A biopharma sponsor was bringing new therapeutics to market for neurobehavioral disorders. They gathered a large collection of audio and video recordings as part of a clinical study to develop a transdiagnostic aid for patients with depression and anxiety.

The challenge was to extract biologically meaningful measurements from the audio and video data that could predict patient reported outcome measures, resulting in candidate digital biomarkers for further analysis and validation.

Our solution offered configurable data processing workflows to transform and quality assure raw video and audio data into measurements that could be connected back to prior studies in the biomedical literature.

The resulting data highlighted a key quality factor that helped distinguish subjects that adhered to the protocol from those that did not.

We were also able to discover a variety of predictive measures for further analysis, which stratified subjects on multiple scales for improved recruitment and enrollment and accelerated assessment of therapeutic response in future clinical trials.