Characterization of neural responses with stochastic stimuli

Eero P. Simoncelli, Liam Paninski, Jonathan W. Pillow, Odelia Schwartz
In The Cognitive Neurosciences, III, 327-338. MIT Press (2004)

A fundamental goal of sensory systems neuroscience is the characterization of the functional relationship between environmental stimuli and neural response. The purpose of such a characterization is to elucidate the computation being performed by the system - a precise description of the mechanisms underlying the response is of secondary importance. Qualitatively, this notion is exemplified by the concept of the ``receptive field'', a quasi-linear description of a neuron's response properties that has dominated sensory neuroscience for the past 50 years. Receptive field properties are typically determined by measuring responses to a highly restricted set of stimuli, parameterized by one or a few parameters. These stimuli are typically chosen both because they are known to produce strong responses, and because they are easy to generate using available technology. While such experiments are responsible for much of what we know about the tuning properties of sensory neurons, they typically do not provide a complete characterization of neural response. In particular, the fact that a cell is tuned for a particular parameter, or selective for a particular input feature, does not necessarily tell us how it will respond to an arbitrary stimulus. Furthermore, we have no systematic method of knowing which particular stimulus parameters are likely to govern the response of a given cell, and thus it is difficult to design an experiment to probe neurons whose response properties are not at least partially known in advance. This chapter provides an overview of some recently developed characterization methods. In general, the ingredients of the problem are: (a) the selection of a set of experimental stimuli; (b) selection of a model of response; (c) a procedure for fitting (estimation) of the model. We discuss solutions of this problem that combine stochastic stimuli with models based on an initial linear filtering stage that serves to reduce the dimensionality of the stimulus space. We begin by describing classical reverse correlation in this context, and then discuss several recent generalizations that increase the power and flexibility of this basic method.

online publications