Poster N18 - Contextualizing Sleep Movements

Introduction: Sleep is an integral part of life, one that many people struggle with, such as the 5-10 % of people who experience restless type sleep disorders such as restless leg syndrome (RLS) or periodic limb movement disorder (PLMD). Modern technology has provided tools which are capable of monitoring sleep. Actigraphy is one such technology which uses a wrist-mounted accelerometer to measure nighttime movement to predict sleep-wake transitions. This technology is particularly advantageous because it is relatively inexpensive, non-intrusive, and can be used at home, unlike standard sleep assessments which use the gold-standard of polysomnography (PSG). The American Academy of Sleep Medicine (AASM) recommends the use of Actigraphy for the diagnosis and monitoring of several sleep conditions; however, despite measuring movement directly, the AASM does not recommend the use of Actigraphy in the diagnosis of restless type disorders like PLMD. This is primarily due to actigraphy-type devices underestimating the movements of the lower limb. Traditional actigraphy devices contain a single accelerometer capable of capturing linear movements in a single direction, though more modern devices have moved towards inertial measurement unit (IMU) based devices which include triaxial accelerometers and triaxial gyroscopes. These additional sensors not only capture linear accelerations in three dimensions, but also capture angular velocities in three dimensions, which makes the device more sensitive to movements. Furthermore, standard actigraphy only uses a single sensor. This present work explores the use of multiple IMU-based actigraphy devices to understand how movements differ across limbs.

Materials and

Methods: One participant wore an Axivity AX6 IMU on each limb (right wrist and ankle, left wrist and ankle), accounting for a total of four IMUs which were worn simultaneously. The IMUs were configured for a collection rate of 100 Hz and accelerometer and gyroscope ranges of 16 G and 1000 degrees/second, respectively. The participant wore the Axivity IMUs during a typical night which resulted in approximately six hours of continuous sleep data collected from each limb. The data from each limb was evaluated and scored individually to determine instances when a movement occurred. Any instance in which the peak acceleration changed by 0.1 g along any axis and lasted at least 0.5 seconds was considered a movement. After each limb was scored individually, the timeframes in which movements occurred were compared across limbs to determine the amount of times movements were coordinated between two or more limbs.

Results, Conclusions, and Discussions: A total of 151 individual limb movements were captured by the IMU data while the participant was asleep. Of these 151 movements, 120 of these movements were coordinated across all four limbs, accounting for 29 movement events. An additional 6 movement events occurred in which movements were coordinated only between the upper limbs or only between the lower limbs, which accounts for an additional 12 individual limb movements. The remaining 19 individual limb movements only occurred in a single limb, though usually such movements proceeded coordinated limb movements. Approximately half of the uncoordinated movements occurred during a single five-minute window of particularly restless sleep. Interestingly, most uncoordinated movements occurred in the participant’s dominate arm and leg.
This preliminary analysis demonstrates that traditional actigraphy can be improved through the inclusion of additional sensors. The simple inclusion of actigraphy devices on all limbs allows for the technology to better capture restless movements during sleep. While further work and validation is needed, using an actigraphy device on each limb might allow actigraphy to be a viable option to enable restless-type sleep disorders to be evaluated at home and alleviate wait times for clinical PSG evaluations. This is particularly impactful for populations who are prone to restless type sleep disorders, such as children with autism spectrum disorder (ASD).
The present analysis is part of an exploratory preliminary analysis. A limitation is that only one participant was considered and only a single night of data was included in the analysis. Additionally, the participant is not known to have a restless-type sleep disorder. Another limitation is the lack of gold standard comparison to confirm the occurrence of movements. These limitations are planned to be addressed in future work where the multi-limb IMU-based actigraphy set-up will be validated against clinical PSG. This data will be collected from children with ASD who have been referred for a clinical sleep evaluation for the purpose of diagnosing restless sleep.
In conclusion, having multi-limb measurements obtained from advance IMU sensors could improve upon traditional actigraphy and thereby reveal insightful discoveries about movement in sleep.

Acknowledgements (Optional): We would like to acknowledge funding sources BCCHRI, MITACS Accelerate, and CFI.