Inactivation of area MT has separate influences on the spiking of single neurons and neuron populations in primate V1

Abstract

For decades it has been debated whether visual information is carried by the spike rate of independently-responding neurons or by patterns in the spike trains of neuron populations. Several studies have suggested that both modes of neural coding may be complementary in processing visual information. The relative roles of feed-forward, horizontal and feedback pathways in the generation of these potential neural codes, however, remain unclear. In this study, we examined the influence of feedback on the spiking properties of single neurons and pairs of neurons in V1 of the anesthetized, paralyzed bush baby by inactivating the middle temporal (MT) motion-sensitive area with Muscimol. MT sends feedback projections to V1 that have been proposed to modulate V1 function. Area MT initially was identified by optical imaging. Subsequently a 100-electrode array was inserted into V1 and the orientation, temporal and spatial frequency and direction selectivity of all well-isolated neurons were examined. Next MT was blocked and responses re-examined. No statistically significant differences in any of these single neuron tuning properties were found following MT inactivation. Spike-time and spike-count correlations between pairs of neurons were also examined before and after MT inactivation. After inactivation the probability of detecting significant spike-time correlation peaks (21.8% vs. 39.7%; N = 2278 pairs) and their amplitudes (0.0169±0.0015 vs. 0.0122±0.0018; ±SEM; p[[lt]]0.05) was reduced although spike count correlations were not significantly reduced (0.149±0.012 vs. 0.173±0.012; ±SEM; p[[lt]]0.10). These data suggest that inactivation of higher area feedback has a greater influence on spike timing between neurons than on the tuning properties of individual neurons.