Brett Talbot, Co-Founder and Chief Clinical Officer at Videra Health (Orem, UT) discusses AI technologies that were originally developed for behavioral health are now showing strong potential in diagnosing and monitoring movement disorders.
Editor's note: TechBuzz welcomes Utah tech leaders to share insights about their technologies as solutions to persistent challenges affecting the community. Today, Brett Talbot, Ph.D., Co-founder and Chief Clinical Officer of Videra Health, discusses AI tools from behavioral health that are being successfully adapted to improve detection and monitoring of movement disorders like Parkinson's.
Orem, Utah - July 8, 2025
By Brett Talbot, Co-founder and CCO, Videra Health
The progression of artificial intelligence in healthcare has largely occurred in silos, with innovations developing within specific medical domains rather than across them. Yet some of the most promising advancements have emerged when we apply technological breakthroughs from one field to another.
This is precisely the case with AI screening technologies originally developed for behavioral health that now show remarkable potential for movement disorder assessment and monitoring.
The Parallel Challenges
Behavioral health conditions and movement disorders share a fundamental trait: they manifest through observable patterns that are difficult to quantify objectively using traditional clinical assessment methods. Both domains have historically relied on subjective rating scales, periodic clinical observations, and patient self-reporting — all valuable, but inherently limited by their episodic nature and potential variability.
Take Parkinson's disease assessment, which typically uses the Unified Parkinson's Disease Rating Scale (UPDRS) during infrequent clinical visits. Compare this to depression screening, which often uses the PHQ-9 questionnaire at similar intervals. Both approaches capture only brief windows of patient experience, potentially missing critical patterns that emerge between appointments.
AI's Initial Success in Behavioral Health
The breakthrough in behavioral health came when we began applying computer vision, natural language processing, and acoustic analysis to detect subtle patterns in patient expression, speech, and behavior that correlate with conditions like depression, anxiety, and cognitive impairment. By analyzing facial micro-expressions, voice modulation, linguistic patterns, and even response latency, AI systems can now detect indicators of mental health conditions with sensitivity that rivals or exceeds traditional screening methods.
Recent studies have demonstrated promising applications of AI-driven assessments in detecting depression among adults. A hybrid deep learning model combining textual and audio features achieved 98% accuracy for audio-based depression detection and 92% accuracy for text-based detection in adult participants. These results show that audio CNN performs well for depression detection.
The technology's ability to analyze multiple modalities of data — including speech patterns, linguistic features, and behavioral markers — provides a comprehensive approach to mental health screening that could augment traditional clinical assessments.
The Logical Extension to Movement Disorders
The leap to applying these technologies to movement disorders is both logical and scientifically sound. Conditions like Parkinson's disease, essential tremor, Huntington's disease, and various dystonias present with observable motor symptoms that AI systems are uniquely equipped to quantify.
Recent advances in computer vision and machine learning show promise for early Parkinson's disease detection. Voice biomarker analysis achieved 92% accuracy with SVM and 94% with Random Forest algorithms in differentiating patients from healthy individuals. For neuroimaging, SVM with non-linear kernel achieved 96.14% detection rates using SPECT data, while another study reported 97.86% accuracy. These AI approaches analyze multiple biomarkers simultaneously, significantly outperforming traditional clinical assessments during early disease phases.
Beyond Detection: Continuous Monitoring and Outcome Measurement
The true value of this cross-application extends beyond initial screening. The same AI systems that can detect behavioral health conditions can now be repurposed to continuously monitor movement disorder progression and treatment response.
Consider a patient with Parkinson's disease using a smartphone application that analyzes:
This continuous data collection creates a rich longitudinal profile far more detailed than periodic clinical assessments alone. For pharmaceutical companies developing new therapeutics, this represents an unprecedented opportunity to measure drug efficacy through objective, continuous data rather than relying solely on subjective patient reports or infrequent clinical measurements.
Implications for Clinical Trials and Drug Development
For pharmaceutical researchers and clinical trial sponsors, this technology offers several transformative advantages:
A recent phase 3 trial incorporated a smartphone application as an exploratory endpoint alongside traditional MDS-UPDRS scoring, with participants completing finger tapping, gait, and cognition tests (both at-home and in-clinic). In another example, Roche has developed a smartphone application incorporating motor tasks such as finger tapping that can differentiate individuals with Parkinson's disease from those without, and has included it as an exploratory endpoint in a phase 2 clinical trial. During a 6-month phase 1b clinical trial, participants completed six daily motor active tests (sustained phonation, rest tremor, postural tremor, finger-tapping, balance, and gait) via smartphone, demonstrating that smartphone-sensor technologies provide reliable, valid, clinically meaningful, and highly sensitive phenotypic data in Parkinson's disease.
Implementation Challenges and Ethical Considerations
Despite its promise, this cross-application approach faces several hurdles:
The Future Landscape
The convergence of behavioral health and movement disorder assessment technologies represents just the beginning of a broader trend toward unified digital biomarker platforms. As we develop more sophisticated AI systems, we'll likely see further cross-pollination between medical domains that have traditionally operated independently.
For pharmaceutical companies and clinical researchers, early adoption of these cross-application technologies offers significant competitive advantages in drug development efficiency, cost reduction, and the ability to demonstrate treatment efficacy through novel, more sensitive endpoints.
Conclusion
The cross-application of AI screening technology from behavioral health to movement disorders exemplifies how breaking down silos between medical specialties can accelerate innovation. By recognizing the fundamental similarities in how these seemingly different conditions can be measured and monitored, we open new pathways for detection, treatment, and understanding of complex neurological and psychiatric disorders.
For those involved in pharmaceutical development and clinical research, this emerging approach offers potentially transformative changes in how we conceptualize disease assessment and treatment response measurement. The companies and researchers who recognize and leverage these connections will be positioned at the forefront of the next wave of therapeutic innovation.
Brett Talbot, Co-founder and CCO of Videra Health, is a distinguished clinical psychologist, technology innovator, and revered figure in the behavioral health community. Prior to Videra Health, Talbot was the Chief Clinical Officer and Executive Director across several prestigious healthcare organizations. His pioneering efforts led to the creation of trailblazing video-based depression, anxiety and trauma clinical assessments.
See Brett's previous TechBuzz article here.