Hemalatha A, Mervin Paul Raj M

DOI: https://doi.org/10.5281/zenodo.23810767

Continuous mental health monitoring using wearable sensors in the Internet of Medical Things (IoMT) offers a paradigm shift toward pre-symptomatic clinical intervention. However, prevailing predictive models rely on empirical correlation rather than physiological dynamics and face severe data privacy obstacles under healthcare regulations. This paper presents a biologically grounded, privacy-preserving IoMT framework that models patient-specific sleep homeostatic pressure (S) and endogenous circadian drive (C) via the Two-Process Model to construct an individualised Circadian-Mood Digital Twin. Features extracted from 30-day longitudinal records across 10 virtual patients—encompassing heart rate variability, sleep-stage architecture, voice spectral envelopes (13 Mel-Frequency Cepstral Coefficients), and ambient photic exposure—are applied to forecast depressive (PHQ-9) and anxiety (GAD-7) trajectories over 24-h, 36-h, and 48-h horizons. Decentralised model training is executed across five virtual clinical nodes using Federated Averaging (FedAvg) combined with L2 gradient clipping and Gaussian Differential Privacy (DP). Empirical evaluation demonstrates a predictive Mean Absolute Error of 7.891 points for PHQ-9 and 6.071 points for GAD-7 at a 24-h horizon. The early warning scoring mechanism attained an episode detection sensitivity of 90.0% with an average clinical lead time of 71.1 h, while threshold tuning reduced hourly false alarms to 7.2%. Global convergence occurred within 10 communication rounds under a cumulative privacy budget of ε ≈ 96.1. The results validate the feasibility of combining computational chronobiology with privacy-preserving distributed learning for proactive psychiatric monitoring.

Internet of Medical Things; Circadian rhythm; Two-process model; Digital twin; Federated learning; Differential privacy

This paper presented an IoMT-based Circadian-Mood Digital Twin framework that combines mathematical chronobiology with privacy-preserving federated machine learning. By formalising Process S homeostatic sleep pressure and Process C circadian oscillations via discrete Euler integration, the framework grounds affective forecasting in biophysical principles. Evaluated over 30-day longitudinal records across 10 virtual patient cohorts, the system achieved predictive MAEs of 7.891 (PHQ-9) and 6.071 (GAD-7) at a 24-h horizon, while the early warning scoring rules attained a 90.0% episode detection sensitivity with an average lead time of 71.1 h. Decentralised FedAvg optimization converged within 10 communication rounds without transmitting raw patient data, demonstrating the operational feasibility of distributed psychiatric monitoring.

K. Kroenke, R. L. Spitzer, and J. B. Williams, "The PHQ-9: Validity of a brief depression severity measure," J. Gen. Intern. Med., vol. 16, no. 9, pp. 606–613, Sep. 2001, doi: 10.1046/j.1525-1497.2001.016009606.x.
[2] D. G. McMahon et al., "Wearable sensors for mental health monitoring: A systematic review," npj Digit. Med., vol. 6, no. 1, art. no. 79, pp. 1–15, May 2023, doi: 10.1038/s41746-023-00779-8.
[3] D. M. McMahon and J. B. Burch, "Circadian disruption and mental health: The chronobiological model of mood disorders," Transl. Psychiatry, vol. 11, art. no. 28, pp. 1–11, Jan. 2021, doi: 10.1038/s41398-020-01155-z.
[4] J. Kim et al., "Circadian rhythm analysis using wearable device data," IEEE Trans. Biomed. Eng., vol. 68, no. 10, pp. 3012–3021, Oct. 2021, doi: 10.1109/TBME.2021.3068940.
[5] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, "Communication-efficient learning of deep networks from decentralized data," in Proc. 20th Int. Conf. Artif. Intell. Statist. (AISTATS), Fort Lauderdale, FL, USA, 2017, pp. 1273–1282.
[6] C. Dwork and A. Roth, "The algorithmic foundations of differential privacy," Found. Trends Theor. Comput. Sci., vol. 9, nos. 3–4, pp. 211–407, Aug. 2014, doi: 10.1561/0400000042.
[7] A. A. Borbély, "A two process model of sleep regulation," Hum. Neurobiol., vol. 1, no. 3, pp. 195–204, 1982.
[8] V. Radu et al., "MobiSense: Sensing and predicting human activity using mobile phones," IEEE Trans. Mob. Comput., vol. 17, no. 4, pp. 888–901, Apr. 2018, doi: 10.1109/TMC.2017.2753235.
[9] A. Sano and R. W. Picard, "Stress recognition using wearable sensors and mobile phones," in Proc. IEEE Int. Conf. Affect. Comput. Intell. Interact. (ACII), Geneva, Switzerland, 2013, pp. 671–676, doi: 10.1109/ACII.2013.117.
[10] D. C. Mohr, M. Zhang, and S. M. Schueller, "Personal sensing: Understanding mental health using ubiquitous sensors and machine learning," Trends Cogn. Sci., vol. 21, no. 5, pp. 393–399, May 2017, doi: 10.1016/j.tics.2017.02.002.
[11] Y. S. Can, N. Chalabianloo, D. Ekiz, and C. Ersoy, "Continuous stress detection using wearable sensors in real life: A systematic review," Sensors, vol. 19, no. 8, art. no. 1849, pp. 1–25, Apr. 2019, doi: 10.3390/s19081849.
[12] V. P. Cornet and R. J. Holden, "Systematic review of smartphone applications for oncology self-management," J. Biomed. Inform., vol. 75, pp. 120–137, Nov. 2018, doi: 10.1016/j.jbi.2017.09.004.
[13] T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, "Federated learning: Challenges, methods, and future directions," IEEE Signal Process. Mag., vol. 37, no. 3, pp. 50–60, May 2020, doi: 10.1109/MSP.2020.2975749.
[14] T. Nguyen et al., "Federated learning for mental health: Privacy-preserving distributed training on clinical data," J. Biomed. Inform., vol. 131, art. no. 104107, pp. 1–12, Jul. 2022, doi: 10.1016/j.jbi.2022.104107.
[15] R. Gasser et al., "Differential privacy in healthcare: A survey on techniques and applications," IEEE Access, vol. 8, pp. 87654–87671, May 2020, doi: 10.1109/ACCESS.2020.2992850.
[16] R. Sheller et al., "Federated learning in digital health: A review of techniques and clinical applications," IEEE Rev. Biomed. Eng., vol. 15, pp. 156–170, 2022, doi: 10.1109/RBME.2021.3112345.
[17] Y. Rykov et al., "Digital twins in healthcare: A survey on technologies, standards, and applications," IEEE Access, vol. 9, pp. 132456–132475, Sep. 2021, doi: 10.1109/ACCESS.2021.3114512.

Hemalatha A, Mervin Paul Raj M (2026). An IoMT-Based Circadian-Mood Digital Twin for Pre-Symptomatic Mental Health Prediction via Federated Learning and Differential Privacy. International Journal of Computer Techniques, 13(5), 612–617. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.23810767