Abstract: Transcranial direct current stimulation (tDCS) exerts beneficial effects on motor recovery after stroke, presumably by enhancement of adaptive neural plasticity. However, patients with extensive damage may experience null or deleterious effects with the predominant application mode of anodal (excitatory) stimulation of the damaged hemisphere. In such cases, excitatory stimulation of the non-damaged hemisphere might be considered. Here we asked whether tDCS exerts a measurable effect on movement quality of the hemiparetic upper limb, following just a single treatment session. Such effect may inform on the hemisphere that should be excited. Using a single-blinded crossover experimental design, stroke patients and healthy control subjects were assessed before and after anodal, cathodal and sham tDCS, each provided during a single session of reaching training (repeated point-to-point hand movement on an electronic tablet). Group comparisons of endpoint kinematics at baseline-number of peaks in the speed profile (NoP; smoothness), hand-path deviations from the straight line (SLD; accuracy) and movement time (MT; speed)-disclosed greater NoP, larger SLD and longer MT in the stroke group. NoP and MT revealed an advantage for anodal compared to sham stimulation of the lesioned hemisphere. NoP and MT improvements under anodal stimulation of the non-lesioned hemisphere correlated positively with the severity of hemiparesis. Damage to specific cortical regions and white-matter tracts was associated with lower kinematic gains from tDCS. The study shows that simple descriptors of movement kinematics of the hemiparetic upper limb are sensitive enough to demonstrate gain from neuromodulation by tDCS, following just a single session of reaching training. Moreover, the results show that tDCS-related gain is affected by the severity of baseline motor impairment, and by lesion topography.