Instrumentation & Measurement Magazine 24-6 - 74

Fig. 4. Examples from three different sensor types used to detect nighttime behaviors in participant homes. (a) The passive infrared motion sensors mounted on
the wall of each room detect participant motion around the home. An algorithm is able to detect extended sedentary (i.e., little motion) periods in the bedroom
with limited motion in other rooms in order to estimate when the subject is " resting " in their room overnight as a surrogate for sleep estimation. (b) The activity
monitoring wristwatch provides an estimate of sleep periods. (c) The pressure sensitive bedmat provides an estimate of sleep periods (shown) and records heart
rate, respiration rate, and movement activity (not shown) during the night. The bedmat was not initially included in the CART platform. It was added to the CART
platform after about a year and a half in order to capture sleep data even when participants would remove their activity monitoring wristwatches before going to
sleep.
drawbacks in sensors selected for the measurement of specific
health domains. Additionally, data on specific areas of
health may be collected by multiple sensors types, such as
sleep. In the CART platform, sleep related outcome measures
are captured by PIR motion sensors, wearables, and a pressure
sensitive bedmat (Fig. 4). Although the motion sensors are
able to provide estimates of time in bed and nighttime activity
[15], reliable information from this type of sensor in the bedroom
is limited to single person homes due to their inability to
identify motion of a specific individual. The Withings Activité
provides individual information on sleep time and can therefore
be used in multi-person homes. As a wearable device, it
suffers the drawback of participant compliance for using the
device, which may result in incomplete sleep data. Compliance
with wearing this device in a previous study was 60% of
nights for individuals with MCI and AD versus 65% in their
care partners [32]. Pressure sensitive mats placed under the
mattress offer certain advantages over wearable devices. They
rest continuously under the mattress and do not require adherence
to wearing the device, they can monitor an individual's
sleep in a multi-person home, and they provide more accurate
knowledge of bed occupancy. More invasive methods,
such as in-home polysomnography, offer more detailed information
on sleep stages but are not practical for continuous use
in a longitudinal study and also rely on patient adherence for
nightly use. However, this type of device may be helpful for
74
short periods of use to help validate the ability of other sensors
to collect accurate sleep related data [36]. More complete information
on a specific health domain could be obtained with
fusion of data from multiple sensors that would otherwise not
be available with a single sensor type. For example, comprehensive
details about nighttime behavior could be obtained
from a pressure sensitive mat that detects bed occupancy and
sleep measures in combination with PIR wall mounted sensors
that show movement in the home and room location after
a bed exit.
Collaborations with International
Research Groups
Institutional Regulations and Research Ethics
The CART initiative has worked with international collaborators
to ensure that collection of health-related and sensor data
adheres to standards within each country. For Canadian researchers
in the AGE-WELL network, storage of de-identified
clinical and sensor data on the CART patient and data management
system on US based servers was achieved within
Personal Information Protection and Electronic Documents
Act (PIPEDA) guidelines by keeping personal health information
(PHI) in Canadian physical locations or data servers.
This involved setting up participant/home and data management
systems to ensure patient privacy, which has been
IEEE Instrumentation & Measurement Magazine
September 2021

Instrumentation & Measurement Magazine 24-6

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 24-6