RAPID COMMUNICATION Signaling of Object Approach by the DCMD Neuron of the Locust

Rind, F. Claire and Peter J. Simmons. Signaling of object ap16 mW/cm, and dark shapes of 4 mW/cm in intensity moved against it. Stimuli were controlled with a Cambridge Research proach by the DCMD neuron of the locust. J. Neurophysiol. 77: 1029–1033, 1997. The locust descending contralateral motion deSystems VSG2/1 board and RG2 raster generator, providing a refresh rate of 200 Hz, at a resolution of 437 lines by 438 dots. tector (DCMD) responds to movements anywhere within a wide visual field, but responds most strongly to the images of apIn 17 experiments, the locust viewed a Sony VDU monitor controlled by a Silicon Graphics Indigo-2 Computer. Images were proaching objects. It has been claimed that the response peaks before the end of an approach movement, providing a signal that refreshed at 72 Hz and screen resolution was 1,024 1 1,248 pixels. In 14 of these experiments the eye was 150 mm from the screen, anticipates collision. However, we find that when the locust eye is presented with appropriate computer-generated images of apwhich subtended 99 1 867 at the eye. Dark squares of intensity measured at the eye of 25 mW/cm, moved against a green (540 proaching objects, the response builds up until after movement has stopped. Premature peaking in the response is due to failure to nm) background of 32 mW/cm in intensity. In a further three experiments the stimulus conditions matched those used by Hatsostimulate the eye with sufficiently small and frequent jumps in image edges. We conclude that the DCMD signals impending collipoulos et al. (1995). The eye was 70 mm from the screen, which subtended 149 1 1297 at the eye. Dark squares of intensity measion by tracking edge motion throughout object approach. sured at the eye of 0.02 or 6.49 mW/cm moved against a green background of either 8.05 or 14.54 mW/cm, respectively. Spikes were recorded extracellularly and converted to 0.1-ms I N T R O D U C T I O N pulses with a height discriminator. They were collected to disk in Neural mechanisms for anticipating collision and trigreal time and analyzed with the use of Cambridge Electronic Degering avoidance movements are not well understood. Resign ‘‘Spike2’’ Software. Different types of stimuli were intercently, Hatsopoulos et al. (1995) described a novel mechaleaved at intervals of 90 s. nism for anticipating collision, in which the response of the descending contralateral motion detector (DCMD) neuron R E S U L T S of the locust peaks and declines before an object finishes its Previously (Rind and Simmons 1992) we employed imapproach toward the eye. They claim this peak would proages of approaching rectangles, so that the corners of the vide a cue for the timing of escape behavior. images move much faster over the retina than the center of The DCMD is a widefield visual neuron that responds each straight edge. We have now also generated images of briskly to any novel movement, but only generates maincircular disks, in which all movements over the retina are tained vigorous responses when an object approaches the at an equal speed, and we find that, for a variety of object eye (Rind and Simmons 1992). Increases in both the extent sizes (Fig. 1, B and C) and speeds, the DCMD response and the velocity of travel of image edges across the eye are continues until several tens of milliseconds after the end of critical cues for the selective responses to approaching obimage movement. In these stimuli, like those of rectangles jects (Simmons and Rind 1992). Our previously published we have used previously, the image stops expanding before recordings (Rind 1996; Rind and Simmons 1992) have not its edges reach the limit of the screen, so that the object shown that the response by the DCMD to object approach appears to stop moving when it reaches the screen rather declines before the end of movement, and we have therefore than coming closer to the locust (see Fig. 1A) . We have repeated and extended our previous experiments. We cannot also recorded responses to images of rectangles that expand substantiate the central finding of Hatsopoulos et al. (1995), to the limit of the screen. For objects smaller than the screen, and suggest that the graphics computer system they emthis mimics the approach by the object to within a few milliployed had insufficient temporal resolution to generate the seconds of collision (Fig. 1A) . Again, for a variety of speeds illusion of an approaching object for the DCMD. (Fig. 1, D–F) and sizes of object, the response usually continues to increase after movement has ended. Even at M E T H O D S slow simulated approach speeds, the response shows little In general, methods were similar to those described previously decline until well after movement has stopped (Fig. 1D) . (Rind and Simmons 1992). Stimulus conditions are summarized Neither image contrast nor the absolute brightness of the in Fig. 1A. The right eye of a locust, usually Locusta migratoria background (up to 35 mW/cm) affects this feature of the but occasionally Schistocerca gregaria, viewed a monitor screen response. while the left eye was shielded with black card. Image size was When we displayed images on a graphics computer VDU monitored by the output of a digital to analog converter (DAC). screen, we found that for squares up to 40 mm across the In 10 experiments, the monitor screen was a Kikusui COS1611 X response also continued to increase after the end of moveY display (100 1 80 mm; P31 phosphor) at 100 mm from the eye. Usually, the background had an intensity, measured at the eye, of ment. However, for larger squares, although the response