Figure 5.

Regression model of changed network metrics explains changes in cognitive function. (A) Bootstrap line (blue) closely approximates to the ideal line in red (adjusted R² = 76%), when observed and expected responses are equal. The vertical lines indicate the prediction standard deviation for each patient across 1,000 runs. (B) Bootstrap analysis: for each different split size, we saved the mean predictions after 1,000 runs and repeated this process 10 times. The vertical lines indicate the standard deviation of the adjusted R² and nRMSE after repeating the bootstrap analysis 10 times. (C) Distribution of adjusted R² and nRMSE by taking random selection of different sets of discriminative network metrics between early and late post-injury stages 1,000 times.