Abstract
{\it Purpose:} Receiving the Arm Ability Training for three weeks improves motor function among stroke patients with mild arm paresis. There is, however, a considerable variability in recovery among patients receiving the Arm Ability Training. The study investigated whether intersubject differences in movement-related electric brain activity explain the variability in motor recovery.
{\it Methods:} In a sample of 9 patients receiving the Arm Ability Training both motor performance using a standardised arm function test (TEMPA) and movement-related electric brain activity, i.e. slow cortical potentials (DC) as well as event-related desynchronisation of alpha (alpha-ERD) and beta activity (beta-ERD), were assessed before the Arm Ability Training was started. Stepwise multiple regression was used to establish the best predictive model for motor improvement scores (TEMPA difference from pre to post test three weeks later).
{\it Results:} Stepwise multiple regression indicated that electric brain activity during movement preparation explained the variance of motor improvement scores completely. Further, electric brain activity during movement as well as baseline motor performance accounted each for 50% of the variance of motor improvement scores.
{\it Conclusions:} Activation of sensorimotor areas during movement preparation and deactivation of other cortical areas during movement execution seem to be factors that predict a favourable outcome after training.