How ambient noise may shape peripheral auditory sensitivity: a theoretical model on the trade-off between signal detection and recognition

Abstract

Ambient noise affects hearing in natural environments and may therefore affect the evolution of animal acoustic signals and auditory sensitivity. An earlier fitness model examining variable ambient noise conditions predicted higher sensitivity as the best strategy for species living in quiet habitats as opposed to lower sensitivity for the ones in noisy habitats. The trade-off between detection and recognition of acoustic signals appeared to be a key factor determining hearing sensitivity for acoustic communication in the presence of noise. The original model focused on the best auditory range of two U-shaped audiograms differing in sensitivity only (i.e., low and high). Here the model is extended by employing additional sensitivity levels and investigating the full range of hearing in order to examine (a) conditions favoring auditory sensitivity in a model characterized by the presence of multiple sensitivity and sound levels, and (b) the importance of other audiogram features, such as bandwidth and shape, which are assessed by analyzing the fitness consequences associated with bandwidth variation in the auditory system displaying best sensitivity. The model also allows theoretical insights into the importance of the detection-recognition trade-off for the evolution of the auditory critical bandwidths and ratios. The model predicted that a successful receiver should evolve an auditory system of either low or high sensitivity, but not intermediate. When low sensitivity pays, the audiogram shape should follow the profile of maximum noise levels encountered in the receiver’s environment. When high sensitivity pays, the audiogram should maximize bandwidth. A high sensitivity system with larger critical bandwidths may be successful only if the ensuing cost due to lower detection performance is outweighed by the benefit of improved sound recognition.