Fig. 5: Stronger segregation during natural stimuli.

A Signal correlation for within-module and otherwise connections. Note that some neurons may not belong to any module, otherwise connections can involve multiple scenarios. *p < 0.05, **p < 0.01, ***p < 0.001, rank-sum test, one-sided, n = 7 mice. Error bars represent 95% confidence interval. Connected neurons partitioned into the same module tend to have higher signal correlations than the rest connected neurons, demonstrating our module partition provides insight into not only the connectivity pattern but also functional similarity to some extent. B Number of modules with normalized size, coverage or purity higher than the threshold. Normalized size is the size of module normalized by the total number of neurons in the network, insets show the plots on a log-log scale. C Number of modules with purity higher than the threshold for each visual area during all visual stimuli. D, E Properties of the modular structure during natural scene and natural movie presentations. We examined four different aspects of the case where neurons from a single visual area are divided into multiple modules (in which they are the dominant area), with signal correlation indicating the functional similarity along with connection probability, positive and negative fractions demonstrating the validity of our module partitioning algorithm. ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, rank-sum test, one-sided, n = 7 mice. Error bars represent 95% confidence interval. F Distance in degree between centers and overall correlation of receptive fields for connected and disconnected neuron. ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, rank-sum test, one-sided, n = 7 mice. Error bars represent 95% confidence interval.