Introduction and ObjectivesMotor evoked potentials (MEPs) have been postulated to be useful in predicting recovery in patients with motor impairment. We aimed to investigate whether MEPs elicited by transcranial magnetic stimulation (TMS), serum brain derived neurotrophic factor (BDNF) and its genotype have prognostic value on stroke recovery in patients with hand paresis due to stroke. MethodsThis was an observational cohort study. Patients underwent TMS with MEPs from abductor digiti minimi evaluation between 2–14 (D0) and 30 days (D30) after stroke and their impact on motor function of the upper limb and general outcome was assessed after 3 months (D90). The presence of a BDNF gene polymorphism was determined and serum BDNF concentrations were measured at D0, D30 and D90. ResultsThe presence of MEPs and their amplitude at rest and in effort significantly correlated with improvement of upper-limb paresis and general outcome after 3 months. Resting motor threshold did not have prognostic value. Central motor conduction time and MEP latency less consistently predicted stroke outcome or motor deficit improvement. Neither BDNF polymorphisms nor BDNF concentration at D0, D30 and D90 corresponded with the degree of paresis or the independence of patients 3 months after stroke. ConclusionsThe presence of MEPs and their amplitude are useful predictors of upper-limb motor function recovery and general outcome after stroke. BDNF concentration and its genotype had no prognostic value. Further studies conducted on large cohorts are necessary to determine the usefulness of these methods in motor recovery and stroke outcome prediction.