Pairwise correlations of spiking activity changes along the ventral visual cortical pathway of macaque monkeys

<jats:title>Abstract</jats:title><jats:p>Spiking activity often correlates across neurons in the cerebral cortex. Precise spike-timing correlation at millisecond scale reflects underlying neural circuitries. Correlations of stimulus evoked spiking activities affects the distributions of spike counts in a multi-neuron space and reflects stimulus encoding ability of neuron populations. The present study compared the degree of spiking-activity correlation among three areas of the macaque visual system to determine whether neural circuitry and the way to encode stimulus-related information by neuron populations are unique to each area. Spiking activity of multiple single neurons was recorded in macaque primary visual cortex (V1), visual association cortex (V4), and inferior temporal cortex (IT). Cross-correlation of spike trains revealed that single-neuron pairs in IT exhibited the highest incidence of precisely correlated firing, whereas V1 pairs showed the lowest incidence. Although noise correlation, which quantifies similarity in trial-to-trial response fluctuations, differed among cortical areas, signal correlation, which quantifies similarity in stimulus preferences, did not. The degree of temporally precise correlation was positively related to the signal-correlation strength in all three cortical areas and the relation of IT pairs was stronger than that of V1 pairs. Temporally precise correlations between single-neuron and population-level activity also differed among cortical areas and V1 neurons exhibited the lowest incidence of correlated firing with population activities. The differences in spiking-activity correlations among visual cortical areas suggest that each cortical area has unique neuronal circuitry for representing information, and which performs unique neuronal computations.</jats:p><jats:sec><jats:title>Significance Statement</jats:title><jats:p>Although hierarchical information processing plays a crucial role in the brain and in artificial neural networks, whether hierarchically connected brain areas share neural circuitry and information representation remains unclear. Here, we examined spiking activity across the ventral visual system in macaque monkeys and found that cortical areas differed in how correlated their spiking activity was, suggesting that neural circuitry and the way to encode stimulus-related information by neuron populations are unique to each cortical area.</jats:p></jats:sec>