The existence of direct projection from the cerebral cortex to the subthalamic nucleus has been demonstrated with anatomical techniques. By use of the Nauta Method or the Fink-Heimer method, somatotopically organized projections from the precentral or motor cortex have been described in monkey 4, 14, 1s and in cat1, 4. More recently, autoradiographic studiesS, la, 12 have shown that the cortical projection mainly arises from the precentral gyrus. The labeled axons terminate in ipsilateral subthalamic nucleus in a patch-like pattern. An ultrastructure study by Romansky et al. 18 on the subthalamic nucleus confirmed the above findings of light microscopy. It was reported that the lesions of the primary sensorimotor areas in cats produced degeneration of terminal boutons containing small vesicles synapsing mainly on spines and small dendrites of subthalamic neurons 18. In spite of these ever mounting anatomical studies on the cortico-subthalamic projection, there has been no electrophysiological study to determine the functional significance of the cortical afferents to the subthalamic nucleus. In this study the combined technique of intracellular recording and intracellular labeling of neurons with HRP was employed to assess the nature of cortical inputs to the subthalamic nucleus. Intracellular recordings were made from subthalamic neurons in rats anesthetized with Ketamine (initial dosage 100 mg/kg ip supplemented 70 mg/kg im as needed). The subthalamic nucleus was approached from the dorsal surface. Bipolar stimulation electrodes were placed stereotaxically in the motor cortical area for jaw and forearm area 6 and in the substantia nigra (AP+ 2.7, Lat 1.8 and 2.6). Substantia nigra stimulation was used to activate subthalamic neurons antidromically. Current pulses of 0.05-0.1 msec were used for stimulation. Recording microelectrodes were filled with 4~ HRP in 0.4 M Tris. HC1 buffer (pH 7.6) and KC1 (0.5 M). Electrode resistances varied from 40 to 60 MfL HRP was ejected using 300 msec pulses of 2-5 nA depolarizing current 1.5 pulse/sec for 2-5 min. The perfusion and tissue processing procedures for HRP were similar to those described previously 16. All stimulation sites were identified histologically.