confidence in the mechanism for gamma rhythm production, and allows us to relate dynamics to the network’s visual functions and response statistics.  This is joint work with Lai-Sang Young and Robert Shapley.