Deep Learning based Automated Ultrasonic Vocalizations Scoring: DLAS

Ultrasonic vocalizations (USVs) provide an excellent behavioral measure that can be used to understand the effects of traumatic stressors on behavior and can be used to screen and identify therapeutic treatments. This provides a strong justification for the use of rodent USVs as a model for emotional processing in PTSD. Since there are no commercially available USV assessment software programs that reliably account for high variability in recording environments and signal-to-noise ratios across USV laboratories, manual analysis has remained the gold standard for accurate USV assessments.To address the critical need of an automated USV assessment program and the limitations of current solutions, Intelligent Automation, Inc. (IAI), along with our collaborator Prof. Christine Duvauchelle from the University of Texas at Austin (UT), proposes to develop a Deep Learning based Automated Ultrasonic Vocalizations Scoring program (DLAS). DLAS offers the key advantage that it is highly adaptive and flexible that it can be used for any user specified test environment and user defined set of USV classes, incorporated in DLAS profiles.