We aimed to determine whether neuropsychological deficits and brain atrophy could predict the efficacy of non-pharmacological interventions.Forty-six participants with mild-to-moderate dementia were monitored for 6 months; 25 underwent an intervention involving physical exercise with music, and 21 performed cognitive stimulation tasks.Participants were categorized into improvement (IMP) and no-IMP subgroups.In the exercise-with-music group, the no-IMP subgroup Hardware performed worse than the IMP subgroup on the Rivermead Behavioural Memory Test at baseline.
In the cognitive-stimulation group, the no-IMP subgroup performed worse than the IMP subgroup on Raven’s Colored Progressive Matrices and the cognitive functional independence measure at baseline.In the no-IMP subgroup, voxel-based morphometric analysis at baseline revealed more extensive gray matter loss in the anterior cingulate Wooden Figure gyrus and left middle frontal gyrus in the exercise-with-music and cognitive-stimulation groups, respectively.Participants with mild-to-moderate dementia with cognitive decline and extensive cortical atrophy are less likely to show improved cognitive function after non-pharmaceutical therapy.