In order to understand the complex functional organization of the motor system, it is essential to know the anatomical and functional connectivity among individual motor areas. Clinically, knowledge of these cortico-cortical connections is important to understand the rapid spread of epileptic discharges through the network underlying ictal motor manifestation. In humans, however, knowledge of neuronal in vivo connectivity has been limited. We recently reported a new method, 'cortico-cortical evoked potential (CCEP)', to electrically track the cortico-cortical connections by stimulating a part of the brain through subdural electrodes and recording the cortical evoked potentials that emanate from a distant region of the cortex via neuronal projections. We applied the CCEP methodology to investigate in vivo cortico-cortical connections between the lateral motor cortex [LMCx; sensorimotor (SM) and lateral premotor areas] and the medial motor cortex [MMCx; supplementary motor area proper (SMA), pre-SMA and foot SM]. Seven patients with intractable partial epilepsy were studied. These patients had chronic implantation of subdural electrodes covering part of the lateral and medial frontal areas. As a part of the routine pre-surgical evaluation, comprehensive cortical mapping was performed by electrical stimulation of the subdural electrodes, and the precise localization of the subdural electrodes was defined by MRI co-registration. Single-pulse electrical stimuli were delivered to MMCx (7 patients) and LMCx (4), and CCEPs time-locked to the stimuli were recorded by averaging electrocorticograms from LMCx and MMCx, respectively. Short-latency CCEPs were observed when stimulating MMCx and recording from LMCx (mean latency: 21.6 ms, range: 9-47 ms) and vice versa when stimulating LMCx and recording from MMCx (mean latency: 29.4 ms, range: 11-57 ms). In terms of the location of these stimulus sites and CCEP responses along the rostrocaudal axis, regression analysis revealed a consistent correlation between the sites of stimulation and maximum CCEP for stimulation of both MMCx and LMCx. Functionally, stimulation of the positive motor areas in MMCx elicited CCEPs at the somatotopically homologous regions in LMCx (71%). The same findings were observed in MMCx (82%) upon stimulation of LMCx. In four subjects in whom bi-directional connectivity was investigated by stimulating both MMCx and LMCx, reciprocality was observed in the majority of connections (78-94%). In conclusion, the present study demonstrated a human motor cortico-cortical network connecting (i) anatomically homologous areas of LMCx and MMCx along the rostrocaudal cognitive-motor gradient; and (ii) somatotopically homologous regions in LMCx and MMCx in a reciprocal manner.