Navigated transcranial magnetic stimulation (nTMS) is becoming a popular tool in pre-operative mapping of functional motor areas. The stimulation intensities used in the mapping are commonly suprathreshold intensities with respect to the patient’s resting motor threshold (rMT). There is no consensus on which suprathreshold intensity should be used nor on the optimal criteria for selecting the appropriate stimulation intensity (SI). In this study, the left motor cortices of 12 right-handed volunteers (8 males, age 24-61 years) were mapped using motor evoked potentials with an SI of 110 and 120 % of rMT and with an upper threshold (UT) estimated by the Mills-Nithi algorithm. The UT was significantly lower than 120 % of rMT (p < 0.001), while no significant difference was observed between UT and 110 % of rMT (p = 0.112). The representation sizes followed a similar trend, i.e. areas computed based on UT (5.9 cm(2)) and 110 % of rMT (5.0 cm(2)) being smaller than that of 120 % of rMT (8.8 cm(2)) (p ≤ 0.001). There was no difference in representation sizes between 110 % of rMT and UT. The variance in representation size was found to be significantly lower with UT compared to 120 % of rMT (p = 0.048, uncorrected), while there was no difference between 110 % of rMT and UT or 120 % of rMT. Indications of lowest inter-individual variation in representation size were observed with UT; this is possibly due to the fact that it takes into account the individual input-output characteristics of the motor cortex. Therefore, the UT seems to be a good option for SI in motor mapping applications to outline functional motor areas with nTMS and it could potentially reduce the inter-individual variation caused by the selection of SI in motor mapping in pre-surgical applications and radiosurgery planning.