Fig. 3

Split-data representational dissimilarity matrices. Results are for the same participants shown in Fig. 2 (separated by column), ages 8, 10, and 20, respectively. The top row (A, B and C) shows the results for the affected limb, and the bottom row (D, E and F) shows the results for the unaffected limb. In each panel the diagonal represents within-movement dissimilarity between the data-splits. The off-diagonals represent between-movement dissimilarities. As expected, most between-movement dissimilarity values are larger (lighter shading) than the within-movement dissimilarity values (darker shading). Our observation from Fig. 2, that the first example participant had poor classification of pinch in their affected limb is supported here; the within-pinch dissimilarity for their affected limb (A) is larger than the dissimilarity between pinch and key, as well as between pinch and point. We can also observe that poor dissimilarities between-movements are not restricted to the affected limb. The unaffected limb of the second example participant (E) shows fairly high within-pinch dissimilarity values, relative to the dissimilarities between pinch and all other movements.