CORNERSTONE RESEARCH GROUP INC — Department of Defense STTR Phase I: DHP16C-003

Description

Cornerstone Research Group (CRG) and the University of Texas (UT) will team to develop a fast and accurate automated analysis program for USV scoring. This automated tool will enable greater research efficiency and throughput allowing greater strides in developing treatments for post-traumatic stress disorder through rodent-based research. Building off prior work by UT on a first generation auto-scoring algorithm (WAAVES), CRG and UT will develop an improved program which addresses various shortcomings with the original. Instead of the narrowly constructed scoring ruleset used the original WAAVES program, CRG will develop an automated analysis code based on machine learning which will not only improve speed and accuracy but also broaden the capabilities of the USV scoring program to include more call types, additional animals including but not limited to mice, and support various recording environments. This will be accomplished by incorporating improved waveform analysis techniques; developing ab initio identification, grouping, and classification algorithms; and providing statistical feedback on the accuracy of the automatically scored calls.UT will provide input on the manual scoring rules needed for an expanded library of calls and provide feedback on the accuracy of the enhanced algorithm by comparing to hand scored data.