Direction tuning of macaque MT neurons: a reverse correlation study

Abstract

We compared direction tuning curves of motion sensitive neurons in area MT of the macaque monkey by using two different methods. First we presented a large patch of a random pixel array with motion impulses (8–13 ms) pseudo-randomly moving in 8 different directions. Spike trains of the neurons were recorded and reverse-correlated to the stimulus sequence. The temporal response profiles were biphasic with an excitatory response (in the range of 40–70 ms) and a delayed inhibitory response (40–60 ms relative to the excitatory response). For some neurons the response to different directions had different response latencies. We compared directional tuning of these reverse correlation experiments with conventional long duration motion trials, by repeatedly presenting one large patch of a random pixel array moving pseudo-randomly in 8 different directions for one second (10–15 repetitions). For both the reverse correlation and conventional method we fitted the direction tuning curves by gaussian functions, taking the response probability (reverse correlation) or the average response frequency (conventional method) for each direction. The results show that direction tuning curves are sharper in case of repeated long stimulation (conventional method). The sharpening of direction tuning in case of longer stimulus presentation compared to motion impulses might be due to temporal summation of motion impulses.