Functional properties of rotation-sensitive neurons in the posterior parietal association cortex of the monkey

We studied the functional properties of rotation-sensitive (RS) neurons of the posterior parietal association cortex in detail. We classified 58 neurons as RS neurons on the basis of statistical analysis, to indicate that their responses to rotary movement were significantly greater (P<0.01) than those to linear movement of the same stimulus. We calculated rotation index, 1 — (L/R), in 82 cells, where L/R is the ratio of net response to linear movement to that to rotary movement. All the RS neurons had rotation index greater than or equal to 0.3. The recording site of these RS neurons was localized in the posterolateral part of area PG (area 7a of Vogt), on the anterior bank of the caudal superior temporal sulcus (STS), in the region partly overlapping the medial superior temporal (MST) area. We compared the response of RS neurons to rotation with that to shearing movement as well as to linear movement. In the majority of RS neurons the ratio of shearing response to rotation response (S/R) was smaller than the ratio of linear response to rotation response (L/R), indicating that the response to rotation was not due to a simple combination of linear movements in the opposite direction. Most of the RS neurons responded to the rotary movement of a single spot as well as that of a slit, although the response was smaller (average 70%) for the former. Most of the RS neurons had large receptive fields (60–180° in diameter) and their responses were independent of the position within the receptive field. The responses of most RS neurons increased monotonically with the increase in angular velocity and were also dependent on the size of the stimulus, although the rate of increase was small when the length was more than 10°. The majority of RS neurons (37/58) responded better to rotation in depth than to that in the frontoparallel plane. Some of them (12/37) responded to diagonal rotation rather than to sagittal or horizontal rotation. We found that some depth RS neurons showed reversal in the preferred direction when we used a trapezoidal window-like plate as the rotating stimulus in the monocular viewing condition, just as occurs in the case of the Ames window illusion. The response of some RS neurons (5/7) was enhanced by tracking eye movement. The enhanced responses were observed during rotary tracking but not during linear tracking. Other RS neurons (n = 2) showed maximum response to the rotation of the monkey chair in the light, as a result of convergence of visual and vestibular signals. We concluded that the continuous change of direction of movement was the most important cue for RS neurons to respond selectively to rotary movement in contrast to linear translational movement, and that these neurons were likely to discriminate the direction and orientation of the plane of rotation of the object in space.