Theory and Experiment in the Brain: Beyond Mean-Field Theory in the Neurosciences
Leading experts in theoretical neuroscience meet at The Bernstein Centre for Computational Neuroscience Göttingen from June, 3- 5, 2015
A theoretical understanding of neural activity in the brain requires powerful mathematical techniques capable of handling a, nonlinear and noisy behavior. Throughout the past decades, many approaches originally developed for theoretical physics were adapted for problems in the neurosciences.
Perhaps one of the most successful techniques to be refitted is the so-called mean-field theory for population dynamics. This approach was pioneered by Grossberg (1967) and Wilson, Cowan (1972) and others. Their key results were instrumental to important findings and predictions in computational neuroscience. Later on, the theory was extended with the inclusion of second-order statistics by van Vreeswijik and Sompolinski (1996) describing the observed irregular activity of excitatory and inhibitory neurons in cortex, so-called the balanced state.
Although adapted mean‐field approaches are successful in providing a good base for many experimental and theoretical observations, their limited “averaging” scope fails to capture many important features of populations dynamics. For instance, phenomena that include spike dependent learning rules or network dynamics driven by external stimulations remain elusive. Moreover, it is unclear what mathematical techniques are needed to address these shortcomings.
This workshop will focus on the inherent difficulties of neural population dynamics problems and newly arising research topics. The meeting aims to serve as a forum for key researchers working with mean-field approaches and/or their emerging alternatives to exchange their tools and views.