[Probcogsci] Fwd: Toby Berger Seminar - Friday Feb 5

[Angela Yu]

Sounds like a very interesting talk, in case you haven't already  
gotten the announcement...

Angela

---------------------------------------------------------
Angela J. Yu
Assistant Professor
Department of Cognitive Science
UCSD, Mail Code 0515
9500 Gilman Drive
La Jolla, CA 92093-0515

Email: ajyu at cogsci.ucsd.edu
Phone: 858-822-3317
Fax: 858-534-1128
Website: www.cogsci.ucsd.edu/~ajyu
---------------------------------------------------------

Begin forwarded message:

> From: Terry Sejnowski <terry at salk.edu>
> Date: January 25, 2010 1:39:25 PM PST
> To: "all_science at salk.edu" <all_science at salk.edu>
> Cc: "everyone at cogsci.UCSD.EDU" <everyone at cogsci.UCSD.EDU>, "ai at cs.ucsd.edu 
> " <ai at cs.ucsd.edu>, "psych-talks at psy.ucsd.edu" <psych-talks at psy.ucsd.edu 
> >, "inc at ucsd.edu" <inc at ucsd.edu>, "neuro-faculty at ucsd.edu" <neuro-faculty at ucsd.edu 
> >
> Subject: Toby Berger Seminar - Friday Feb 5
>
> Crick-Jacobs Center for Theoretical and Computational Biology Seminar
>
> Toby Berger
>
> Dept. of Electrical & Computer Engineering
> University of Virgina
>
> "A Mathematical Theory of Energy Efficient Sensory Neurons"
>
> Friday, February 5, 11 am
>
> Salk Institute
> Trustees Room, East Building
>
> Host:  Terry Sejnowski
>
> -----
>
> Abstract:  An idealized neuroscience-based mathematical model of how a
> neuron stochastically processes data and communicates information is
> introduced and analyzed.  The model neuron continually transmits
> information about the intensity of excitation it is experiencing from
> neural spike trains delivered to its synapses by thousands of other
> neurons.  It encodes this excitation history into a sequence of time
> instants at which it generates neural spikes of its own.  By  
> propagating
> these spikes along its axon to thousands of other neurons, the neuron
> functions as what communication and information theorits would call a
> multiaccess, partially degraded broadcast channel that employs a
> time-continuous version of pulse position modulation.
>
> The mathematical model features three parameters, denoted by m,  
> kappa, and
> b, that largely characterize the neuron as an engine of computation  
> and
> communication.  Each set of values of these parameters corresponds  
> to a
> long term maximization of the bits the neuron conveys to its targets  
> per
> joule it expends doing so.  It is shown that this bpj ratio is  
> maximized
> by distributing the random duration between successive spikes the  
> neuron
> generates aaccording to a gamma pdf with parameters kappa and b, and  
> by
> distributing b/A according to a beta probability density with  
> parameters
> kappa and m-kappa, where A is the random intensity of the effectively
> Poisson process of spikes that arrive to the union of the neuron's
> synapses. Possible ways to analyze generalized versions of the model  
> that
> better correspond to real neurons in sensory cortex also will be
> discussed.
>
> -----
>