Single-unit responses to large-field movement (angular velocity, w = 0.25-42 degrees/s) of sine-wave gratings of different spatial wavelength (lambda = 5.2-41 degrees) and contrast have been recorded in the nucleus of the basal optic root (nBOR) of the accessory optic system (AOS) of the pigeon. 2. The steady-state response to moving sine-wave gratings increases with increasing contrast to reach a saturation level at 25%. 3. Generally the steady-state responses of the cells passed through a maximum when stimulated at various velocities. In 12 of the 15 cells tested with six different velocities and four different spatial wavelengths, the location of the response maximum on the velocity scale depended on the spatial wavelength (lambda) used. That is, in these cells the response depends on the temporal frequency (tf = w/lambda) of the stimulus and not on its velocity alone. This is in agreement with the prediction of the theory of motion detection according to the basic version of the correlation scheme. 4. The temporal frequency for maximal response of individual cells shifts to higher values when the contrast of the sine-wave gratings is reduced to 5%. 5. The steady-state response of 16 of the recorded directional selective cells (53) is modulated with the temporal frequency of the stimulus, regardless of the phase of the grating at the beginning of its movement. 6. In phasic-tonically responding cells, the phasic response peak decays to the steady-state level with a time constant that becomes shorter as the temporal frequency of the stimulus increases. 7. The basic version of the correlation scheme includes only the time constant of one low-pass filter. Therefore the phasic response is expected to decay to the steady-state level with one and the same time constant, and the position of the maximal response on the temporal frequency scale should not be influenced by a change of pattern contrast. According to the model, phase-dependent modulations of the steady-state response should occur only when the spatial wavelength of the stimulus pattern is large compared with the sampling base of the underlying detector. Consequently the results given in points 4-6 cannot be described by a basic version of the correlation scheme.
