Treatment-induced plastic changes were investigated in the brains of stroke patients in the subacute stage of illness. Nine patients participated in 1 week of conventional physiotherapy. In the subsequent week conventional physiotherapy was combined with forced-use therapy. Focal transcranial magnetic stimulation was used to determine the motor output areas of the abductor pollicis brevis muscles prior to the treatment and after the first and after the second week. Motor performance was evaluated using the Nine Hole Peg Test, the Frenchay Arm Test and vigorometry to measure the grip strength. Before treatment the cortical representation area of the paretic hand muscle was significantly smaller than the contralateral side. This difference persisted after the first week of physiotherapy. In contrast, the motor output map in the affected hemisphere was significantly enlarged after forced-use therapy. This increase in motor cortex excitability was accompanied by a significant improvement in dexterity. Across the two treatment weeks the centres of the motor output maps shifted significantly stronger in the affected hemisphere than in unaffected hemisphere, suggesting the recruitment of adjacent brain areas. We conclude that the combination of forced-use therapy and conventional physiotherapy enhances motor cortex excitability and improves motor performance compared to a preceding conventional physiotherapy alone. Due to the small number of patients and the lack of a control group, these results are preliminary observations and require replication in a larger sample.