*** Modeling Nonlinear Operations in Dendrites of Neocortical Neurons

How does a neuron respond to a complex spatiotemporal pattern of synaptic input? I have been using biophysically detailed models to address this question; modeling this problem in detail really requires two classes of model, working together:

1) a sufficiently detailed model of the single neuron we wish to study; building this model requires a thorough review of the literature to build a "best estimate" model of the kinetics and distribution of cellular structures such as ion channels, pumps and transporters, and calcium buffers.
2) a simplified model of the other neurons that provide input to the cell that we wish to model in detail; if studying a neuron in the visual system, for example, we want to present to the model a series of images -- a "movie" -- that results in a spatiotemporal pattern of synaptic input to our cell generally similar to that seen by a cell in the intact animal.

Work on both of these issues continues; software I have written to solve problem (2) is publicly available as a package for use with the NEURON simulation environment.

	Schematic of model providing synaptic input to a single neuron. A linear-nonlinear cascade model computes the time-varying responses of afferent neurons to a movie stimulus; a "synapse map" then specifies which afferents project what types of synapses to what locations on the biophysically detailed model.
	Mean synaptic event rates and raster plots of event trains generated by model in response to a movie containing a 250 msec presentation of a thin bar stimulus.

In earlier work, we examined two computations in visual "complex" cells that might be performed in dendritic subunits: translation- and contrast-polarity-invariant tuning to stimulus orientation (Mel et al. 1998) and binocular disparity (Archie and Mel 2000). Current work focuses on the the effects of attention on responses of neurons to multiple simultaneous stimuli in V2 and V4; an early summar is available as Archie and Mel (2001).