Insect stereopsis demonstrated using a 3D insect cinema.

Vivek Nityananda,Stuart Crichton,Jenny Read,Ghaith Tarawneh,Ronny Rosner,Judith Nicolas

doi:10.1038/srep18718

Abstract

Stereopsis - 3D vision – has become widely used as a model of perception. However, all our knowledge of possible underlying mechanisms comes almost exclusively from vertebrates. While stereopsis has been demonstrated for one invertebrate, the praying mantis, a lack of techniques to probe invertebrate stereopsis has prevented any further progress for three decades. We therefore developed a stereoscopic display system for insects, using miniature 3D glasses to present separate images to each eye, and tested our ability to deliver stereoscopic illusions to praying mantises. We find that while filtering by circular polarization failed due to excessive crosstalk, “anaglyph” filtering by spectral content clearly succeeded in giving the mantis the illusion of 3D depth. We thus definitively demonstrate stereopsis in mantises and also demonstrate that the anaglyph technique can be effectively used to deliver virtual 3D stimuli to insects. This method opens up broad avenues of research into the parallel evolution of stereoscopic computations and possible new algorithms for depth perception.

Highlights

Depth-perception, given the relative simplicity of their nervous systems
In this study we attempted to develop a new technique that would enable the use of arbitrary binocular stimuli to investigate insect stereopsis
Given these encouraging results regarding effective crosstalk, we examined whether the mantis would experience illusory stereoscopic depth due to binocular disparity cues presented via spectral filters

Summary

Introduction

Depth-perception, given the relative simplicity of their nervous systems. In this study we attempted to develop a new technique that would enable the use of arbitrary binocular stimuli to investigate insect stereopsis. We used a spectrophotometer to measure the light transmitted through both blue and green filters when the monitor displayed green and blue stimuli, and weighted this by the mantis spectral sensitivity[27]. We found a relatively low level of physical crosstalk for both the green and the blue filters (< 1–11% at maximum luminance; Fig. 2a) after weighting by the mantis spectral sensitivity.

Results

Conclusion