Response of Visual Cortical Neurons of the cat to moving sinusoidal gratings: response-contrast functions and spatiotemporal interactions.

I. Recordings were obtained from 108 cells near the border of areas 17 and 18 in anesthetized, paralyzed cats. Emphasis was placed on analyzing responses to moving sinusoidal gratings, whose spatial frequency, velocity, and contrast were varied systematically. In the presence of drifting gratings, about 30% of the cells responded with a periodic discharge at virtually all effective spatial and temporal frequencies. These neurons tended to be “simple,” to have very low levels of spontaneous activity, and to have the narrowest spatialand temporal-frequency response functions. Another 30% of the cells responded aperiodically at all but the lowest spatial and temporal frequencies. These cells tended to be “complex” and to have the broadest tuning curves. The remaining cells responded periodically over part of the range of effective frequencies, and could be simple, complex, or intermediate and difficult to classify. 2. For most of the neurons that responded appreciably to both directions of movement at the optimal orientation, the peak frequency (and shape) of spatialand/or temporal-frequency response functions differ by more than 0.3 octave for the two directions of movement. The peak frequency and/or bandwidth of the temporal response function for either direction may also depend on whether there is a pause between movements in the two directions. 3. The response of almost all cells increases linearly with contrast up to a saturation level. The threshold contrast shows little dependence on spatial frequency, but is a function of temporal frequency. The slopes of response-contrast functions are steepest at the optimal spatial or temporal frequencies and diminish at lower or higher frequencies. 4. The peak frequency and shape (bandwidth and high-frequency roll-off) of temporal-frequency response functions do not depend significantly on the spatial frequency used to make the measurement and the converse is true. Accordingly, it is possible to use a spatial-frequency tuning curve measured at fixed temporal frequency and a temporal curve measured at fixed spatial frequency to predict the spatial-frequency tuning curve that is obtained when all spatial frequencies are presented at the same velocity. This implies that knowledge of one spatial an .d one temporal curve of a cell allows one to calcu late its response to any combination of spatial and temporal frequencies. 5. The higher a cell’s best spatial frequency, the lower is the optimal velocity of movement at the best spatial frequency. The equation for the regression line that summarizes this correlation has velocity proportional to spatial frequency raised to the -5/4 power.