TrackUSF, a novel methodology for automated analysis of ultrasonic vocalizations, reveals modified social communication in a rat model of autism

Abstract

Rodents emit various social ultrasonic vocalizations (USVs), which reflect their emotional state and mediate social interaction. USVs are usually analyzed by manual or semi-automated methodologies categorizing discrete USVs according to their structure in the frequency-time domains. This laborious analysis hinders effective use of USVs for screening animal models of human pathologies associated with modified social behavior, such as autism spectrum disorder (ASD). Here we present a novel, automated methodology for analyzing USVs, termed TrackUSF. To validate TrackUSF, we analyzed a dataset of mouse mating calls and compared the results with a manual analysis by a trained observer. We found that TrackUSF was capable of detecting most USVs, with less than 1% of false-positive detections. We then employed TrackUSF to social vocalizations in Shank3-deficient rats, a rat model of ASD and found, for the first time, that these vocalizations exhibit a spectrum of deviations from pro-social calls towards aggressive calls.