Fig. 5: PMd is dominated by task-dependent activity and S1 by task-independent activity.
From: Regional specialization of movement encoding across the primate sensorimotor cortex
a We used dPCA to decompose neural population activity of one session into demixed neural modes (dNM), which were then decomposed into task-dependent and task-independent parts. We refer to demixed neural modes as task-dependent or task-independent according to whether the variance is mostly accounted for (>50%) by task-dependent or task-independent parts. b Plots show M1 neural population peri-gait activity of three leading task-independent and task-dependent modes for each task in Mk-Ek session S7. All task-independent or task-dependent modes spanned the task-independent or task-dependent subspace respectively. c Neural population activity projected into a space spanned by the leading task-independent neural modes. d Bars show the portion of the variance explained by each neural mode, and the portion of explained variance belonging to its task-dependent (light green) or task-independent (dark green) parts for the same session as in (b). In this representative example in Mk-Ek, variance in the leading PMd demixed neural modes is largely explained by their task-dependent parts. In contrast, variance is largely explained by the task-independent parts in M1 and S1. e Bar plots show the mean portion of variance explained by the task-independent subspace across all sessions and both monkeys (n = 14; S1: 68% ± 3%; M1: 56% ± 3%; PMd: 20% ± 3%; PMd vs. M1: p = 0.0005; PMd vs. S1: p = 0.0005; M1 vs. S1: p = 0.0005). Dots show the value for individual sessions in both monkeys (left column: Mk-EK; right column: Mk-Nt.). Black line shows the chance level with grey tube showing the values not significantly different from chance at p ≥ 0.05 (chance level: PMd: 4.122 ± 0.22; M1: 5.31 ± 0.40; S1: 7.00 ± 0.58; measured value vs. chance: PMd: p = 0.0005; M1: p = 0.0005; S1: p = 0.0005). Blue line shows the estimated noise contribution with light blue tube shows the values not significantly different from noise contribution at p ≥ 0.05 (noise contribution: PMd: 91.73 ± 0.45; M1: 92.55 ± 0.45; S1:91.92 ± 0.45; measured value vs. noise contribution: PMd: p = 0.0005; M1: p = 0.0005; S1: p = 0.0005). Pie charts show the portions of variance explained by task-independent and task-dependent parts of all the demixed neural modes, respectively. Bars: mean ± s.e.m. *** p < 0.001, one-sided Monte Carlo permutation test. Source data are provided as a Source Data file.
