DeepSqueak: a deep learning-based system for detection and analysis of ultrasonic vocalizations

Neuropsychopharmacology volume  44, pages 859–868 (2019 )Cite this article

Rodents engage in social communication through a rich repertoire of ultrasonic vocalizations (USVs). Recording and analysis of USVs has broad utility during diverse behavioral tests and can be performed noninvasively in almost any rodent behavioral model to provide rich insights into the emotional state and motor function of the test animal. Despite strong evidence that USVs serve an array of communicative functions, technical and financial limitations have been barriers for most laboratories to adopt vocalization analysis. Recently, deep learning has revolutionized the field of machine hearing and vision, by allowing computers to perform human-like activities including seeing, listening, and speaking. Such systems are constructed from biomimetic, “deep”, artificial neural networks. Here, we present DeepSqueak, a USV detection and analysis software suite that can perform human quality USV detection and classification automatically, rapidly, and reliably using cutting-edge regional convolutional neural network architecture (Faster-RCNN). DeepSqueak was engineered to allow non-experts easy entry into USV detection and analysis yet is flexible and adaptable with a graphical user interface and offers access to numerous input and analysis features. Compared to other modern programs and manual analysis, DeepSqueak was able to reduce false positives, increase detection recall, dramatically reduce analysis time, optimize automatic syllable classification, and perform automatic syntax analysis on arbitrarily large numbers of syllables, all while maintaining manual selection review and supervised classification. DeepSqueak allows USV recording and analysis to be added easily to existing rodent behavioral procedures, hopefully revealing a wide range of innate responses to provide another dimension of insights into behavior when combined with conventional outcome measures.

Rodents engage in social communication through a rich repertoire of ultrasonic vocalizations (USVs, vocalizations >20 kHz). Rats and mice produce complex sequences of USVs throughout development and in a variety of social and motivational contexts [1,2,3,4,5,6]. These sequences are made up of a number of uniquely shaped syllables across a wide range of frequencies (20–115 kHz), and they appear to have a form of syntax which is contextually dependent [3, 7, 8]. Since their discovery, there has been a concerted effort to assess the significance of USVs in rats and mice and to utilize these innate responses as an indicator of the subjective experience of the animal. For example, high frequency (~50 kHz) USVs have been associated with positive affect in rats, while lower frequency (~22 kHz) USVs have been associated with negative affect [1, 9,10,11,12]. However, within rat 50 kHz USVs there is considerable variability in syllable type and sequence structure that may encode valuable information, and has yet to be deciphered [13]. Mouse USVs are less clearly tied to affective state, but the shape and sequence of syllables vary greatly across genetic strains [14,15,16,17], behavioral and social contexts [3, 18, 19], genetic manipulations [20], and development [4, 21].

Leave a Comment