Extended Data Fig. 5: Correlating pre- to post-amputation multivoxel finger activity patterns.

(a) Visualization depicting the inter-session Pearson correlations of individual fingers within the BA3b hand region. (b) Inter-session correlations for the left (top row) and right hands (bottom) in the contralateral hand ROI. Line colors indicate session pairings (indicated in the legend). For case-study participants, dashed line denotes the affected hand; solid line unaffected hand. Violin plots reflect able-bodied control’s Pre – Post (6 m) values. (c) Between-session differences in finger correlation coefficients. Difference values are depicted for the (left) missing or non-dominant hand of controls and (right) intact or dominant hand of controls. The difference values are ordered to reflect the increasing gap between sessions: [Pre1-Pre2] minus [Pre Avg. – Post1 (3 m)] minus, [Pre1-Pre2] minus [Pre Avg. – Post2 (6 m)] and [Pre1-Pre2] minus [Pre Avg. – Post3 (1.55/y)]. Each violin plot reflects an individual finger. When testing whether the case-study participants showed a unique reduction in the average correlation, across fingers, relative to controls, for the missing hand, only P3, at the 3-month timepoint, for the missing hand (not intact), showed a significant pre-post reduction in the average correlation coefficient, relative to controls (t(15) = −2.59, puncorr=0.02). However, this difference returned to the typical range of controls when later tested at the 6-month timepoint (t(15) = −1.23, puncorr=0.23). All other annotations are as in Fig. 2. We only report statistics when significant.