Item

Abstract

Visual movement detection has been investigated both under photopic and scotopic light conditions by measuring the optomotor turning responses of walking flies,Musca domestica. From the data it is concluded that the spatial sampling pattern underlying movement detection changes with the average stimulus brightness. At high luminance nearest-neighbour interactions clearly dominate whereas at very low light intensities interactions between receptors having one, two and three times the minimum angular separation contribute with about equal strength to the response (Figs. 6, 7). This change in the spatial interaction pattern may be based on neuronal recruitment of wide-angle movement detectors at low light levels or, alternatively, on neural pooling of signals from neighbouring receptors prior to the movement-specific interactions. Both mechanisms may provide a gain in absolute light sensitivity at the cost of spatial acuity.

The temporal properties of movement detection also change with stimulus brightness. High grating speeds are detected less efficiently at low luminance (Fig. 3). These temporal changes may be attributed to equivalent changes in the photoreceptor responses.

Negative optomotor responses may be elicited by a pair of test stimuli separated in visual angle by about 15° corresponding to 7–8 rows of ommatidia (Figs. 9, 10). This unexpected behaviour is suggested to reflect the influence of lateral inhibition which extends, in the periphery of the visual system, with decreasing strength over a range of at least 5 rows of ommatidia. Movement-specific interactions on the other hand do not appear to extend beyond 4–5 rows of ommatidia.