Distance Discrimination in a Spider

Among invertebrates and especially spiders, the Central American hunting spider Cupiennius salei has one of the best studied visual systems, which is well developed and consists of four pairs of lens eyes. Many data are available from intracellular recordings of the photoreceptors to determine the spectral sensitivity up to visual guided behaviour concerning target detection and discrimination. During behavioural studies we obtained good evidence that there might be an additional ability in visual detection. The spiders seem to display distance discrimination favouring a close over a more distant target. This is extremely rare among invertebrates, and may be unique for an animal which cannot use stereoscopic cues because there is no binocular overlap of the left and right eyes. When these animals walk towards an object to find either prey or a retreat, they do that in a zigzag style causing self-induced motion parallax between target and background, whose extent might be enough to perform this task. We want to test this observation using a locomotion compensator. The animal is walking in-place on top of a large actively driven sphere (diameter 90cm), where the motion of the animal will be completely compensated by the counter-motion of the sphere. A cylindrical 360 projection screen will surround the animal where all kinds of visual stimuli can be presented. Moreover, the visual stimulation software will generate changing images, according to the movement of the spider. This setup allows closed- and open-loop experiments with arbitrary objects and backgrounds at any "virtual" distance. With this setup we can also discriminate between motion parallax and retinal expansion. To ensure that the behaviour of animals walking on the sphere can be compared to freely walking animals, we want to perform experiments with animals walking in an arena, where they have to orient towards real targets. In a twofold simultaneous choice test, two targets of different distances in front of variable backgrounds can be tested. We can also cover different pairs of eyes to find out which are responsible for this task. This is highly interesting considering that the anterior median eyes are known to be involved in target discrimination and the secondary eyes in motion detection. This resembles the wide-spread separation of visual information into different cues such as colour, form, motion, and depth described for the mammalian visual system. We want to combine these behavioural experiments with electrophysiological telemetric measurements of eye muscle activity. The anterior median eyes have muscles that move the retina and a changing activity of these muscle during distance discrimination might indicate the involvement of these eyes in this behavioural context. Additional observations make our findings even more interesting. While the animal is approaching a target, the amplitude of zigzagging seems to be variable, which indicates a possible active adjustment as a response to the distance of the target. If this turns out to be true, it would be a hitherto unique ability for active sensing of distance.

Cupiennius salei, a Central American hunting spider, has one of the best studied visual systems among the Arachnida, with respect to its development, anatomy and physiology, but also with respect to visual guided behavior. Preliminary behavioral experiments suggested that, among many other visual discrimination abilities, C. salei is capable of visual depth discrimination. The aim of this project was to confirm these preliminary results and to create a novel setup for behavioral experiments to study the mechanisms of depth perception in more detail. In a first step we conducted arena experiments that corroborated the preliminary results. C. salei is capable of visual discrimination between a close and a far object, as long as they are presented in front of a visually structured background. Because of the almost lacking binocular overlap of the frontal eye pairs, binocularity can be dismissed as a likely depth perception mechanism. Among the number of possible cues that can be used to judge distances, motion parallax was the cue that was most compatible with our experimental results. However, many depth cues (e.g. retinal expansion, relative size, and parallax) are linked and cannot be separated experimentally in the real world experiments. In a next step, we developed a virtual reality setup (VR), based on a motion compensator and a digital visual presentation. This setup allowed testing spiders in a real-time closed-loop virtual world. The first experiments were aimed at testing whether experiments in the VR setup can reproduce previous real-world experiments. We could show that the previous results can be reliably reproduced and thus the VR setup is an adequate substitute for real- world experimental arenas. In the next step, we programmed objects with uncoupled depth cues, to narrow down the possible mechanism by which C. salei makes its decisions. Even in absence of all size related cues, spiders chose the closer object. This confirms that the most likely depth cue for C. salei is motion parallax between fore- and background. One remarkable finding was that the principal eyes were necessary for the discrimination task. Usually, motion is assumed to be detected solely by the secondary eyes in spiders and therefore this finding was surprising. Further investigation of the perception and neural computation of the depth cues is pending until suitable electrophysiological or optical methods are available. In the course of the project we conducted further experiments, both with and without the use of the VR setup. We could show, for the first time, that C. salei is capable of visual associative learning, a cognitive skill that was so far only known from jumping spiders. Further experiments, e.g. on distance discrimination, color vision, orientation in naturalistic environments and the retinal perception and computation of visual stimuli were conducted and could serve as starting point for future projects.