MEASUREMENTS OF INFORMATION RATES IN MONKEY MT NEURONS IN RESPONSE TO TIME-VARYING STIMULI.

G. Buracas, A. Zador, M. DeWeese, and T. Albright

Society for neuroscience, 1996

Abstract

The traditional notion of a tuning curve relates the time-average firing rate of a neuron to the constant characteristics of the stimulus in its receptive field; this firing rate is typically measured in windows of several hundreds of milliseconds. How does the appropriate time window change when the stimulus changes with time? Inspired by work on a motion-sensitive neuron in the fly (Bialek et al, Science, 1991 252:1854-7), we presented visual stimuli with rich temporal structure to alert fixating macaque monkeys while recording extracellularly from neurons in area MT. The stimulus consisted of a Gabor patch whose direction of motion performed a random walk. For one stimulus ensemble, the spike train exhibited modulation on a 10 millisecond time scale (cf. Bair and Koch, 1996); under these conditions, the spike train encoded nearly all of the 30 bits/second available in the stimulus. This information could be extracted only when the spikes are detected with a precision of order 10 millisecond. By contrast, a traditional analysis based on counting the number of spikes in much longer epochs would have underestimated the information rate by an order of magnitude. These results suggest that the notion of a tuning curve can be extended to the analysis of time-varying stimuli, provided that the window size is appropriately small. These experiments may have implications for understanding cortical processing of natural time-varying stimuli. Supported by EY07605 and Salk-Sloan center for Theoretical Neurobiology.

Go back to Tony's home page