Personal Meaning and Experimental Medicine in Autogenic mHealth:
What Works for Me, Why, and When?

Autogenic mHealth offers opportunities for individuals to work with intelligent systems to monitor, understand, and manage their own health. Realizing this potential requires connecting personal data with personal meaning and supporting systematic experimentation with behavior change strategies. This presentation considers how experimental medicine and the operating conditions framework can inform participant-driven health management. First, interpreting sensor and self-report data requires understanding the purposes, experiences, and contexts that give these observations meaning for an individual. Second, a personal experimental medicine approach involves identifying and measuring candidate mechanisms of change, selecting or developing techniques to influence those mechanisms, and testing whether changes in mechanisms lead to changes in personally meaningful outcomes. Evaluating these outcomes requires attention to personal baselines, usual fluctuations, and meaningful magnitudes of change. Third, a personal operating conditions framework distinguishes circumstances that influence whether a strategy engages its intended mechanism from circumstances that influence whether changing that mechanism changes behavior. Together, these approaches provide a foundation for autogenic mHealth systems that help individuals investigate what works for them, why it works, when it works, and when to try something different.

TBA

Tools That Serve Us and Are Ours to Correct: A Design Fiction and Backcast for Person- and Community-Centered Autogenic mHealth

Autogenic mHealth aims to shift health management from expert-driven interventions toward person- and community-directed approaches, but realizing this vision requires more than technological advances in sensing, adaptive algorithms, and large language models. This talk presents a design fiction set in City Heights, San Diego, in 2040, using insights from research on adaptive physical activity support, LLM-supported coaching, and community co-design to explore what it means to build intelligent health systems that serve people and remain open to their correction. Through a backcasting approach, the presenters identify six design requirements: goals defined by the people and communities being served; representations that remain faithful and correctable; clear distinctions between inferences and facts; personal information that remains private unless intentionally shared; attention to costs imposed on others without reducing them to scores or judgments; and sustained relationships over time. The talk connects these principles to existing evidence, identifies gaps in supporting health management at community and regional levels, and discusses opportunities to build on the YourMove dataset and community partnerships to advance person- and community-centered Autogenic mHealth.

Autogenic interventions for attention deficits and neurodevelopmental applications?

A significant need exists for non-pharmaceutical interventions for attention deficits in kids, where challenges to typical interventions include heterogeneity, as well as tolerance, adherence and resources. In this talk I will introduce efforts using wearable and non-wearable sensing in natural environments to address the challenge of heterogeneity and will extend the discussion to the potential role of autogenic interventions in the next step of this program. Special emphasis will be placed on defining unique constraints – including identification of causal mechanisms and target outcomes – in the domain of mental health and neurodevelopmental populations, that make this domain both highly challenging but also a distinct opportunity for interactive, within-individual, dynamic intervention approaches.

Time Series Foundation Models in mHealth Applications

Mobile health technologies generate rich, longitudinal streams of physiological, behavioral, and contextual data, offering unprecedented opportunities to understand and support health in everyday life. Yet translating these complex, heterogeneous signals into actionable, personalized insights remains a fundamental challenge. Time series foundation models offer a promising approach by learning generalizable representations from large-scale sensor data that can transfer across individuals, sensing modalities, and downstream health tasks. In this talk, I will discuss emerging advances in time series foundation models and their potential to transform mHealth.

Autogenic mHealth: Challenges and Opportunities

Over the past fifteen years, mHealth research has evolved from developing health sensing devices and predictive models to optimizing contextually tailored adaptive interventions, including JITAIs. JITAIs aim to determine the optimal type, amount, and timing of support by adapting to an individual’s changing internal and contextual state. However, current methodology remains largely focused on a process where behavioral science experts pre-specify a set of intervention options grounded in behavioral theory. Algorithms then tailor the selection of options to observed contexts. Recent advances in AI are poised to disrupt the current JITAI framework in the behavioral health space by enabling individuals to engage in self-directed, AI agent-mediated experimentation and intervention. This talk will provide an overview of the opportunities and challenges represented by this new paradigm, which we refer to as Autogenic mHealth.

Beyond Episodic Care: Sensor-Triggered, Self-Directed Offloading to Protect the Diabetic Foot Between Clinic Visits

Diabetic foot ulcers are a major cause of morbidity and mortality, yet successful healing depends on consistent use of offloading devices to relieve pressure on the affected foot. Neuropathy prevents patients from feeling the harm associated with removing these devices, while cognitive decline and daily demands can undermine adherence. Because clinic visits are often weeks apart, lapses in offloading may go undetected until wounds deteriorate. This talk presents a sensor-enabled approach to diabetic foot care that combines wearable monitoring, real-time feedback, and context-triggered conversational coaching to support adherence between clinical visits. In a five-site, three-arm randomized trial involving 210 patients, a sensor-instrumented removable cast walker with real-time smartwatch feedback achieved 58% healing at 12 weeks, compared with 49% for both a standard removable cast walker and one supplemented with staff-delivered education. In a subsequent study, context-triggered prompts initiated when sensors detected boot removal elicited responses within 60 minutes in 71-77% of cases, compared with 31-41% for scheduled prompts. In a 10-patient subsample, context-triggered prompts changed behavior, and all wounds healed within 12 weeks, compared with 58% healing in the full smart boot cohort. These findings highlight the potential of sensor-triggered, patient-directed interventions to identify and address high-risk behaviors as they occur, extending care beyond episodic clinical encounters. The talk will discuss design principles for autogenic foot care, including context-aware feedback, patient and family engagement, and criteria for escalating concerns to clinicians.

TBA

From Sleep Coach to Self-Coach: Toward Autogenic mHealth for Sleep in Dementia Family Caregivers

Family caregivers of persons living with dementia often sleep poorly, partly because of the nighttime behaviors of the person they care for. Prior and ongoing clinical trials have tested a behavioral sleep intervention that uses key strategies drawn from cognitive behavioral therapy for insomnia: stimulus control, sleep compression, relaxation, and sleep hygiene. Trained sleep coaches help caregivers build the knowledge and skills to improve both their own sleep and their loved ones’ sleep. However, this expert-driven model takes a lot of staff time and is a burden for caregivers who have little time to spare. That limits real-world dissemination, and it is unknown whether the benefits last after coaching ends. Drawing on lessons from these trials, this talk makes the case for participant-driven, autogenic mHealth for caregiver sleep. It describes how continuous self-monitoring through wearables and brief self-reports, paired with intelligent systems grounded in this behavioral sleep program, could help caregivers understand their own and their care recipients’ sleep patterns. Caregivers could then try and refine strategies and adapt as caregiving demands change, with human coaching saved for when it is most needed. Such an approach could make evidence-based sleep care more scalable and sustainable for this underserved population.

Beyond Symptom Tracking: Using mHealth to Understand and Address Fear Dysregulation in PTSD

Exposure to traumatic events such as natural disasters, assault, and sudden, catastrophic medical events is common, and posttraumatic stress disorder (PTSD) is the quintessential trauma-related mental disorder that can onset following these experiences. A key pathological feature of PTSD is dysregulated fear responding—particularly difficulty suppressing fear in safe environments. Indeed, PTSD has been conceptualized as a disorder of pathological fear learning, and PTSD is characterized by responses reflecting persistent excitation and exaggerated fear expression, poor inhibition of fear, and impaired safety signal processing. Not only is this heightened fear reactivity and related arousal characteristic of PTSD (across neural and autonomic markers), but these responses may underlie, in part, its adverse consequences for physical health, including heightened risk of developing cardiovascular disease. Identifying and monitoring dysregulated fear responding as a key intermediary phenotype of PTSD thus holds promise for promoting mental and physical health outcomes after trauma. mHealth technologies and methodologies have the potential to advance these efforts in several ways, from helping to shed light on underlying mechanisms, to tracking these processes in real-time in individuals’ natural environment, to designing and implementing interventions to offset risk in vulnerable individuals.