QUANTIFYING SPECTRAL AND TEMPORAL FEATURES OF BIRDSONGS USING A LIGHT-WEIGHT IMPLANTABLE DEVICE

Abstract

Social cues modulate all sorts of communication in a wide range of species, and these modulatory changes make communication signals more salient. The songbird vocal system has emerged as thebest-developed model for the neural basis ofvocal communication,speech acquisition and production since songbirds are among the very few animals that learn their vocalizations through vocal imitation, like humans. In order to better understand the basis of vocal communication in a social context, it is important to observe and collect data from specimen in natural, stress-free environment, but most importantly at individual-level. While microphones are a good tool to record the vocalizations of songbirds in isolation, they are not able to record the vocalizations of multiple birds within the same environment due to sound cross- contamination. In this work, we aim to implant piezoelectric accelerometers on the skulls of songbirds which record bone-conducted recordings that are uncontaminated by airborne sounds, thereby rendering the study of vocal communication of multiple songbirds in the same environment possible. The implantable accelerometer along with its required circuitry for signal processing and filtering would allow for recordings of free-moving, single bird vocalizations in group settings and under different study conditions. This approach makes it possible to obtain vocalizations irrespectively of conspecific and background noises. The aim is to study the changes in the temporal and spectral features of song (syllables’ duration, pitch, frequency modulation, entropy, etc...) using accelerometers under various conditions, and compare these bone-conduction recordings to the standard microphone recordings in an attempt to assess the integrity and efficiency of the implantable device.