Paired associative stimulation (PAS) can increase motor cortical excitability, possibly by long-term potentiation (LTP)-like mechanisms. As the capability of the cortex for plasticity decreases with age, we were interested here in testing interindividual variability and age-dependency of the PAS effect. Motor-evoked potentials (MEPs) were recorded from the resting right abductor pollicis brevis muscle before and for 30 min after PAS in 27 healthy subjects (22-71 years of age). PAS consisted of 225 pairs (rate, 0.25 Hz) of right median nerve stimulation followed at an interval equaling the individual N20-latency of the median nerve somatosensory-evoked cortical potential plus 2 ms by transcranial magnetic stimulation of the hand area of left primary motor cortex (PAS(N20+2)). The PAS(N20+2)-induced changes in MEP amplitude (ratio post PAS/pre PAS) were highly variable (1.00 +/- 0.07, range 0.36-1.68). Fourteen subjects showed the expected LTP-like MEP increase (responders) while 13 subjects showed a long-term depression (LTD)-like MEP decrease (non-responders). Responders had a significantly lower resting motor threshold (RMT) and minimum stimulus intensity to elicit MEPs of 1 mV (MEP(1 mV)) than non-responders. RMT and MEP(1 mV) correlated significantly negatively with the PAS(N20+2) effect. The absolute PAS(N20+2) effect size irrespective of its direction decreased with age (r = -0.57, P = 0.002), i.e., LTP-like and LTD-like plasticity were large in young subjects but substantially smaller in elderly subjects. In conclusion, measures of motor cortical excitability (RMT, MEP(1 mV)) and age determine direction and magnitude of PAS effects in individual subjects.