Figure 1.
Results of standard second level group analyses.
Reliable (ICC>0.33) and unreliable (ICC≤0.33) brain (de)activations (p<0.05, corrected) for the non-symbolic numerosity comparison task (left) and the non-symbolic exact calculation task (right). Red: Reliable activations; transparent red: unreliable activations; deactivations are analogously depicted in blue.
A: Brain (de)activation of control children (TD).
B: Brain (de)activation of dyscalculic children (DD).
C: Brain activation differences between dyscalculic and control children. (Un-) reliable higher activations of the dyscalculic children are depicted in (transparent) red, lower activations are depicted in (transparent) blue.
Figure 2.
Overall extent of significant differences in brain activation pattern.
Brain areas, where at least one dyscalculic child shows a significant difference (p<0.05, corrected) in brain activation in comparison to the control group as detected by means of the single-case comparison test by Crawford et al. [23] are conjointly visualized with areas that showed significant (p<0.05, corrected) between group differences as detected by the standard GLM (see Figure 1C).
Relatively stronger or weaker activations as detected by means of the single-case comparison test by Crawford et al. [23] are shown in violet or green, respectively, whereas group effects follow the same color convention as in Figure 1C.
Figure 3.
Significant differences in brain activation pattern for individual dyscalculic children.
Brain areas where an individual dyscalculic child shows a significant (p<0.05, corrected) difference compared to the control group as detected by means of the single-case comparison test by Crawford et al. [23] visualized with a different color per child.
Top row: relative over-activation for comparison (left) and calculation (right) task;
Bottom row: relative under-activation for comparison (left) and calculation (right) task.
Figure 4.
Combinations of ROIs leading to the best classification rate in the SVA.
A. two combinations of 7 ROIs for the comparison task; B. two combinations of 4 ROIs for the calculation task, C. one combination of 6 ROIs for the concatenated vector, D. ROIs that led most frequently to the best classification.
Figure 5.
Results of hierarchical cluster analyses.
Dendrograms of complete linkage hierarchical cluster analyses based on the vector of contrast beta weights per child within the reliability mask. D = dyscalculic child, C = control child.
A. comparison task; B. calculation task; C. conjoint data of comparison and calculation task, cluster C1: red, cluster C2: blue, cluster C3: black.
Figure 6.
Second level group analysis of two clusters.
Second level group analysis data tested against baseline for the two groups that were obtained from a conjoint cluster analysis (see Figure 4C). Brain (de)activations of the predominantly dyscalculic (C1) and the predominantly control (C2) cluster are depicted in red and blue respectively, overlap is depicted in pink.