The multichannel mechanism in color vision has been studied. In this paper, temporal sensitivity of each channel measured using DOG color patterns modulated with single square waves of various time widths are reported. Further, frequency characterisitics were calculated from the measured results. The channel sensitive to higher spatial frequency showed temporal characteristics of wider bandwidth.
To Investigate interactions between movements of different spatial frequencies, we measured the perceived direction of a counterphasing sinusoidal grating (target) superimposed on a drifting grating (inducer). Target was perceived to move in the same direction as inducer (i.e., motion assimilation) when the spatial frequency (SF) of inducer was lower than target SF, and the temporal frequency (TF) of inducer was low. On the other hand, when inducer SF was higher than target SF, or when inducer SF was low but its TF and contrast were sufficiently high, target was perceived to move in the direction opposite to inducer (i.e., motion contrast).
The smooth pursuit eye movement system shows frequency characteristics like a simple lowpass filter when eyes track sinusoidal taqrget motions. However, if the target motion contains more than two frequency cpomponents, the response amplitude markedly reduces while the phase advances for lower frequency components. We conjectured such a non-linear nature is resulted from an adaptive control mechanism in the system, that is, the system adaptively changes the stiffness of extra-ocular muscles to increase the response amplitude against the higher frequency component which is relatively easily predictable. The control mechanism of the smooth pursuit system explained using a feed-forward model.
We have analyzed the inter-saccade intervals during binocular gaze at a brightness edge, a depth discontinuity edge, or a rim; right and left retinal images match in the edge, but not due to occlusion in the rim. In the case of edges, the histograms of the intervals are unimodal and have the same distribution, not depending on the existence of texture. In contrast, the histograms for the rim show clear differences depending on the existence of texture; the histograms in the case of no texture are similar to that for the edges, but the histograms in the case of texture are bimodal and tend to shirt toward lower value of the interval. We have conjectured that the saccade planning schema may include a, mechanism in the early stage of visual processing, which detects occlusion by matching both retinal images and causes earlier saccades.
Some characteristics of view direction generalization in the recent memory of places and faces were studied in order to examine whether three dimensional objects are represented in two dimensional or are represented in three dimensional in the human brain and what kind of transformations are carried out to recognise the input images. And we got the following results: (1) Three dimensional world is representedin two dimensional in the human brain. (2) A mirror transformation, an expansion/ contraction in the horizontal direction and so on may exist in the human brain. We may need to introduce a fourth hypothesis to explain these psychological results.