Praying mantids use complex, perceptual algorithms to identify prey objects

We assessed the visual algorithms used by two species of mantis, Mantis religiosa and Cilnia humeralis to identify potential prey items by presenting the animals with computer-generated visual stimuli that varied in several predetermined parameters (i.e., size, orientation, speed, movement pattern, contrast to background). The degree to which the mantids recognized the stimuli as prey was indicated by the rates at which they responded to the stimuli with three appetitive behaviors: visual tracking, approaching, and striking. In all cases we found that the mantids simultaneously assessed a specific number of stimulus parameters to determine if a moving object represents a potential prey item. For Mantis religiosa these included the overall size of a compact stimulus (e.g., a disk), movement pattern, apparent stimulus speed, and geometric configuration in relation to direction of movement. In addition, Cilnia humeralis recognized moving objects as potential prey based in part on stimulus-to-background contrast. This species also responded to the synchronous movement of a random dot pattern as if the pattern was a single, large moving object. The latter is a perceptual ability shared with vertebrates, including humans.