The Next Wave of Innovation: Silence Speaks Volumes in Clinical Research

---

In 2018, at a National Academy workshop, Stroud and colleagues presented a compelling case that digital biomarkers and the field of mobile assessment were at risk of “overpromising.”¹ They warned researchers and technology developers to avoid hype in favor of rigorous methods if they hoped to bring meaningful innovation to clinical research. In the six years that have followed, while some interesting and intriguing developments have surfaced in the use of AI-based methods and complex, multimodal measures of behavior, for the most part, these have been peripheral to drug development. It is fair to say that the fervor around digital approaches has not yet brought substantive change to measurement in neuroscience clinical trials, particularly in psychiatry.  

In 2024, however, several developments with the potential to alter the landscape have begun to surface, notably in mood disorders and psychosis. First, groundbreaking work published by Cohen et al. on the development and validation of vocal biomarkers as tools for enrichment in depression demonstrated a significant potential for the selection and characterization of patients.² The post-hoc analysis of a bipolar depression trial showed that speech latency, a simple, generalizable feature, can be reliably calculated across countries, languages, and cultures. Their measure of pauses in speech, specifically the delay in speech generation, was shown to improve signal detection by 50 to 100% when applied at screening.³ 

A follow-up study by Cohen, Kirkpatrick, and colleagues used data from Reviva Pharmaceuticals’ Phase III trial of brilaroxazine, a novel antipsychotic treatment.⁴ In that study, they demonstrated similar results using a modification of the same speech latency measure, showing dramatic results not only on total pathology, but also on negative symptoms and functioning as well.  

The story of the last six years in digital biomarkers has been one of machine learning models and “black box” proprietary technologies. We submit that this approach has not yielded significant solutions to the enduring problems of neuroscience drug development. Simplicity, transparency, a return to the fundamentals of psychometrics, and old-fashioned methodological rigor will do what complexity and AI have not yet been able to do. Look out for novel, but low-burden approaches to the assessment of clinically relevant features of speech, motor function, and human behavior.   

In 2025, we predict that the wearables will come off, the apps will be uninstalled, and the fanfare will fade. The next wave of innovation in this space might be focused on things as simple and understated as silence, but the impact will be deafening.  

---

References:

1. Stroud C, Onnela JP, Manji H. “Harnessing digital technology to predict, diagnose, monitor, and develop treatments for brain disorders.” NPJ Digit Med. https://pmc.ncbi.nlm.nih.gov/articles/PMC6550158/.  
2. Siegel, J., Cohen, A., Szabo, S., Tomioka, S., Opler, M., Kirkpatrick, B., Hopkins, S. “Enrichment using speech latencies improves treatment effect size in a clinical trial of bipolar depression.” Journal of Psychiatry Research, Volume 340, 116105, ISSN 0165-1781. https://doi.org/10.1016/j.psychres.2024.116105. 
3. Cohen, A., Rodriguez, Z., Opler, M., Kirkpatrick, B., Milanovic, S., Piacentino, D., Szabo, S., Tomioka, S., Ogirala, A., Koblan. K., Siegel, J., Hopkins, S. “Evaluating speech latencies during structured psychiatric interviews as an automated objective measure of psychomotor slowing.” Journal of Psychiatry Research, Volume 340, 116104, ISSN 0165-1781. https://doi.org/10.1016/j.psychres.2024.116104. 
4. Cohen, A., Kirkpatrick, B., Opler, M., Tatsumi, K., Bhat, S., Bhat, L. “Enrichment Based on Speech Latency Enhances Treatment Effects in a Phase III Study of Brilaroxazine.” Poster Presented at the 2024 CNS Summit.   

Related Insight:

Ethics in Clinical Research

Under the Microscope: Biomarker and Diagnostic Tests as FDA-Regulated Devices

Whitepapers