Professor William Bialek
Princeton University and The CUNY Graduate Center, USA

Title: "Statistical mechanics for networks of real neurons"


Percepts, thoughts, memories, and actions depend on the coordinated activity of thousands of neurons in the brain.  How do we describe this coordination, and the collective behaviors that can emerge from neural networks?  In the inanimate world, emergent phenomena are described in the language of statistical physics, and it is an old dream that this would also provide a productive approach to thinking about the brain.   Now that experiments can monitor, simultaneously, the electrical activity of thousands of neurons, we should demand that these theoretical ideas make closer contact with experiment.   I’ll describe efforts that my colleagues and I have made in this direction, using the mouse hippocampus as an example.  We will see that very simple statistical physics models provide a strikingly accurate description of ~100 neurons, while ideas from the renormalization group hint at an underlying simplicity on the scale of ~1000 neurons.  I hope to make clear why these somewhat abstract results are exciting for a theoretical physicist, and what they might tell us about biological function.
This is joint work with L Meshulam, JL Gauthier, CD Brody, and DW Tank.


About Neuroscience Seminars

Neuroscience seminars at the QBI play a major role in the advancement of neuroscience in the Asia-Pacific region. The primary goal of these seminars is to promote excellence in neuroscience through the exchange of ideas, establishing new collaborations and augmenting partnerships already in place.

While seminars in the QBI Auditorium have been suspended due to COVID-19, we will still be holding seminars via Zoom. The days and times of these seminars will vary depending on the time zone of the speaker. Please see each seminar listed below for details. 


Neuroscience Seminars archive 2005-2018