Aphasia Therapy
Pre-surgical planning

Dyslexia Research

The lab has mainlt been involved in morphometric analyses, revealing the presence of various endophenotype in the dyslexic population.

VBM (Human Brain Mapping, 2009, 30(7))

Contrary to mainy groups, our VBM analysis failed to find reliable differences in volumes between groups (39 control subjects vs 38 dyslexic readers). However, significant correlations and anti-correlations were observed. In particular, pseudoword reading showed significant effect across groups mainly over the STS, the left/right fusiform gyri and the cerebellum. Over this last area, both groups show significant effects without overlapping, suggesting a remapping in dyslexic readers. Differences between groups were observed in their correlations with a phoneme deletion task and an irregular spelling task (lexicon). Note that no within group effect was observed in this last task and difference should be interpreted cautiously.

Correlations local brain volumes and performences in control and dyslexic subjects (reading) and control vs dyslexic subjects (spelling and phonological deletion).


Patient Classification (BMC Neuroscience)

The absence of volume differences observed above seemed to be explained by a high variance in the data. Using a different (non linear) approach, we found that 2 areas show significant difference in all dyslexics compared to controls. In fact, about half of the dyslexic readers show lower right declive volumes than controls and the other half high cerebellar declive volumes. Note that the right cerebellar declive is the exact area found by Stoodley & Schmahmann that present fMRI activation meta-analysis results during language tasks (NeuroImage 2009). A similar result was observed over the lentiform nucleus, leading to observe 4 brain phenotypes. Interestingly, although this classification into 4 groups only relies on brain measurements, it revealed itself relevant as significant differences exist between subgroups in term of behaviour, and especially in term of phonological and lexicon performances.

Illustration of the right cerebellar declive cluster where we observed 100% of dyslexic subjects being different from controls. Here the percentage map of difference (PMD) is thresholded at 95% revealing a larger cluster. When compared with the meta-analysis from Stoodley and Schmahmann (NeuroImage, 44, 2009), our cluster fall inside the right cluster observed for language and not the motor or sensorimotor clusters.

Illustration of the differences between controls and dyslexic subgroups. The classification revealed 4 dyslexic subgroups which differed mainly in terms of phonological and lexicon access performances (red histograms). Furthermore, performances varied linearly across groups (green adjustments) although the linear adjustments were not enough to explain all of performance distributions (blue histograms). Overall, these results suggest that is exists different brain phenotypes which differ behaviourally but follow a normal distribution of performances (green lines) with best performances observed for optimal grey matter volumes (controls) and lower performances for lower or higher grey matter volumes.