Abstract Body

Predicting the capacity to recover from speech and language impairments after stroke is essential for rehabilitation planning, expectations and goal-setting. Nevertheless, accurate predictions have been difficult to generate because there are so many factors that affect recovery. I will briefly review the challenges faced and then present the results of an investigation of several hundred adult stroke survivors that illustrates how well speech and language outcome can be predicted by a combination of lesion location, lesion size and the initial severity of impairment. I will show which lesion sites (A) cause consistently and persistently poor speech, (B) have temporary, albeit devastating early impact followed by consistently good recovery, and (C) have variable outputs. I will then discuss theoretical explanations for these three different types of lesion effects and consider how such theories might help us to improve our predictions and explanations in future.