We evaluated a rehabilitation programme for executive deficits in multiple sclerosis patients by comparing outcome scores of a cognitive intervention group (CIG; n = 11) with those of a placebo group (n = 14) and an untreated group (n = 15). Executive functioning and verbal learning improved significantly more in the CIG. The treatment effect on verbal learning was still present at 1-year follow-up. Baseline brain atrophy, quantified by the brain parenchymal fraction, was associated with treatment effects for one aspect of executive functioning. Consequently, cognitive intervention may be beneficial and baseline brain atrophy has some predictive value in determining treatment outcome for executive functioning.