The relationship between behavioral and psychological symptoms of dementia and sleep patterns is dynamic and forms a vicious cycle, with bidirectional Granger causality observed between irritability and number of awakenings.
Key Findings
Results
Unidirectional Granger causality was found running from number of awakenings on the previous day to irritability on the following day.
P=.03 for the effect of previous day's number of awakenings on next-day irritability
This relationship was unidirectional, meaning irritability also fed back to awakenings (making the overall irritability-awakenings relationship bidirectional)
Analysis was conducted using a panel vector autoregressive model over a 14-day observation period
Sample consisted of 154 older adults with dementia
Results
Unidirectional Granger causality was found running from number of awakenings on the previous day to appetite or eating disorders on the following day.
P=.04 for the effect of previous day's number of awakenings on next-day appetite or eating disorders
This temporal effect ran from sleep disturbance to behavioral symptom
Data were collected from June 2018 to June 2020 in community and institutional settings
Sleep data were collected via wrist actigraphy worn continuously for 2 weeks
Results
Delusions on the previous day had Granger-causality effects on multiple subsequent sleep parameters.
Delusions had a Granger-causality effect on total sleep time (P<.001)
Delusions had a Granger-causality effect on wake after sleep onset (P=.01)
Delusions had a Granger-causality effect on number of awakenings (P=.006)
These findings indicate that a behavioral/psychological symptom temporally preceded changes in three distinct objective sleep metrics
Results
Irritability on the previous day had a Granger-causality effect on the number of awakenings the following night.
P=.007 for the effect of previous day's irritability on subsequent number of awakenings
Combined with the finding that awakenings also predicted next-day irritability (P=.03), this constitutes a bidirectional relationship
Caregivers recorded behavioral and psychological symptoms of dementia in a daily symptom diary alongside actigraphy data
Results
Bidirectional Granger causality was specifically observed between irritability and number of awakenings.
Number of awakenings predicted next-day irritability (P=.03) and irritability predicted next-night number of awakenings (P=.007)
This bidirectional relationship suggests a self-reinforcing cycle between this sleep parameter and this behavioral symptom
Authors describe this as a 'vicious cycle' between sleep disturbances and behavioral and psychological symptoms of dementia
Panel vector autoregressive modeling was used to capture these temporal dynamics in daily data
Methods
The study used wrist actigraphy and caregiver-recorded daily symptom diaries over 14 days to capture dynamic sleep-BPSD interactions in 154 older adults with dementia.
154 older adults with dementia participated, recruited from community and institutional settings
Data collection spanned June 2018 to June 2020
Participants wore wrist actigraphy devices continuously for 2 weeks to objectively measure sleep patterns
Caregivers recorded behavioral and psychological symptoms of dementia (BPSD) daily in a symptom diary
A panel vector autoregressive (PVAR) model was used to analyze temporal interdependencies
What This Means
This research suggests that sleep problems and behavioral symptoms in people with dementia are closely linked in a two-way, self-reinforcing cycle. The study tracked 154 older adults with dementia over two weeks using wrist-worn sleep monitors and daily caregiver diaries. By analyzing these day-by-day records, the researchers found that disrupted sleep — specifically waking up frequently during the night — predicted more irritability and eating problems the next day. In the other direction, delusions (false beliefs) and irritability one day were associated with worse sleep the following night, including less total sleep, more time awake after falling asleep, and more nighttime awakenings.
The most striking finding was a bidirectional loop specifically between nighttime awakenings and irritability: poor sleep led to more irritability, and more irritability led to worse sleep the next night. This cycle could compound over time, making both problems progressively worse. The study is notable for using objective sleep measurements and daily tracking rather than relying on single time-point assessments, which allowed the researchers to detect these temporal cause-and-effect patterns.
This research suggests that in dementia care, treating sleep disturbances and behavioral symptoms together — rather than in isolation — may be important for breaking this cycle. Caregivers and healthcare providers may benefit from monitoring both sleep quality and behavioral changes on a daily basis, as improvements in one area could potentially help the other. The findings also highlight the burden placed on caregivers, who tracked symptoms daily, and underscore the need for early interventions targeting both sleep and behavioral symptoms simultaneously.
Cho E, Hwang S, Yang M, Kim E, Cho J, Park C. (2026). Dynamic Relationship Between Sleep Patterns and Behavioral and Psychological Symptoms of Dementia: Longitudinal Observational Study.. JMIR aging. https://doi.org/10.2196/80422