RT Journal Article
SR Electronic
T1 Heterogeneous structure of the population code in V4 shapes pair-wise interactions on different time scales
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 645135
DO 10.1101/645135
A1 Veronika Koren
A1 Ariana R. Andrei
A1 Ming Hu
A1 Valentin Dragoi
A1 Klaus Obermayer
YR 2019
UL http://biorxiv.org/content/early/2019/05/29/645135.abstract
AB In visual areas of primates, neurons activate in parallel while the animal is engaged in a behavioral task. In this study, we examine the structure of the population code while the animal performs delayed match to sample task on complex natural images. The macaque monkeys visualized two consecutive stimuli that were either the same or different, while recorded with laminar arrays across the cortical depth in V1 and V4 cortical areas. We decoded correct choice behavior from the activity of single neurons as well as from neural populations of simultaneously recorded units. Comparing the predictive power of the activity of single neurons with the high-dimensional model of the population activity, we find that the high-dimensional read-out predicts correct choices better than an average single neuron from the same recording session. Utilizing decoding weights, we divide neurons in informative and uninformative, and show that informative neurons in V4, but not in V1, are more strongly synchronized, coupled and correlated than uninformative neurons. As neurons are divided in two coding pools according to their coding preference for matching and non-matching stimuli, in V4, but not in V1, spiking synchrony, coupling and correlations within the coding pool are stronger than across coding pools. In summary, our analysis points out that the heterogeneous structure of responses is important for decoding and that this structure shapes pairwise interactions on different time scales.HighlightsResponses of a population of neurons in V1 and V4 to complex naturalistic images contain more information about correct choice than an average single neuron.Informative neurons are more strongly coupled, correlated and synchronized than uninformative neurons in V4, but not in V1.Neurons are more strongly coupled, correlated and synchronized within coding pools compared to across coding pools in V4, but not in V1.The structure of population responses contributes substantially to the performance of the decoder. In V1, the structure across all neurons matters, while in V4, only the structure within the coding pool is important.