Abstract
Grouping in the Lorenceau and Shiffrar occluded diamond illusion depends on stimulus factors such as the presence or absence of an occluder and the configuration of the moving elements (Lorenceau and Alais, Nature Neuroscience, 2001). The occluded diamond stimulus is bistable even when the occluder and inducer configurations are favorable — the moving elements can either appear to move independently or they can group together. We used this property of the illusion to study form-contingent motion grouping in patients with strabismus. We measured the relative proportion of time that 11 normal vision observers reported coherent versus independent (incoherent) motion of the elements and compared this to the same proportions in a group of 8 patients with mild to moderate strabismic amblyopia (20/25 to 20/63). Each eye was tested. Two different pairs of oscillation frequencies were tested, a faster pair (2.0 and 3.6 Hz) and a slower pair (1.5 and 2.0 Hz) because we had noted in pilot testing that coherence is easier to obtain at slower speeds. Choices were indicated via button presses during trials that lasted 17.5 sec for the faster pair and 14 sec for slower pair of oscillation frequencies. The proportion of coherent vs incoherent reports was larger at the slower speed in all eyes tested. Of greater interest is that the proportion of coherent vs incoherent reports was higher for normals than for either eye of the patients at both speeds. The normal acuity fellow eyes of the strabismus patients had the same durations of coherence as did the amblyopic eyes and therefore the decreased grouping is not due to amblyopia, per se. The presence of grouping failures in both eyes of the patients suggests that normal binocular experience is necessary for complete development of mechanisms that integrate form and motion information for scene layout.