Striate cortex responses to periodic patterns with and without the fundamental harmonics

Abstract

1. The visual system has been modelled as a set of independent linear channels each tuned to a limited band of spatial frequency with the average bandwidth being approximately 1 octave. A great deal of psychophysical and physiological evidence supports this basic notion. However, Henning, Hertz & Broadbent (1975) have shown reciprocal masking between a fundamental frequency (1F) and a complex grating composed of higher harmonics several octaves removed ((4+5+6)F); their results clearly indicate a lack of independence.2. We recorded the activity of cells in the striate cortex of monkeys and cats using stimuli similar to those of Henning et al. to make comparisons with their psychophysical data and to test specific physiological predictions.3. We found that cells tuned to the fundamental frequency did not produce an excitatory response to the (4+5+6)F pattern. However, the response of such cells to 1F could be reduced by simultaneous presentation of (4+5+6)F. Similarly, the response of cells tuned to high frequencies, when presented with (4+5+6)F, was reduced by simultaneous presentation of 1F. However, this reciprocal inhibition could be produced between single harmonics (e.g. 1F and 4F) and was not dependent upon a special relationship between 1F and (4+5+6)F.4. When cells tuned to high frequencies were presented with the (4+5+6)F pattern they generated predictable responses in the higher harmonics (4, 5, 6) but they also generated an unexpected, non-linear, response at the fundamental frequency, 1F, even though no such low frequency component was present in the stimulus. This effect is due to the response rectification which striate cells show.5. In support of the linear independent spatial frequency channel model, we find (a) striate cells provide an excitatory response to only a limited range of frequencies, (b) they do not provide such responses to the ;apparent' yet ;missing' fundamental in the (4+5+6)F beating pattern, and (c) the response wave form to complex stimuli like (4+5+6)F is reasonably predictable (at least for simple cells) from the model. Against the model we find that (a) frequencies outside the excitatory bandpass can produce inhibition and (b) the rectification of the response wave form introduces harmonics not present in the stimulus.