GATSBY COMPUTATIONAL NEUROSCIENCE UNIT
UCL Logo

Xiao-Jing Wang
Center for Complex Systems, Brandeis University, USA

 

 

Wednesday 12 October 2005

16:00

 

Seminar Room B10 (Basement)

Alexandra House, 17 Queen Square, London, WC1N 3AR

 

Microcircuit Reverberation Underlying Working Memory, Decision Making and Selective Attention

 

It is a major challenge in neuroscience today to elucidate the neural mechanisms of cognitive functions at the cellular and microcircuit levels.  We have used biophysically-based network modeling approach to investigate a local cortical network, such as in prefrontal cortex, that is capable of mnemonic persistent activity during working memory.  In this talk, I will first summarize the main results, insights, and predictions from this work.  Then, I show that the same model can also be used to understand decision making computations, such as accumulation of sensory evidence and categorical choice.  The model reproduces quantitatively both neurophysiological data and behavioral psychophysics in a visual motion discrimination experiment.  Finally, I will discuss an extended model in which a working memory network is in reciprocal connection with a sensory network (MT).  We found that this model can account for interesting experimental observations on selective visual attention.  Surprisingly, our model automatically operates according to the `feature-similarity gain principle' of selecitve attention, initially proposed by psychologists and monkey physiologits.  These findings suggest that a single `cognitive-type' cortical network can underlie both internal representation and cognitive-type processing.