Chimamanda Udodi ’28
Each day, often without our awareness, our phones collect small clues about our lifestyle, including typing speed, sleep-wake patterns, phone use, daily routes, and social interactions. Experts in digital mental health have uncovered a new possibility for this data: the ability to identify subtle mood shifts before individuals become aware of them. This method, known as digital phenotyping, is reshaping mental health research by enabling earlier intervention, more personalized treatment planning, and continuous emotional tracking.
Digital phenotyping involves the continuous tracking of human behavior through data collected from phones and other personal devices, capturing the nuances of how people interact with technology on a daily basis (Zakai & Alharthi, 2025). Because smartphones are ubiquitous, these devices gather constant, naturalistic updates about real-world behavior,including information that clinicians rarely observe during traditional office visits. While a single smartphone can gather vast behavioral information each day, therapy sessions typically occur only a few times per month and are shorter in duration. This disparity raises a transformative question: could mental health assessment shift from rare check-ins to continuous reflections of everyday life as it unfolds?
Growing evidence suggests that changes in mental state are reflected in how individuals use their devices. Reduced physical mobility or increased phone usage may reflect declines in mood, often before individuals perceive significant changes in their mental state (Saeb et al., 2015). Saeb et al. (2015) established this link in a study of 40 participants, finding statistically significant negative correlations between depression and GPS-based biomarkers—specifically location variance (r = -0.58) and regularity in 24-hour movement rhythms (r = -0.63). Building on such digital phenotypes, a subsequent study used machine-learning models to analyze combinations of smartphone-derived behavioral signals, including screen status, internet connectivity, and activity patterns, demonstrating strong predictive performance for depressive states (Asare et al., 2021). Among individuals with bipolar disorder, accelerated speech, altered keystroke dynamics, or sudden bursts of activity can serve as early indicators of approaching manic or depressive episodes (Zulueta et al., 2018). Additionally, subtle habits, such as sluggish typing, constant phone-checking, or late-night browsing could point to underlying stress or inner turmoil.
One major goal of digital phenotyping is suicide prevention. Preliminary studies suggest that acute behavioral shifts—such as abrupt social withdrawal or unusual sleeping patterns—often precede psychiatric crises (Onnela & Rauch, 2016). By identifying these signals, clinicians may be able to act rapidly, even before individuals verbalize their distress. Instead of relying solely on the patient’s self-reporting, automated alerts could highlight concerning behavioral patterns, ensuring that support arrives at critical moments.
Moreover, digital phenotyping holds strong potential to make mental health care more personalized. Because people use their devices in different ways, digital behavior is inherently idiosyncratic. For some, sending fewer messages may signal depressive withdrawal, whereas for others it may simply reflect typical texting habits. By leveraging machine learning to establish an individual’s unique digital baseline, algorithms can detect individually relevant deviations rather than comparing to population-level trends (Torous et al., 2016). This approach aligns with the field’s movement toward precision psychiatry, in which clinical strategies are tailored to patients’ actual behavioral patterns, rather than traditional diagnostic labels.
Despite the broad potential of digital phenotyping, the ability to monitor behavior through personal devices introduces significant ethical considerations, particularly around data privacy. Smartphones capture deeply intimate information, including social circles, daily routines, and fluctuations in emotional well-being. As Martinez-Martin et al. (2021) emphasize, existing ethical and regulatory frameworks in mental healthcare do not clearly extend to digital phenotyping, leaving critical gaps in protecting sensitive information. In light of these concerns, it is essential that patients maintain data sovereignty, with transparent control over what information is collected and how it is shared.
Furthermore, there is a risk of over-reliance on automated decision-making. These systems are only as effective as the data used to train them; biased or noisy data can lead to clinical errors, particularly for underrepresented populations. Without careful oversight, these tools could inadvertently widen existing disparities in mental health support (Onnela & Rauch, 2016). Finally, the psychological impact of passive surveillance cannot be ignored, as the feeling of being constantly monitored may itself induce stress in the very individuals the technology is intended to support.
Despite these hurdles, digital tracking is transforming the landscape of mental healthcare more profoundly than many other contemporary tools. Its primary clinical utility lies in its ability to help providers identify emotional shifts early, customize care based on individual behavioral patterns, and intervene before a psychiatric breakdown occurs. Unlike expensive neuroimaging or genetic testing, this method utilizes the data of daily life—communication, mobility, and sleep—making it a highly scalable and cost-effective solution for global mental health (Zakai & Alharthi, 2025).
The central question is no longer whether digital phenotyping will influence the trajectory of psychiatry, but when it will be integrated into routine care. We may soon see a future where longitudinal digital data complements traditional clinical interviews during routine checkups. Furthermore, these tools could empower patients to monitor their own psychological trends with greater clarity. However, the pace of adoption will likely depend on how effectively privacy and ethical risks are addressed over the next decade.
At present, one reality is clear: personal devices already capture subtle patterns related to mood, often without individuals’ conscious awareness. In the near future, these same technologies may serve as a vital safety net—helping recognize warning signals early enough to enable timely interventions and prevent harm
Chimamanda Udodi is a staff writer at The Princeton Medical Review. She can be reached at cu8222@princeton.edu.
References
Faurholt-Jepsen, M., Bauer, M., & Kessing, L. V. (2018). Smartphone-based objective monitoring in bipolar disorder: status and considerations. International Journal of Bipolar Disorders, 6(1), 6. https://doi.org/10.1186/s40345-017-0110-8
Martinez-Martin, N., Greely, H. T., & Cho, M. K. (2021). Ethical development of digital phenotyping tools for mental health applications: Delphi study. JMIR mHealth and uHealth, 9(7), e27343. https://doi.org/10.2196/27343
Onnela, J. P., & Rauch, S. L. (2016). Harnessing Smartphone-Based Digital Phenotyping to Enhance Behavioral and Mental Health. Neuropsychopharmacology, 41(7), 1691–1696. https://doi.org/10.1038/npp.2016.7
Saeb, S., Zhang, M., Karr, C. J., Schueller, S. M., Corden, M. E., Kording, K. P., & Mohr, D. C. (2015). Mobile Phone Sensor Correlates of Depressive Symptom Severity in Daily-Life Behavior: An Exploratory Study. Journal of Medical Internet Research, 17(7), e175. https://doi.org/10.2196/jmir.4273
Zakai, J. G., & Alharthi, S. A. (2025). Harnessing Digital Phenotyping for Early Self-Detection of Psychological Distress. Healthcare, 13(16), 2008. https://doi.org/10.3390/healthcare13162008Zulueta, J., Piscitello, A., Rasic, M., Easter, R., Babu, P., Langenecker, S. A., McInnis, M., Ajilore, O., Nelson, P. C., Ryan, K., & Leow, A. (2018). Predicting Mood Disturbance Severity with Mobile Phone Keystroke Metadata: A BiAffect Digital Phenotyping Study. Journal of Medical Internet Research, 20(7), e241. https://doi.org/10.2196/jmir.9775
Leave a Reply