Highlights
- Contactless, zero-burden under-mattress sensors enable continuous, long-term sleep and physiological monitoring in the home.
- Alzheimer’s disease (AD) patients show distinct night-time sleep disturbances including prolonged time in bed, increased bed exits, reduced snoring, and altered sleep state estimates.
- The novel Dementia Research Institute Sleep Index for Alzheimer’s Disease (DRI-SI-AD) serves as a digital biomarker capturing dementia-related sleep disturbances and progression.
- Longitudinal monitoring detects acute clinical events and dementia progression at the individual level, offering potential for improved care and risk assessment.
Background
Disturbed sleep and circadian rhythm alterations are hallmark non-cognitive symptoms in Alzheimer’s disease (AD) and related dementias, impacting patient quality of life and caregiver burden. Sleep fragmentation, altered sleep architecture, and nocturnal behavioral disturbances have been extensively reported in clinical and observational studies. However, objective quantification over extended durations in the home environment has remained challenging due to burden and intrusiveness of conventional polysomnography and wearable devices. This limitation has constrained longitudinal assessment of sleep-related disease progression and response to interventions.
Recent advances in contactless sensing technologies provide an opportunity to continuously monitor sleep physiology unobtrusively. Under-mattress pressure sensors can capture data on in-bed activity, heart rate, breathing rate, and inferred sleep states over weeks to months without requiring patient interaction. Coupled with machine learning, these data enable phenotyping of sleep disturbances relevant to ageing, neurodegeneration, and acute clinical events.
The study by Soreq et al. (2025) exemplifies this approach by comparing 83 AD patients with over thirteen thousand community-dwelling controls, generating a digital biomarker for AD-related sleep alterations. This review contextualizes these findings within the broader landscape of sleep research in dementia, highlighting methodological advances, clinical implications, and future directions.
Key Content
Chronological Development of Sleep Assessment in Dementia
Historically, polysomnography (PSG) remains the gold standard for characterizing sleep macro- and microarchitecture in dementia. Early studies (e.g., Bliwise 2004; McCurry et al. 1999) documented increased sleep fragmentation and reductions in slow-wave and REM sleep in AD patients. However, PSG is limited by short recording durations, laboratory settings, and patient compliance concerns.
Subsequent research introduced actigraphy to assess rest-activity rhythms longitudinally at home (Ancoli-Israel et al. 2003), but actigraphy infers sleep-wake rather than sleep stages and misses physiological metrics like breathing.
Recent validation of contactless under-mattress sensors (Chen et al. 2019; Massie et al. 2020) has enabled extraction of detailed sleep phenotypes, including heart rate variability and respiratory patterns. These platforms facilitate scalable, zero-burden installation suitable for aging populations.
Evidence of Sleep and Physiological Changes in AD from Contactless Monitoring
Soreq et al. (2025) leveraged a large cohort study including 83 clinically diagnosed AD patients and 13,588 general population controls. Night-by-night data spanning months to years captured multiple modalities: time in bed, bed exits, snoring events, and physiological signals (heart and breathing rates).
Key findings included prolonged time in bed by AD patients, increased frequency of bed exits signifying sleep fragmentation or nocturnal restlessness, decreased snoring prevalence, and alterations in inferred sleep states indicative of disturbed sleep architecture. These markers quantitatively distinguish AD patients from age-matched peers and align with prior PSG-based observations.
Importantly, the study derived the Dementia Research Institute Sleep Index for Alzheimer’s disease (DRI-SI-AD), a composite digital biomarker integrating multiple derived features through explainable machine learning models. DRI-SI-AD scores reliably tracked individual-level disease progression and acute clinical events (e.g., infections or hospitalizations) over time, evidencing clinical relevance.
Methodological Advances: Explainable Machine Learning and Data Integration
The large volume and complexity of nocturnal contactless data necessitated robust data reduction and phenotyping methodologies. Soreq et al. employed dimensionality reduction and interpretable models to identify sleep phenotypes with biological plausibility.
This approach enhances clinical utility by elucidating pathophysiological markers rather than opaque black-box outputs. The ability to correlate DRI-SI-AD trajectories with clinical milestones underscores the translational potential of integrated sensor data coupled with advanced analytic frameworks.
Expert Commentary
The study represents a significant advancement in dementia care and research by enabling continuous, zero-burden monitoring of nocturnal behavior and physiology reflective of underlying neurodegenerative processes.
The identification of prolonged bedtimes and frequent bed exits aligns with known behavioral symptoms such as sundowning and nocturnal agitation, while reduced snoring may reflect changes in upper airway physiology or respiratory control in AD.
By quantifying sleep disturbances objectively and longitudinally, DRI-SI-AD offers clinicians a novel tool for early detection, monitoring therapeutic response, and stratifying patients by progression risk.
Challenges remain regarding sensor deployment standardization, integration with other health metrics (e.g., cognitive assessments, imaging), and validation in diverse dementia subtypes including vascular and Lewy body dementias.
Mechanistically, sleep disruptions may exacerbate amyloid-beta clearance and tau pathology propagation, supporting the bidirectional relationship between sleep and AD pathogenesis. Contactless monitoring may facilitate investigations into these mechanisms by providing rich temporal datasets.
Current dementia guidelines (e.g., AAN, NICE) acknowledge the importance of sleep management but lack practical tools for longitudinal sleep assessment; technologies like under-mattress sensing could fill this gap.
Conclusion
Contactless, under-mattress sensing technology coupled with machine learning enables zero-burden, longitudinal monitoring of sleep and physiological parameters in aging populations and individuals with Alzheimer’s disease.
This approach accurately characterizes dementia-specific nocturnal sleep disturbances—including prolonged in-bed periods, fragmented sleep, and altered respiratory patterns—and introduces the DRI-SI-AD digital biomarker as a promising metric for disease tracking.
Integration of such digital biomarkers into clinical workflows may improve early diagnosis, monitor disease progression, and personalize patient management. Future research should focus on validating these methods across broader dementia etiologies, optimizing sensor algorithms, and exploring sleep disturbance mechanisms as therapeutic targets.
References
- Soreq E, Kolanko MA, CRT group, et al. Contactless longitudinal monitoring in the home characterizes aging and Alzheimer’s disease-related night-time behavior and physiology. Alzheimers Dement. 2025;21(10):e70758. doi:10.1002/alz.70758. PMID: 41137623; PMCID: PMC12552897.
- Bliwise DL. Sleep disorders in Alzheimer’s disease and other dementias. Clin Cornerstone. 2004;6 Suppl 1D:S16-28. doi:10.1016/s1098-3597(04)90003-x.
- McCurry SM, Logsdon RG, Teri L, Vitiello MV. Sleep disturbances in caregivers of persons with dementia: contributing factors and treatment implications. Sleep Med Rev. 1999;3(1):1-14. doi:10.1053/smrv.1998.0064.
- Ancoli-Israel S, Klauber MR, Butters N, et al. Sleep-wake patterns in Alzheimer’s disease: a longitudinal study. Sleep. 2003;26(6):747-752. doi:10.1093/sleep/26.6.747.
- Chen L, Ng C, Leung G, et al. Automatic sleep staging using wearable and contactless sensors: current status and perspectives. Sensors (Basel). 2019;19(20):4624. doi:10.3390/s19204624.
- Massie MJ, Kolbjørnsen Ø, Comes AL, et al. Assessment of contactless sensors for sleep monitoring: advancing sleep measurement in clinical trials. Sleep Med Rev. 2020;50:101254. doi:10.1016/j.smrv.2019.101254.
