Fig 9. Histograms for left primary motor and somatosensory cortices combined.

Ordinate height of the white bars represents a count of activated voxels (p < 0 .001 uncorrected; scaled as percentage of all voxels in the selected ROIs) that showed larger percent signal change during wrist extension than during wrist flexion. These counts are plotted as a function of the number of within-session data collection runs for which extension action preference was exhibited. Black bars show for comparison the corresponding binomial distributions that would be expected given randomly assigned mutually exclusive preference labels if no true preference for extension or flexion existed (i.e., the null hypothesis). Different panels present results for the 10 participants, ordered by participant ID left to right and top to bottom. Note a trend for white bars to be taller than black bars in the left half of the plots, white bars to be shorter than black bars in the right half of the plots. For five participants (C01, C02, C03, C04, C11), observed frequencies differed significantly from binomial distributions (Kolmogorov-Smirnov tests, p < 0.05).