Recent studies in neurorehabilitation research show that success in aphasia therapy is linked to a high treatment frequency. Computer-aided therapy offers a solution to the dilemma of increasing therapy frequency while maintaining or reducing the load on therapists' resources. Until now it has, however, been unclear which patients can reasonably be treated with computer-aided therapy. The study investigates therapists' indication choices of a new computer-aided training programme designed to supplement conventional speech therapy for aphasics (EvoCare therapy, Dr Hein GmbH, Nuremberg, Germany). The goal was to ascertain which patients were suitable for the training and which (individual) allocation criteria played a role in the therapists' decision for or against the new therapy concept. The study is an explorative prospective application study in inpatient rehabilitation care. To determine the allocation criteria, comprehensive medical, psychosocial and neurolinguistic questionnaires were used. The speech therapists were surveyed separately. Forty-nine of the 75 patients were treated with EvoCare therapy; the others received purely conventional speech therapy. Patients chosen for computer-aided therapy suffered more frequently from problems with everyday mobility and serious neurolinguistic disorders. Type and extent of brain damage, degree of reliance on caregivers, sensomotoric and cognitive deficits and depression were irrelevant to the allocation.