Fig. 8.

Comparison of performance on the category fluency task (across all categories) and the letter fluency task (across the letters F, A and S). Left panels show means and standard errors of the number of correct responses generated for each patient group on each task. Right panels show the relationship between the composite measure of semantic impairment and the number of correct responses generated, normalized by the control mean and standard deviation in each condition. Linear fits between these variables are shown; where the simple correlations are stronger than expected by chance, r² values are also shown.