Dual-track Residential Exercise With AI and Monitoring for Sleep (DREAMS)

National Sun Yat-sen University

Status: Begins enrollment in 2 months

Conditions: Difficulty Maintaining Sleep; Healthy Aging and Independent Living; Chronic Insomnia Characterized; Difficulty Falling Asleep

Treatments: Device: Fit Mirror-Guided Home-Based Exercise Program

Study type: Interventional

Funder types: Other

Identifiers

NCT07513584

11501EC021

Details and patient eligibility

About

As societies rapidly transition toward aging demographics, sleep issues among community-dwelling older adults have emerged as a critical concern affecting healthy aging and independent living. Current single-track exercise intervention models are often difficult to implement due to suboptimal adherence. Therefore, this study aims to utilize artificial intelligence technology combined with a dual-track residential exercise mode to improve sleep quality, thereby enhancing the self-care and independent living abilities of the elderly

Full description

The DREAMS Study addresses the critical public health challenge of chronic insomnia among community-dwelling older adults (aged ≥60), which substantially impacts healthy aging and independent living. Traditional exercise interventions often suffer from suboptimal adherence and rely on subjective self-reporting that fails to capture the physiological "mismatch" between perceived and actual sleep. To get around these problems, this study uses a home-based, closed-loop, dual-track exercise recommendation model that combines wearable ActiGraph monitoring with AI-driven skeletal recognition technology (iMirror). This adaptive framework differentiates between insomnia phenotypes: daytime moderate-intensity training (HIIT or resistance exercise) is prescribed to enhance sleep drive for those with difficulty falling asleep (DFA), while nighttime relaxation training (yoga or Pilates) targets reduced hyperarousal for those with difficulty maintaining sleep (DMS). By utilizing continuous objective data, the system creates a feedback loop that dynamically adjusts exercise prescriptions (frequency, intensity, and timing), reducing the need for on-site professional supervision and ensuring safe implementation within the participant's familiar home environment. Ultimately, the DREAMS Study establishes a scalable, data-driven model for precision health promotion to inform future policies on sleep health in aging populations.

Enrollment: 60 estimated patients

Sex: All

Ages: 60+ years old

Volunteers: No Healthy Volunteers

Inclusion criteria

• Age ≥ 60 years
• Capable of independent mobility (without the use of assistive devices).4
• Meeting one of the following sleep disturbance criteria:

Core symptoms of DSM-5 chronic insomnia (self-reported) for ≥ 3 months. Insomnia Severity Index (ISI) ≥ 15 (moderate-to-severe insomnia). Pittsburgh Sleep Quality Index (PSQI) > 5 (poor sleep quality).

• Basic ability to use a tablet or smartphone (caregivers may assist with login, but exercise must be performed by the participant).
• Mini-Cog score ≥ 3.
• Consent to wear wearable devices and participate in data collection.

Exclusion criteria

• Major cardiovascular events within the past 3 months (e.g., acute myocardial infarction, unstable angina), severe heart failure, or uncontrolled hypertension (e.g., SBP ≥ 180 or DBP ≥ 110 mmHg).
• Severe osteoarticular or neuromuscular diseases that prevent the safe completion of exercise (e.g., recent hip fracture, severe Parkinsonian imbalance).
• Severe psychiatric disorders or substance use disorders that may affect adherence.
• Untreated moderate-to-severe obstructive sleep apnea (OSA) with extreme daytime sleepiness (the study will use objective measurements for preliminary screening).
• Currently receiving structured psychotherapy for insomnia (e.g., CBT-I or BBTi) and not yet stabilized.
• Severe visual or hearing impairment that prevents following voice or visual instructions.

Trial design

Primary purpose: Supportive Care

Allocation: N/A

Interventional model: Single Group Assignment

Masking: None (Open label)

---

60 participants in 1 patient group

AI-Driven Dual-Track Residential Exercise and Physiological Monitoring Arm

Experimental group

Description:

Participants (≥ 60) with chronic insomnia undergo a 12-week home exercise intervention followed by a 12-week follow-up. The system integrates FitMirror AI skeletal recognition for real-time guidance and ActiGraph wearable monitoring for continuous data collection. An adaptive closed-loop mechanism tailors exercise to specific insomnia phenotypes: daytime HIIT or resistance training is prescribed to enhance sleep drive for those with sleep-onset difficulties (DFA), while nighttime yoga or Pilates targets reduced hyperarousal for those with maintenance difficulties (DMS). Effectiveness is evaluated at baseline, 12 weeks, and 24 weeks using linear mixed-effects models to track improvements in multi-dimensional sleep health and functional fitness.

Treatment:

Device: Fit Mirror-Guided Home-Based Exercise Program