Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Personalized entropy-informed deep learning for identifying opioid misuse

Nature Mental Health volume 4, pages 112–124 (2026) Cite this article

Subjects

Abstract

Fluctuations in pain, stress and craving are thought to contribute to opioid misuse. Developing accurate prediction models is vital for intervention and prevention efforts. In this work, we leverage physiological data and semantic analysis of electronic health records to tackle the challenge of detecting opioid misuse. Utilizing personalized hierarchical deep-learning models, we analyze trajectories of predicted pain, stress and craving states with 10,140 hours of heart-rate data collected by wearables from patients on long-term opioid therapy. From these trajectories, we extract entropy features from nonlinear dynamical analysis and develop a novel relevance-based temporal fusion model of opioid misuse risk. We incorporate clinical data into a large language model to enhance opioid misuse risk detection. We then fuse these modalities to achieve an accurate opioid misuse risk assessment with area under the precision-recall curve of 0.94 ± 0.05. This study marks a substantial advancement in personalized prediction of addictive behavior by elucidating the entropic nature of underlying affective state dynamics.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to the full article PDF.

USD 39.95

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Overview of the entire pipeline.
The alternative text for this image may have been generated using AI.
Fig. 2: Visualization of test-time quality.
The alternative text for this image may have been generated using AI.
Fig. 3: Visualization of learned features and the contributions of input components.
The alternative text for this image may have been generated using AI.
Fig. 4: Three-dimensional visualization of Takens’ embedding from the trajectories of affective states.
The alternative text for this image may have been generated using AI.

Similar content being viewed by others

Data availability

The data are available upon request with a signed data access agreement. For the detailed permission to access the data, please contact egarland@health.ucsd.edu.

Code availability

This software is © 2025 The Regents of the University of California. It is made available for educational and research use; for permission to use it in such contexts, please contact the corresponding author. For commercial use or licensing inquiries, please contact the Office of Innovation and Commercialization at innovation@ucsd.edu.

References

  1. Ciotti, M. et al. The COVID-19 pandemic. Crit. Rev. Clin. Lab. Sci. 57, 365–388 (2020).

    Article  PubMed  Google Scholar 

  2. Drug overdose deaths: facts and figures. NIDA https://nida.nih.gov/research-topics/trends-statistics/overdose-death-rates (2023).

  3. Mann, B. & Pattani, A. In 2023 fentanyl overdoses ravaged the US and fueled a new culture war fight. NPR https://www.npr.org/2023/12/28/1220881380/overdose-fentanyl-drugs-addiction (2023).

  4. Garland, E. L., Froeliger, B., Zeidan, F., Partin, K. & Howard, M. O. The downward spiral of chronic pain, prescription opioid misuse, and addiction: cognitive, affective, and neuropsychopharmacologic pathways. Neurosci. Biobehav. Rev. 37, 2597–2607 (2013).

    Article  PubMed  PubMed Central  Google Scholar 

  5. Lazarus, R. S. Stress and Emotion: A New Synthesis (Springer, 2006).

  6. Price, D. D. Psychological and neural mechanisms of the affective dimension of pain. Science 288, 1769–1772 (2000).

    Article  PubMed  Google Scholar 

  7. Giuliani, N. R. & Berkman, E. T. Craving is an affective state and its regulation can be understood in terms of the extended process model of emotion regulation. Psychol. Inq. 26, 48–53 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  8. Kuppens, P. It’s about time: a special section on affect dynamics. Emot. Rev. 7, 297–300 (2015).

    Article  Google Scholar 

  9. Vanhasbroeck, N., Ariens, S., Tuerlinckx, F. & Loossens, T. Computational Models for Affect Dynamics (Springer, 2021).

  10. Schupp, H. T., Junghöfer, M., Weike, A. I. & Hamm, A. O. The selective processing of briefly presented affective pictures: an ERP analysis. Psychophysiology 41, 441–449 (2004).

    Article  PubMed  Google Scholar 

  11. Buu, A. et al. The association between short-term emotion dynamics and cigarette dependence: a comprehensive examination of dynamic measures. Drug Alcohol Depend. 218, 108341 (2021).

    Article  PubMed  Google Scholar 

  12. Freeman, W. J. Societies of Brains: A Study in the Neuroscience of Love and Hate (Psychology Press, 2014).

  13. Haken, H. Synergetics of brain function. Int. J. Psychophysiol. 60, 110–124 (2006).

    Article  PubMed  Google Scholar 

  14. Apkarian, A. V., Baliki, M. N. & Geha, P. Y. Towards a theory of chronic pain. Prog. Neurobiol. 87, 81–97 (2009).

    Article  PubMed  Google Scholar 

  15. McEwen, B. S. Allostasis and Allostatic Load: Implications for Neuropsychopharmacology (Routledge, 2013).

  16. Thayer, J. F. & Lane, R. D. Claude Bernard and the heart–brain connection: further elaboration of a model of neurovisceral integration. Neurosci. Biobehav. Rev. 33, 81–88 (2009).

    Article  PubMed  Google Scholar 

  17. Thayer, J. F. & Lane, R. D. A model of neurovisceral integration in emotion regulation and dysregulation. J. Affect. Disord. 61, 201–216 (2000).

    Article  PubMed  Google Scholar 

  18. Kim, H.-G., Cheon, E.-J., Bai, D.-S., Lee, Y. H. & Koo, B.-H. Stress and heart rate variability: a meta-analysis and review of the literature. Psychiatry Investig. 15, 235–245 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  19. Alinia, P. et al. Associations between physiological signals captured using wearable sensors and self-reported outcomes among adults in alcohol use disorder recovery: development and usability study. JMIR Form. Res. 5, 27891 (2021).

    Article  Google Scholar 

  20. Gullapalli, B. T. et al. On-body sensing of cocaine craving, euphoria and drug-seeking behavior using cardiac and respiratory signals. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 3, 46 (2019).

    Article  Google Scholar 

  21. Bertz, J. W., Epstein, D. H. & Preston, K. L. Combining ecological momentary assessment with objective, ambulatory measures of behavior and physiology in substance-use research. Addict. Behav. 83, 5–17 (2018).

    Article  PubMed  Google Scholar 

  22. Määttänen, I. et al. Positive affect state is a good predictor of movement and stress: combining data from ESM/EMA, mobile HRV measurements and trait questionnaires. Heliyon 7, e06243 (2021).

  23. Murray, D. W., Ridenour, T. A., Swingler, M. M., Morgan, A. & Hegarty-Craver, M. Feasibility of combining biosensor and ecological momentary assessment to measure stress experiences among economically disadvantaged adolescents. Stress Health 39, 684–689 (2023).

    Article  PubMed  Google Scholar 

  24. Jiang, M. et al. Ultra-short-term analysis of heart rate variability for real-time acute pain monitoring with wearable electronics. In Proc. 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 1025–1032 (IEEE, 2017).

  25. Kalkman, G. A. et al. Monitoring opioids in Europe: the need for shared definitions and measuring drivers of opioid use and related harms. Eur. Addict. Res. 28, 231–240 (2022).

    Article  PubMed  Google Scholar 

  26. Gudin, J. A., Mogali, S., Jones, J. D. & Comer, S. D. Risks, management, and monitoring of combination opioid, benzodiazepines, and/or alcohol use. Postgrad. Med. 125, 115–130 (2013).

    Article  PubMed  PubMed Central  Google Scholar 

  27. Gullapalli, B. T. et al. OpiTrack: a wearable-based clinical opioid use tracker with temporal convolutional attention networks. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 5, 102 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  28. Benyamin, R. et al. Opioid complications and side effects. Pain. Physician 11, S105–S120 (2008).

    Article  PubMed  Google Scholar 

  29. Tejeda, H. A. & Bonci, A. Dynorphin/kappa-opioid receptor control of dopamine dynamics: implications for negative affective states and psychiatric disorders. Brain Res. 1713, 91–101 (2019).

    Article  PubMed  Google Scholar 

  30. Luo, Y. et al. Dynamic clustering via branched deep learning enhances personalization of stress prediction from mobile sensor data. Sci. Rep. 14, 6631 (2024).

    Article  PubMed  PubMed Central  Google Scholar 

  31. Stark, J., Broomhead, D. S., Davies, M. E. & Huke, J. Takens embedding theorems for forced and stochastic systems. Nonlinear Anal. Theory Methods Appl. 30, 5303–5314 (1997).

    Article  Google Scholar 

  32. Yan, Y. et al. Topological nonlinear analysis of dynamical systems in wearable sensor-based human physical activity inference. IEEE Trans. Hum. Mach. Syst. 53, 792–801 (2023).

    Article  Google Scholar 

  33. Wolf, A., Swift, J. B., Swinney, H. L. & Vastano, J. A. Determining lyapunov exponents from a time series. Phys. D Nonlinear Phenom. 16, 285–317 (1985).

    Article  Google Scholar 

  34. Hu, K., Ivanov, P. C., Chen, Z., Carpena, P. & Stanley, H. E. Effect of trends on detrended fluctuation analysis. Phys. Rev. E 64, 011114 (2001).

    Article  Google Scholar 

  35. Qian, B. & Rasheed, K. Hurst exponent and financial market predictability. In Proc. IASTED Conference on Financial Engineering and Applications 203–209 (ACTA Press, 2004).

  36. Garland, E. L., Froeliger, B. & Howard, M. O. Allostatic dysregulation of natural reward processing in prescription opioid misuse: autonomic and attentional evidence. Biol. Psychol. 105, 124–129 (2015).

    Article  PubMed  Google Scholar 

  37. Garland, E. L., Bryan, C. J., Nakamura, Y., Froeliger, B. & Howard, M. O. Deficits in autonomic indices of emotion regulation and reward processing associated with prescription opioid use and misuse. Psychopharmacology 234, 621–629 (2017).

    Article  PubMed  Google Scholar 

  38. Garland, E. L. & Howard, M. O. Prescription opioid misusers exhibit blunted parasympathetic regulation during inhibitory control challenge. Psychopharmacology 238, 765–774 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  39. Roberts, R. L. & Garland, E. L. Association between opioid use disorder and blunted heart rate variability among opioid-treated chronic pain patients. Addict. Biol. 27, 13230 (2022).

    Article  Google Scholar 

  40. Koob, G. F. & Le Moal, M. Drug addiction, dysregulation of reward, and allostasis. Neuropsychopharmacology 24, 97–129 (2001).

    Article  PubMed  Google Scholar 

  41. Shurman, J., Koob, G. F. & Gutstein, H. B. Opioids, pain, the brain, and hyperkatifeia: a framework for the rational use of opioids for pain. Pain. Med. 11, 1092–1098 (2010).

    Article  PubMed  Google Scholar 

  42. Elman, I. & Borsook, D. Common brain mechanisms of chronic pain and addiction. Neuron 89, 11–36 (2016).

    Article  PubMed  Google Scholar 

  43. Friedman, B. H. An autonomic flexibility–neurovisceral integration model of anxiety and cardiac vagal tone. Biol. Psychol. 74, 185–199 (2007).

    Article  PubMed  Google Scholar 

  44. Ottaviani, C., Medea, B., Lonigro, A., Tarvainen, M. & Couyoumdjian, A. Cognitive rigidity is mirrored by autonomic inflexibility in daily life perseverative cognition. Biol. Psychol. 107, 24–30 (2015).

    Article  PubMed  Google Scholar 

  45. Eddie, D., Price, J. L., Bates, M. E. & Buckman, J. F. Substance use and addiction affect more than the brain: the promise of neurocardiac interventions. Curr. Addict. Rep. 8, 431–439 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  46. Zhu, Z. & Bonanno, G. A. Affective flexibility: relations to expressive flexibility, feedback, and depression. Clin. Psychol. Sci. 5, 930–942 (2017).

    Article  Google Scholar 

  47. De Guzman, K. R., Snoswell, C. L., Taylor, M. L., Gray, L. C. & Caffery, L. J. Economic evaluations of remote patient monitoring for chronic disease: a systematic review. Value Health 25, 897–913 (2022).

    Article  PubMed  Google Scholar 

  48. El-Rashidy, N., El-Sappagh, S., Islam, S. R., M. El-Bakry, H. & Abdelrazek, S. Mobile health in remote patient monitoring for chronic diseases: principles, trends, and challenges. Diagnostics 11, 607 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  49. Thompson, J. M. T., Stewart, H. B. & Turner, R. Nonlinear Dynamics and Chaos (Wiley, 1990)

  50. Zhao, W. X. et al. A Survey of Large Language Models. Preprint at https://arxiv.org/abs/2303.18223 (2023).

  51. Huang, J. & Chang, K.C.-C. in Findings of the Association for Computational Linguistics: ACL 2023 1049–1065 (Association for Computational Linguistics, 2023).

  52. Wei, J. et al. Chain-of-thought prompting elicits reasoning in large language models. In Proc. 35th International Conference on Neural Information Processing Systems 24824–24837 (Curran Associates, 2022).

  53. Wang, B. et al. Towards understanding chain-of-thought prompting: an empirical study of what matters. In Proc. 61st Annual Meeting of the Association for Computational Linguistics (eds Rogers, A. et al.) 2717–2739 (ACL, 2023); https://doi.org/10.18653/v1/2023.acl-long.153

  54. Luo, Y., Liu, Y., Cai, R. & Rahman, T. Start simple: progressive difficulty multitask learning. In Proc. 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 4: Student Research Workshop) 48–55 (Association for Computational Linguistics, 2024).

  55. Garland, E. L. et al. Zoom-based mindfulness-oriented recovery enhancement plus just-in-time mindfulness practice triggered by wearable sensors for opioid craving and chronic pain. Mindfulness 14, 1329–1345 (2023).

    Article  Google Scholar 

  56. Myllymaki, T. Stress and Recovery Analysis Method Based on 24-Hour Heart Rate Variability (Firstbeat Technologies Ltd, 2014).

  57. Charlton, P. H. et al. An assessment of algorithms to estimate respiratory rate from the electrocardiogram and photoplethysmogram. Physiol. Meas. 37, 610–626 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  58. Föhr, T. et al. Subjective stress, objective heart rate variability-based stress, and recovery on workdays among overweight and psychologically distressed individuals: a cross-sectional study. J. Occup. Med. Toxicol. 10, 39 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  59. Madison, A., Vasey, M., Emery, C. F. & Kiecolt-Glaser, J. K. Social anxiety symptoms, heart rate variability, and vocal emotion recognition in women: evidence for parasympathetically-mediated positivity bias. Anxiety Stress Coping 34, 243–257 (2021).

    Article  PubMed  Google Scholar 

  60. Myllymäki, T. et al. Effects of vigorous late-night exercise on sleep quality and cardiac autonomic activity. J. Sleep. Res. 20, 146–153 (2011).

    Article  PubMed  Google Scholar 

  61. Palesh, O. et al. Secondary outcomes of a behavioral sleep intervention: a randomized clinical trial. Health Psychol. 38, 196–205 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  62. Rusko, H. et al. Stress and relaxation during sleep and awake time, and their associations with free salivary cortisol after awakening. In Proc. Nordic Ergonomics Society Congress (Nordic Ergonomics and Human Factor Society, 2006).

  63. Teisala, T. et al. Associations of physical activity, fitness, and body composition with heart rate variability–based indicators of stress and recovery on workdays: a cross-sectional study. J. Occup. Med. Toxicol. 9, 16 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  64. Uusitalo, A. et al. Heart rate variability related to effort at work. Appl. Ergon. 42, 830–838 (2011).

    Article  PubMed  Google Scholar 

  65. Butler, S. F. et al. Development and validation of the current opioid misuse measure. Pain 130, 144–156 (2007).

    Article  PubMed  PubMed Central  Google Scholar 

  66. Tobin, R. M., Graziano, W. G., Vanman, E. J. & Tassinary, L. G. Personality, emotional experience, and efforts to control emotions. J. Personal. Soc. Psychol. 79, 656–669 (2000).

    Article  Google Scholar 

  67. Hu, Y. et al. A cross-space CNN with customized characteristics for motor imagery EEG classification. IEEE Trans. Neural Syst. Rehabil. Eng. 31, 1554–1565 (2023).

    Article  PubMed  Google Scholar 

  68. Bansal, S., Gowda, K. & Kumar, N. Multilingual personalized hashtag recommendation for low resource Indic languages using graph-based deep neural network. Expert Syst. Appl. 236, 121188 (2024).

    Article  Google Scholar 

  69. Jiang, M. et al. Personalized and adaptive neural networks for pain detection from multi-modal physiological features. Expert Syst. Appl. 235, 121082 (2024).

    Article  Google Scholar 

  70. Chang, J. et al. PEPNet: parameter and embedding personalized network for infusing with personalized prior information. In Proc. 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining 3795–3804 (Association for Computing Machinery, 2023); https://doi.org/10.1145/3580305.3599884

  71. Guo, P., Lee, C.-Y. & Ulbricht, D. Learning to branch for multi-task learning. In Proc. 37th International Conference on Machine Learning 3854–3863 (PMLR, 2020).

  72. Jang, E., Gu, S. & Poole, B. Categorical reparameterization with Gumbel-Softmax. Preprint at https://arxiv.org/abs/1611.01144 (2017).

  73. Park, J. S. et al. Generative agents: interactive simulacra of human behavior. In Proc. of the 36th Annual ACM Symposium on User Interface Software and Technology (Association for Computing Machinery, 2023).

  74. Wang, G. et al. Optimized glycemic control of type 2 diabetes with reinforcement learning: a proof-of-concept trial. Nat. Med. 29, 2633–2642 (2023).

    Article  PubMed  PubMed Central  Google Scholar 

  75. Muzammil, M. Finetuning EnDevSols/tinyllama-2.5T-Clinical model on Clinical Dataset, Hugging Face (accessed 19 March 2023); https://huggingface.co/muzammil-eds/tinyllama-2.5T-Clinical-v2

  76. Wang, J. et al. Prompt engineering for healthcare: methodologies and applications. Preprint at https://arxiv.org/abs/2304.14670 (2023).

Download references

Acknowledgements

We thank the Supercomputing Center and Halıcıoğlu Data Science Institute at University of California San Diego for providing the working space and the computational resources necessary to conduct all experiments. E.L.G. was supported by R01DA058621 (PI: Garland), R01AT011772 (PI: Garland), R01DA056537 (PI: Garland), and R01DA057631 (PI: Garland) from the National Institutes of Health during the preparation of this paper. Opinions, interpretations, conclusions and recommendations are those of the authors and are not necessarily endorsed by the National Institutes of Health. T.R., Y.L. and B.T.G. were supported by NSF Smart and Connected Health (2320678), Google Research Scholar Award (Gift) and Optum Labs Award (UCSD Proposal Number 30223298; KR 46245).

Author information

Author notes

  1. These authors jointly supervised this work: Eric L. Garland, Tauhidur Rahman.

Authors and Affiliations

  1. Halıcıoğlu Data Science Institute, University of California, San Diego, CA, USA

    Yunfei Luo, Bhanu Teja Gullapalli & Tauhidur Rahman

  2. Manning College of Information and Computer Sciences, University of Massachusetts, Amherst, MA, USA

    Iman Deznabi & Madalina Brostean Fiterau

  3. Ilumivu, Asheville, NC, USA

    Mark Tuomenoksa

  4. Sanford Institute for Empathy and Compassion, University of California, San Diego, CA, USA

    Eric L. Garland

  5. Department of Psychiatry, School of Medicine, University of California, San Diego, CA, USA

    Eric L. Garland

Authors
  1. Yunfei Luo
    View author publications

    Search author on:PubMed Google Scholar

  2. Iman Deznabi
    View author publications

    Search author on:PubMed Google Scholar

  3. Bhanu Teja Gullapalli
    View author publications

    Search author on:PubMed Google Scholar

  4. Mark Tuomenoksa
    View author publications

    Search author on:PubMed Google Scholar

  5. Madalina Brostean Fiterau
    View author publications

    Search author on:PubMed Google Scholar

  6. Eric L. Garland
    View author publications

    Search author on:PubMed Google Scholar

  7. Tauhidur Rahman
    View author publications

    Search author on:PubMed Google Scholar

Contributions

E.L.G. and T.R. as co-principal investigators conceived the original study and signal processing pipeline, and acquired funding and ran the data collection and analysis of the study. M.T. as the data engineer contributed to the preliminary data processing and analysis. M.B.F. as the senior peer contributed to the guidance of notions and methodologies in the clinical and machine-learning fields. Y.L., I.D. and B.T.G. led the research, data analysis, data visualization and machine-learning model development.

Corresponding author

Correspondence to Tauhidur Rahman.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Mental Health thanks Maryam Khodaverdi and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Supplementary Appendices A–G, figures and tables.

Reporting Summary (download PDF )

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Luo, Y., Deznabi, I., Gullapalli, B.T. et al. Personalized entropy-informed deep learning for identifying opioid misuse. Nat. Mental Health 4, 112–124 (2026). https://doi.org/10.1038/s44220-025-00555-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • Issue date:

  • DOI: https://doi.org/10.1038/s44220-025-00555-8

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing