Neuronal plasticity is to be kept within operational limits to serve its purpose as a safe memory system that shapes and focuses sensory and motor representations. Temporal and spatial properties of motor cortical plasticity were assessed in patients with writer's cramp using a model of long-term potentiation (LTP) and long-term depression (LTD) of synaptic efficacy. Paired associative stimulation (PAS) combined repetitive electric stimulation of the median or ulnar nerve (MN or UN) with subsequent transcranial magnetic stimulation of the contralateral dominant motor cortex at 21.5 ms (MN-PAS21.5; UN-PAS21.5) or 10 ms (MN-PAS10). Motor-evoked potentials were recorded from abductor pollicis brevis (APB) muscle and abductor digiti minimi (ADM) muscles in 10 patients with writer's cramp and 10 matched healthy control subjects. Following MN-PAS21.5 or UN-PAS21.5 in non-dystonic subjects, motor responses increased if the afferent PAS-component came from a homologous peripheral region and remained stable with a non-homologous input. In contrast, following either MN-PAS21.5 or UN-PAS21.5, both APB- and ADM-amplitudes increased in patients. Compared with controls, this increase started earlier, its magnitude was larger and its duration longer. Following MN-PAS10 in controls, APB-amplitudes decreased, while ADM-amplitudes increased. In writer's cramp, the decrease of APB-amplitudes started earlier and lasted longer. Of note, ADM-amplitudes were decreased, too. LTP-like as well as LTD-like plasticity is abnormal with respect to both gain and spatial organization. These findings may help to develop a pathophysiological model explaining core features of focal dystonia.