The authors' aim was to determine the cortical mechanisms that underlie the transition from effective performance to its disruption. They thus used transcranial magnetic stimulation (TMS) to study changes of corticospinal excitability after a motor exercise that did not produce overt or perceived neuromuscular fatigue. Forty-four subjects performed either 5 or 10 min of repetitive finger movements paced by tones at 2 Hz, a frequency below the spontaneous movement rate. Changes of corticospinal excitability were assessed with TMS at rest and during motor response preparation (premovement facilitation paradigm). Over time, variability of movement rate increased, while the average movement rate shifted toward self-paced rhythms, without significant changes in other kinematic parameters. Amplitudes of motor evoked potentials at rest decreased depending on task duration and TMS intensity. Moreover, 5-min exercise induced fully compensatory increases in premovement facilitation, while 10-min exercise produced partially compensatory increases with loss of temporal modulation. Our findings suggest that protracted exercise induces significant decrements in corticospinal excitability with initial impairment of the phasic motor neurons that are recruited at higher stimulus intensities. Changes in premovement facilitation likely represent compensation of premotor areas for decreased efficiency of the primary motor cortex induced by exercise.