The velocity dependence of direction selectivity of visual cortical neurones in the cat.

1. The range of velocities, yielding direction-selective responses, was investigated in a total of 167 direction-selective cells from areas 17 and 18 of the cat, using a high-contrast light bar moving at velocities ranging from 0.6 to 900 deg s-1. 2. 11% of the cells were direction selective over the full range of velocities tested. Most cells (66%) gave only responses at low velocities and thus were not direction selective at high velocities. The remaining cells gave responses over a broad range of velocities but the direction selectivity was limited to either high or intermediate velocities (18 and 5% of the cells, respectively). Cells with direction selectivity at high but not at low velocities had large receptive fields with non-overlapping 'on' and 'off' subregions and they responded quickly and phasically to stationary flashes. This suggests that the latter cells relied on fast and brief interactions over large distances. 3. In thirty cells the spatial and temporal limits of direction selectivity were investigated using a stroboscopically illuminated moving light bar. In all cells direction selectivity depended both on the interflash distance and the interflash time interval. Area 17 cells with large receptive field at high eccentricity tolerated much larger interflash spacings than area 17 cells with small receptive fields near the area centralis. For eleven of the thirty cells the effective interflash distance could be larger than the width of the receptive field. The largest effective interflash time interval varied between 35 and 250 ms. 4. Eight of the thirty cells were direction selective at high but not at low velocities. These eight cells all remained direction selective over large interflash distances and they required brief interflash intervals (less than or equal to 65 ms). 5. Responses to single stroboscopic flashes within the sequence were observed in ten cells, which all responded well at high apparent velocities. While most cells (eight out of ten) showed both response increments in the preferred direction and response decrements in the non-preferred, the decrements constituted the dominant element in the direction selectivity of six out of ten cells while the remaining four cells relied mainly on response increments. 6. It is concluded that the range of direction-selective velocities of some cat visual cortical cells can be predicted from a knowledge of the spatial extent and the time course of the direction-selective interactions.(ABSTRACT TRUNCATED AT 400 WORDS)