PubMed1981
If an eye movement signal is fed through a galvanic mirror, to move a projected image which a subject is inspecting, prominent objects in the image may seem to jiggle or jump with the the eye when the gain is just below the threshold for detecting a jump of the entire image (Brune and Lücking 1969). We have refined and extended this observation with both naive and practiced subjects, finding results which contradict all of the current theories about the mechanism of stabilization of the visual world and suggest that cognitive factors in perception important influences on the stabilization process. Using this method with a paired photocell system to detect horizontal eye movements, some subjects saw a prominent object in the display jump slightly while the rest of the scene remained stable. The task was done first with landscape slides, then repeated with Escher prints where two sets of alternating figures completely filled the image. Subjects could concentrate on one set of forms as the "figure" and the other as the "ground", and reverse the two at will. In a majority of practiced subjects and in smaller proportion of naive subjects, motion of part of the "figure" was seen regardless of which alternative set of forms constituted it. Reversibility of the effect controlled for influence of object size, brightness, etc. in inducing the selective jump. These and related observations show that cognitive or attentional variables are as important as image properties or gain alone in determining subjective stabilization of the visual world, though current theories (inflow, outflow, cancellation, etc.) consider image position to be simple variable. Another experiment showed that image movement on the retina during saccades cannot explain saccadic suppression of displacement.