The graph-paper effect: subjective stereoscopic patterns induced by moving gratings.

Smooth tracking across an oblique grid pattern produced hallucinations of vertical and/or horizontal striations which moved with the eyes. The effect was produced by single or multiple gratings, monocularly or binocularly, but in the latter case it appeared to lie steresoscopically in the plane of fixation. Gratings containing thin lines or sawtooth edges of moderate contrast were particularly effective stimuli but sine or square waves were not. The subjective stripes had an apparent edge polarity which was opposite to that of the inducing edges or, with thin inducing lines, was determined by the lines' polarity and movement direction. Conventional explanations (eg strobe, afterimage, or moiré effects) can be ruled out. Neither the present effect nor the "pincushion-grid illusion" are due to the presence of spurious Fourier components in the stimulus pattern. An extension of a previous model, involving disinhibitory interaction between movement and pattern channels, accounts for many aspects of this elaborate phenomenon. The detailed dependence of polarity on spatial waveform and movement direction implies: (i) the spatial second harmonic is a necessary component, (ii) the generating is approximately linear and has 90 degree phase preference, and (iii) there is a movement-induced phase lag of about 45 degrees in the response to the second harmonic.