The objectives of this study were:  (1) to quantify the variability of neural response to robot and human manual brushing, and (2) find parameters of robot brushing that mimic human manual brushing.

A robot (Dobot Magician, Shenzhen, PRC) was programmed to brush (sable hair paint brush with 0.5 cm width) the hind paw of an anesthetized rat. Multichannel electrodes (NeuroNexus, Ann Arbor, MI) were placed in the ipsilateral superficial dorsal horn (SDH; lamina II-III within the spinal segment L4-6) to monitor single unit firings evoked by the robot and a trained examiner. The proximal and distal boundaries of receptive field, identified manually by the examiner using Von Frey probes, were marked on the skin of the paw, and used as start and end points for brushing. Robot brushing was programed in an arc motion at different depths (0.5, 1, 2 and 3 mm), speeds (50, 100 and 200 mm/s) and directions (proximal-to-distal vs. distal-to-proximal). Firing patterns evoked by robot and manual brushing were compared based on the data recorded from the SDH.

Qualitatively, robot brushing generated consistent and repeatable patterns of multiunit SDH activity, while manual brushing revealed wider variability.  Robot brushing with depth=1 mm, speed=100 mm/s was most like manual forward brushing in terms of firing rate and duration.

Robot brushing appeared to demonstrate a more consistent pattern of response of SDH neurons to afferent input.  Standardization of the sensory stimulus using this novel robot tool may allow for fewer trials of applied afferent input and greater sensitivity to detect subtle changes in response in sensory experiments.