OBJECTIVE: We aimed to use measured input-output (IO) data to identify the best fitting model for motor evoked potentials. METHODS: We analyzed existing IO data before and after intermittent and continuous theta-burst stimulation (iTBS & cTBS) from a small group of subjects (18 for each). We fitted individual synaptic couplings and sensitivity parameters using variations of a biophysical model. A best performing model was selected and analyzed. RESULTS: cTBS gives a broad reduction in MEPs for amplitudes larger than resting motor threshold (RMT). Close to threshold, iTBS gives strong potentiation. The model captures individual IO curves. There is no change to the population average synaptic weights post TBS but the change in excitatory-to-excitatory synaptic coupling is strongly correlated with the experimental post-TBS response relative to baseline. CONCLUSIONS: The model describes population-averaged and individual IO curves, and their post-TBS change. Variation among individuals is accounted for with variation in synaptic couplings, and variation in sensitivity of neural response to stimulation. SIGNIFICANCE: The best fitting model could be applied more broadly and validation studies could elucidate underlying biophysical meaning of parameters.