Acoustic signal detection of tropical insects under noise

This study combines field and laboratory investigations on the ecology, behaviour and sensory physiology of acoustic signal detection under noise in tropical crickets and katydids. Most species of both groups of insect use air-borne sound for communication between the sexes, but also show some kind of bat-avoidance behaviour during flight. The mechanism to distinguish "good" from "bad", however, is different. Crickets show categorical perception and respond to any stimulus above 15 kHz with avoidance, and positive steering towards stimuli below 10 kHz during flight. By contrast, katydids cannot use frequency as a cue for discrimination, since their own mating signals use the high frequency channel as well. The project will investigate the different mechanisms these two groups of insects might use to solve the same problem of detecting and discriminating conspecific mates and predators. For both contexts, however, the nocturnal tropical rainforest with its strong high-frequency background noise imposes a problem for signal detection. In the project, we will investigate the consequences of various sensory properties (absolute sensitivity, tuning of ears, and central nervous mechanisms), which are different between crickets and katydids. Although e.g. the higher sensitivity of katydids produces an advantage because it increases the detection range for echolocation calls of bats, and gives katydids more degrees of freedom to evolve anti-bat behaviours compared to crickets, there is a trade-off between increased sensitivity of a sensory system and the potential for masking by noise. Thus, there is the potential of confounding conspecific calls with bat predators ("false alarms"). The overall aim of this study is thus to describe and compare behavioural, sensory and central nervous adaptations which allow reliable detection of conspecific mates and bats within noisy rainforest habitats.

In insects hearing evolved in two different contexts: (1) communication with mates and rivals, and (2) the detection of predators. We investigated in the tropical rainforest of Panama, how the detection of signals in these contexts is constrained by high background noise levels produced by the many different species of insects and frogs calling at the same time. Thus they suffer from the same problem as humans at a cocktail-party, the masking of signals by background sound. We could identify a number of mechanisms both in the sensory periphery and in the central nervous system which strongly reduce the problem of masking. Some of the mechanisms are surprisingly similar to those used by humans when attending to speech at a cocktail-party. These include (1) the sharpening of filters in hearing organs, (2) mechanisms to selectively attend to a subset of signals interesting for the insect, or (3) the use of directionality for improved signal detection. In response to one of their main predators, echolocating bats, many rainforest crickets exhibit avoidance behavior in flight, either by turning away from the ultrasound which these bats produce, or by stopping flight altogether. Surprisingly, the behavioral threshold for the avoidance response is quite high at about 80 dezibels, which intuitively appears to be maladaptive, because in the sensory arms race between predator and prey, the high threshold produces a disadvantage for the prey because it decreases the detection range for echolocation calls. However, since these crickets do not distinguish between ultrasound echolocation signals of bats, and background sound containing ultrasound, low behavioral thresholds would induce many false avoidance responses to irrelevant signals. We confirmed this hypothesis by experimentally manipulating sensory thresholds in grasshoppers and quantifying the amount of false responses, both in the laboratory and directly in the nocturnal rainforest. Thus the rules for responding to predators depend strongly on the kind of acoustic environment, and even high thresholds for such responses are adaptive under noisy conditions. In addition to avoidance responses in flight, acoustic insects have evolved other defenses against predators. Gleaning bats usually use prey-generated sounds to detect and locate prey. We collected insect remains from roosts of one of these bat species, identified insect remains and quantified the proportion of prey with defenses against predatory bats based on defenses described in the literature. Most remains were from acoustic insects, and half of those were from species with documented defenses against gleaning bats. Our results show that this group of bats can occasionally overcome defenses of their acoustic insect prey.