ArticlePDF Available

Recent Developments in Home Sleep-Monitoring Devices



Improving our understanding of sleep physiology and pathophysiology is an important goal for both medical and general wellness reasons. Although the gold standard for assessing sleep remains the laboratory polysomnogram, there is an increasing interest in portable monitoring devices that provide the opportunity for assessing sleep in real-world environments such as the home. Portable devices allow repeated measurements, evaluation of temporal patterns, and self-experimentation. We review recent developments in devices designed to monitor sleep-wake activity, as well as monitors designed for other purposes that could in principle be applied in the field of sleep (such as cardiac or respiratory sensing). As the body of supporting validation data grows, these devices hold promise for a variety of health and wellness goals. From a clinical and research standpoint, the capacity to obtain longitudinal sleep-wake data may improve disease phenotyping, individualized treatment decisions, and individualized health optimization. From a wellness standpoint, commercially available devices may allow individuals to track their own sleep with the goal of finding patterns and correlations with modifiable behaviors such as exercise, diet, and sleep aids.
International Scholarly Research Network
ISRN Neurology
Volume 2012, Article ID 768794, 10 pages
Review A rticle
Recent Developments in Home Sleep-Monitoring Devices
Jessica M. Kelly,
Robert E. Strecker,
and Matt T. Bianchi
Department of Neurology, Massachusetts General Hospital, 55 Fruit Street, Wang 720, Boston, MA 02114, USA
VA Boston Healthcare System and Harvard Medical School, Brockton, MA 02301, USA
Correspondence should be addressed to Matt T. Bianchi,
Received 22 July 2012; Accepted 13 September 2012
Academic Editors: C. G. Carlotti Jr. and H. Ovadia
Copyright © 2012 Jessica M. Kelly et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and repro duction in any medium, provided the original work is properly cited.
Improving our understanding of sleep physiology and pathophysiology is an important goal for both medical and general wellness
reasons. Although the gold standard for assessing sleep remains the laboratory polysomnogram, there is an increasing interest
in portable monitoring devices that provide the opportunity for assessing sleep in real-world environments such as the home.
Portable devices allow repeated measurements, evaluation of temporal patterns, and self-experimentation. We review recent
developments in devices designed to monitor sleep-wake activity, as well as monitors designed for other purposes that could
in principle be applied in the field of sleep (such as cardiac or respiratory sensing). As the body of supporting validation data
grows, these devices hold promise for a variety of health and wellness goals. From a clinical and research standpoint, the capacity
to obtain longitudinal sleep-wake data may improve disease phenotyping, individualized treatment decisions, and individualized
health optimization. From a wellness standpoint, commercially available devices may allow individuals to track their own sleep
with the goal of finding patterns and correlations with modifiable behaviors such as exercise, diet, and sleep aids.
1. Introduction
The laboratory polysomnogram (PSG) has long been the
gold standard for assessing sleep physiology in health and
disease. The PSG has proven most useful for the diag-
nosis and treatment of obstructive sleep apnea (OSA),
although less common disorders are also readily identified by
laboratory PSG, including narcolepsy, rapid-eye-movement
(REM) sleep behavior disorder, non-REM parasomnias, and
periodic limb movements of sleep. The diag nostic criteria
for certain sleep disorders, such as restless legs syndrome
and insomnia, are purely clinical. For insomnia in particular,
exclusive reliance on self-reported complaints presents two
important challenges in understanding this common dis-
order. First, it is not uncommon to observe a mismatch
between the subjective report of sleep-wake durations and
the objective findings of the PSG. Second, the lack of
objective data constrains the capacity to phenotype insomnia
patients, a limitation that has implications ranging from
epidemiology studies to the development of therapeutic
Despite the clear utility of PSG in clinical sleep medicine,
issues of cost and inconvenience have motivated the devel-
opment of portable devices capable of evaluating sleep in
the home. Clinical home testing is currently targeting sleep
disordered breathing, and the data supporting the use of
home sleep apnea devices has been reviewed recently [1, 2].
Although, to date, home-based sleep measurements have
focused on sleep apnea diagnostics, patients with insomnia
may also benefit from advances in home-sleep-monitoring
devi ces. Wrist actigraphy, which measures limb movement,
has been used for several decades in various contexts, includ-
ing in patients with insomnia; however, actigraphy is not
widely used clinically. For example, recent practice param-
eters suggest actigraphy as an option for circadian rhythm
disorders and potentially for adjunctive assessment of insom-
nia [3]. However, the diagnostic classifications of insomnia
provided by the American Academy of Sleep Medicine and
2 ISRN Neurology
by the psychiatric Diagnostic and Statistical Manual do not
include any objective criteria. Despite the extensive published
experience with wrist movement-based monitors in research
settings [4, 5], actigraphy does not enjoy wide clinical use
because of limited utility outside of circadian rhythm
disorders, which are often evident by clinical history alone.
Nevertheless, commercially available movement-based sleep
monitors are growing increasingly common in the consumer
wellness market. The utility of such devices in a medically
unregulated fashion remains uncertain.
The wellness and clinical markets share an interest for
improved metrics of “real-world” sleep patterns. Longitu-
dinal home monitoring avoids certain limitations of the
laboratory PSG, such as the atypical sleeping environment
and the single-night snapshot. Sleep is a dy namic process
that varies from day to day, and hence it is important to
measure multiple nights of sleep for medical, research,
and wellness reasons. Home monitoring devices oer the
potential to provide a more realistic platform in which many
nights of sleep data can be captured. Longitudinal data
is likely to prove invaluable for the discovery of intrinsic
patterns of sleep variability or to correlate sleep with the
timing of var ious other activities such as exercise, naps,
food, caeine, alcohol, and stress. Since each of these
other act ivities” can vary from day to day, complex eects
and interactions are expected to occur, which creates the
need for large data sets to identify correlations with sleep.
Quantifying sleep in relation to these diverse factors can
only be accomplished through longitudinal data, with the
clinical goal of individualized evaluations and treatment
strategies. The personal wellness goal of sleep-monitoring
in order to optimize health also stands to be achieved
through longitudinal monitoring and self-tracking. This
goal is served by portable monitoring in a variety of contexts
outside of the field of sleep medicine (for review, see [6]).
The present paper reviews recent developments in the
area of devices that can be used for home-based sleep
assessment, some of which are currently available for
direct purchase in the wellness market. Devices developed
explicitly for sleep quality monitoring are reviewed, as are
devices developed for other reasons that could potentially
be adapted for sleep-monitoring. The list is not intended
to be exhaustive, and it is likely that the field will continue
to expand rapidly as new devices are introduced. We do
not review devices used for detection or diagnosis of sleep
apnea [1], nor do we review standard actigraphy devices
[7]. When available, validation information is provided (see
Section 7 for further comments on the metrics of sensitivit y,
specificity, and accuracy). These devices are grouped into
categories based on the type of data collected. For each
device, listed alphabetically within category, the key features
are evaluated including availability of published validation
studies. Finally, we discuss a research agenda for the field of
home-based sleep-monitoring. This paper is not intended to
endorse any particular device or to advise readers medically
regarding diagnostics or therapeutics; in fact it is important
to recognize that nonrefreshing sleep can be associated with
numerous medical and psychiatric conditions, and physician
consultation is suggested for concerned readers.
2. Sleep-Monitoring Based on
Brain Activity Signals
2.1. iBrain (NeuroVigil). This device consists of headgear
that records a single frontal lead EEG. The algorithm used
to process the frontal EEG is based originally on work done
in zebra finch birds, in which 84% accuracy was obtained
when compared to manual sleep-wake scoring in that species
[8]. The company web site indicates ongoing human studies;
however, no validation studies in humans are currently
available. Data from the device can be updated via a USB
drive which also charges the device. The device can record
multiple nights of data.
2.2. Zeo (Available for Consumer Purchase). This device con-
sists of an elastic headband with fabric sensors on the
forehead that detect a combination of electroencephalo-
gram (EEG), frontalis muscle e lectromyogram (EMG), and
electrooculogram (EOG) signals. The headband broadcasts
wirelessly to either an alarm clock receiver station or to an
iPhone for analysis. The main advantage of the Zeo is the
capacity to monitor sleep over time with relative ease. Indeed,
the headband sensor can be used daily for several months
before the sensor pads require replacement.
A proprietary neural network model uses the data
streams to render classifications of wake, light NREM, deep
NREM, and REM sleep in 30-second epochs. Deep NREM
sleep corresponds to slow wave sleep or stage N3. The term
deep is often linked to this stage because of the conspicuous
high-amplitude and low-fre quency EEG signal pattern and
because awakening from this stage of sleep is most dicult.
Although the Zeo algorithm assigns greater weight in their
sleep quality index to this stage of sleep, there is little
evidence in the literature that the amount of deep” sleep
correlates with feeling refreshed. Deep sleep has been shown
to correlate with “homeostatic” sleep pressure: the longer one
has been awake, the more sleep pressure accumulates and
the more deep sleep is observed during subsequent recovery
The light NREM sleep class is actually a combination of
two stages of NREM sleep known as N1 and N2. However,
these are fundamental ly dierent states, with the latter
containing two classic features of sleep, known as sleep
spindles and K-complexes, while the former lacks these
features and is instead characterized by mild slowing of the
EEG (relative to high-frequency and low-amplitude EEG
signals typical of wakefulness). In fact, stage N2 represents
the majority of sleep time in a normal individual. It is worth
emphasizing that there is no clinical or biological basis for
combining these two stages, and the potentially negative
connotation of the term “light” may give users the false
impression that sleep scored in this category is necessarily
abnormal. Although excessive stage N1 sleep may indicate
sleep fragmentation (of any cause), this is not the case for
stage N2. T here is no way to distinguish N1 and N2 using
ISRN Neurology 3
this device, due to combining N1 and N2 sleep into a single
Zeo recently published a validation study of their sleep
staging algorithm, which was initially optimized using a
group of healthy adults aged 21–60 (67% male) and then
tested in a separate group of 26 healthy adults aged 19–60
years (50% male) [9]. Subjects underwent laboratory PSG
with simultaneous headband monitoring, which showed
75% agreement on epoch-by-epoch scoring across all
sleep-wake stages. Of note, the two human experts scoring
the PSG data showed only 83% agreement, which can be
viewed as an apparent upper limit of performance for any
automated scoring algorithm. This modest agreement is
similar to prior literature and serves as a reminder that
sleep stage scoring is characterized by substantial uncer-
tainty. When considering each PSG-scored sleep-wake stage
individually, Zeo correctly identified 71% of deep NREM
sleep and 64% of wake, while it was better at detecting
light NREM and REM sleep (86% each). Another way to
understand the accuracy of the classification is to a sk how
likely a stage reported by the Zeo matches that defined by
the PSG scoring. For example, there was a 75% chance that
an epoch scored as REM by the Zeo was correct. When Zeo
was mistaken about REM sleep, the most likely PSG-defined
stages to be misclassified as REM were lig ht NREM and wake.
Epochs scored as deep NREM had a 69% chance of being
correct, while those scored as wake or REM sleep each had
85% chance of being correct.
The extent to which the classifier algorithm retains accu-
racy with patients suering from medical, neurological,
psychiatric, or sleep disorders is unknown. One might expect
certain medications to alter accuracy, due to eects on the
EEG, EMG, and EOG (esp ecially neuroactive medications).
In addition, the eects of caeine, smoking, and alcohol (all
of which are known to aect sleep physiology) also remain
unknown. Zeo does have the option for researchers to pursue
o-line postprocessing of the recorded signals.
3. Sleep-Monitoring Based on
Autonomic Signals
3.1. Heally Recording System. The Heally system consists of
a shirt with a combination of embedded sensors and wired
adhesive electrodes that measure respiratory and cardiac
physiology, as well as ports for optional EMG and EOG
electrodes [10]. A small study of six healthy male subjects
was conducted at home over multiple nights, in which sleep
versus wake was scored according to nonvalidated criteria (a
human scorer classified sleep-wake state using‘a combination
of video, EOG and EMG signals). Like w rist actigraphy, the
shirt overestimated total sleep time as well as the number
of brief awakenings, compared to the human scoring.
The accuracy across subjects was modest at approximately
80% agreement with human scoring, similar to accuracies
obtained with limb actigraphy [10].
3.2. M1 (SleepImage). This medical device consists of a small
processing unit and wire electrode that attaches to the chest
via adhesive pads. Data signals stored locally in the device
include elect rocardiogram (ECG), actigraphy, and body
position. The trunk actigraphy signal is used to determine
total sleep time, sleep eciency, and the number of awak-
enings that occur within sleep. The sig nals are subjected to
o-line analysis through the SleepImage web site. The ECG
component is used to compute cardiopulmonary coupling
frequencies, a metric that consists of a combination of
respiratory-driven heart rate variability (autonomic func-
tion) and fluctuations in the R-wave amplitude that relate to
mechanical changes of breathing (position of the heart and
lung tissue relative to the skin surface) [11]. This algorithm
distinguishes “stable” versus “unstable” NREM sleep, using
the cardiopulmonary coupling metric rather than the brain,
eye, and muscle activity used for the standard classification
of N1, N2, N3, and REM sleep. The relationship between
“stable and “unstable NREM sleep and conventional EEG-
derived sleep stages is described next.
Stable NREM is associated mainly with stage N3 but also
includes portions of N2 and is associated with a coupling
frequency in the range of the normal respiratory rate, which
is around 0.3 Hz. This pattern is known as high-frequency
coupling (HFC). Unstable NREM sleep is associated mainly
with stage N1 but also portions of stage N2, especially when
N2 sleep is frag mented and/or the “cyclic alternating pattern
is seen [12]. This pattern is associated with coupling in a
lower r ange (0.1 Hz) and is known as low-frequency coupling
(LFC). REM sleep and wakefulness produce similar coupling
frequencies, due to similarly irregular breathing. This pattern
is known as very low-frequency coupling (VLFC) and occurs
at frequencies under 0.01 Hz.
When sleep apnea is present, the contribution of the
LFC component is increased, known as elevated LFC or e-
LFC. Within this e-LFC metric, if the frequency is variable,
this is known as broad-band coupling and is associated with
obstructive sleep apnea. This pattern corresponds to the
observation that obstructive apnea events typically have
variable cycle lengths. However, when the dominant e-LFC
values are very similar over time, this is known as narrow
band coupling and is associated with central sleep apnea
which typically has a short and “metronomic cycle length.
Thus, although the device is not approved for the diagnosis
of sleep apnea, within known sleep apnea patients some
distinction can be achieved between obstructive and central
phenotypes [13]. It is worth noting that sleep that is highly
fragmented for a variety of reasons may be dominated by a
high percentage of the night spent in the LFC pattern.
The M1 can be used for 5–7 nights of recording on two
disposable button batteries. The raw ECG data is stored
locally in the device and is extracted o-line for analysis.
Early validation studies took advantage of the fact that
the coupling algorithm can be a pplied to any ECG signal,
such as those obtained routinely in overnight PSG studies.
Analysis of the large Sleep Heart Health Study database
showed correlations of HFC and LFC with important factors
such as stroke and hypertension [14]. Subsequent studies
showed correlations of coupling metrics with depression and
fibromyalgia [15, 16].
4 ISRN Neurology
One limitation of the device is that certain patient
populations may not be amenable to ECG analysis, including
those with certain types of arrhythmias and potentially
patients with autonomic dysfunction. Also, trunk actigraphy
such as that provided by this device does not have as
much supporting data, compared to the traditional wrist
actigraphy, for estimating sleep and wake.
4. Sleep-Monitoring Based on Movement
4.1. Fitbit (Available for Consumer Purchase). The Fitbit
monitor is a small device that can be worn on the wrist,
clipped to clothing, or carried in a pocket. The features
include a pedometer and altimeter (to count steps or h ills
climbed), a calorie counting feature (extrapolated from the
estimate of steps walked), movement detection by actigraphy,
and a clock. The analysis of movement yields standard sleep-
related metrics such as a distinction between sleep and wake,
total sleep time, sleep latency, and an “arousal index” based
on episodes of movement during presumed sleep time. There
are no published validations of the accuracy of the sleep-
wake metrics of the Fitbit compared to PSG or to standard
actigraphy watch devices.
4.2. Lark (Available for Consumer Purchase). The Lark device
is a wrist-watch actigraphy monitor that features a silent
vibrating alarm. Actigraphy metrics include total sleep
duration, sleep latency, and a “sleep quality index” based
on movements. However, there are no published validation
reports comparing Lark-derived measures with standard
wrist actigraphy or PSG data. The device currently requires
an iPhone or iPad or iTouch to visualize the data, although
the web site indicates that an Android platform is under
4.3. Sleep Cycle Alarm (Available for Consumer Purchase).
The Sleep Cycle alarm clock is an iPhone application that
uses the built-in accelerometer of the iPhone to monitor
movement during the night. The iPhone is placed near ones
pillow. The application reports graphs of total sleep time and
a distinction between light sleep, deep sleep, and wake. There
are no available studies of the device to validate this analysis
of sleep. The application also has a smart-alarm feature to
wake users within thirty minutes of their final alarm by
detecting periods of light sleep based on movement. Like
the other devices making this smart-alarm claim, supporting
validation studies are not available.
4.4. SleepTracker (Innovative Sleep Solutions) (Available for
Consumer Purchase). TheSleepTrackerdeviceisawrist
watch that records movement based on actigraphy. Like
similar movement-based devices, the web site claims a smart-
alarm feature that determines optimal points within sleep
to awaken to feel refreshed. The watch has audio-alarm and
vibrating-alarm options. Sleep data can be viewed through
the web site following USB upload, including total sleep
time and a metric of “sleep quality” based on movement.
Although there are no published validation studies of the
smart-alarm feature or the sleep-wake accuracy, the company
has performed testing in 18 adults who underwent simulta-
neous sleep laboratory monitoring for suspected sleep apnea
(unpublished data, personal communication with Lee Loree,
owner). In this study, the device was >90% accurate in
detecting events of sleep disruption, but the relationship of
the detected e vents to clinically defined sleep parameters is
4.5. Up (Jawbone) (Available for Consumer Purchase). The
Up monitor by Jawbone is a bracelet-like device that interacts
with the iPhone. The de vice serves as a pedometer, and
although it reports a distinction between deep and “light”
sleep, there are no published validation studies that compare
the device to PSG or to wrist actigraphy, and even standard
actigr aphy algorithms do not typically allow such a distinc-
tion. The device also includes a smart-alarm feature that
claims to awaken the wearer at the ”optimal” time, but, again,
this commonly reported feature lacks published validation.
4.6. WakeMate (Available for Consumer Purchase). Wake-
Mate is a wristband device that transmits actigraphy data to
sleep time, sleep latency, number of awakenings, and a “sleep
quality” score based on movements. Compatible interfaces
include iPhone, Android, and Blackberry phones. Like the
above devices, it also makes the smart-alarm claim to
determine the optimal wake time w ithin a window ending in
the final alarm setting. The website indicates that the device
is 95%–98% as accurate as standard actigraphy. However,
supporting validation data is not available for either of these
5. Bed-Based Sleep Monitors
5.1. Air Cushion. This is a thin, air-filled cushion designed to
be positioned on top of a mattress [17]. The pressure-sensing
pad records heart rate, respiration rate, snoring, and body
movement. An automated sleep staging algorithm using
heart rate and movement signals was developed based on 27
overnight recordings from eight university students who had
no subjective sleep complaints. The algorithm demonstrated
the following agreement with PSG data: 82.6% for NREM
sleep, 38.3% for REM sleep, and 70.5% for wake. As is
commonly the case with autonomic metrics, REM sleep and
wake were dicult to distinguish.
5.2. Ear lySense Mattress. This device is a piezoelectr ic sensor
that is placed under a mattress. The system measures
respiration, heart rate, snoring, coughing, and movement. In
a study available on their website of 40 children and 16 adults
(who were being evaluated for sleep complaints), a Bayesian
classifier algorithm that combined features of respiration
with movement signals distinguished sleep versus wake with
modest accuracy compared to concurrent PSG scoring. On
an epoch by epoch basis, sleep was detected with a sensitivity
of 84% but a specificity of only 30% (compared to wake);
wake was more accurately identified (sensitivity of 68%
ISRN Neurology 5
and sp ecificity of 80%). Further distinction of REM versus
NREM sleep was also described, but statistics of accuracy are
not presented. However, REM was reportedly misassigned to
periods of light NREM sleep and adjacent wake epochs.
5.3. Emfit Bed Sensor. This system consists of Emfit foil
electrodes placed under neath a foam mattress which record
movement, respiratory rate, and heart rate data [1820].
These data streams were then subject to machine learning
algorithms to optimize agreement with human PSG scoring
in a sample of 17 healthy adults. The mattress algorithm
showed an agreement of 71% with PSG data on an epoch by
epoch basis [19]. Wake and REM sleep were most challenging
to distinguish, as these two states were most often misclassi-
fied. In a similar study using the Emfit foil electrodes, sleep
staging of eleven healthy female participants was moderately
accurate compared to PSG, with an agreement of 76% [18].
In a separate study of nine females, the Emfit bed sensor was
found to have a 79% agreement with PSG data in determin-
ing wake, NREM, and REM sleep states, but REM sleep was
again dicult to classify [20].
5.4. Home Health Station (TERVA). TheHomeHealthSta-
tion is a comprehensive system intended to be set up in a
patient’s home to record and display blood pressure, axil-
lary temperature, respiration rate, heart rate, activity, and
subjective behavioral diary entries [21]. The system includes
a static-charge-sensitive bed engineered by Biomatt Monitor-
ing systems, which measures heart rate, respiration rate, and
time spent in quiet” sleep based on movement data. Pre-
vious studies have found the accuracy of the static-charge-
sensitive bed to be between 86% to 98% for classifying wake
versus sleep [22]. In addition, the bed sensor has been used
to detect sleep apnea: it dete cted sleep-disordered breathing
during 4% of the night in healthy patients compared to 43%
of the night in patients with known sleep apnea [23, 24].
5.5. Linen Sensor. This system consists of electrodes embed-
ded in the pillow case as well as the linens near the foot of
the bed [25]. Validation was conducted in 30 patients under-
going sleep evaluation for a variety of clinical reasons, as
well as six healthy subjects. Data quality was a concern in
their study, as 20% of the recording time was not usable due
to excess movement and/or poor contact with the sensors. In
the six healthy subjects, the bed sensor classified 82% of the
night as NREM sleep and 19% of the night as REM sleep, and
this was relatively accurate compared to standard PSG, which
classified 78% and 23% of night as NREM and REM sleep,
respectively. The number of arousals (which were not defined
in the paper) was underestimated compared to standard PSG
5.6. SleepMinder (BiancaMed). This device is a radiofre-
quency monitor that uses 5.8 GHz frequencies to detect
body movements [26]. The SleepMinder was studied by
placing the sensor above and lateral to the bed, such as on a
bedside table. It was most accurate when placed within
0.5 meters of the bed, with a maximum distance of 2.5
meters. Distinguishing sleep and wake showed 78% accuracy
in a population of 153 subjects who underwent PSG
monitoring for suspected sleep apnea. Total sleep time was
overestimated, which is commonly the case with movement
detection by wearable actigraphy devices. The device per-
formance was less accurate in distinguishing wake, REM,
and stage N1 sleep but reported 96% accuracy in classifying
slow wave sleep. In a separate study of 176 patients who
underwent overnight PSG monitoring for suspe cted sleep
apnea, the device was able to classify subjects with versus
without sleep apnea, based on a cuto value of AHI
=15, with
a sensitivity of 89% and a specificity of 92% [27].
5.7.Touch-FreeLifeCare(TLC)System. The TLC system is
a bed sensor that can transmit information for remote mon-
itoring. This device can be placed underneath any standard
mattress and wirelessly transmits heart rate, respiratory rate,
and movement data. A sleep quality score is generated
based on a combination of sleep duration, restlessness, heart
rate, and breathing rate. However, validation studies of this
sleep quality metric are not available.
6. Other Devices with Potential for
Sleep Monitoring
6.1. BioHarness (Zephyr). This vest-like device is strapped
across the chest and records respiration rate, heart rate, skin
temperature, motor activity, and body position. The data can
be wirelessly transferred for remote monitoring.
6.2. HealthVest (SmartLifeTech). This is a one piece garment
with electrodes embedded in the shirt. Respiration rate, heart
rate, and body position can be measured and monitored
remotely .
6.3. LifeBed (Hoana). This is a bed used in clinical settings
that displays and records respiration and heart rate, and it
also alerts caretakers when a patient is out of bed.
6.4. LifeShirt (VivoMet rics, Rae Systems). This monitor is
a form-fitting garment that measures multiple aspects of
respiratory a nd cardiac physiology, movement, skin temper-
ature, and body position, through a combination of sensors
embedded in the fabric in combination with either dry [28]
or adhesive electrodes [29], including the capacity for remote
monitoring by wireless transmission. There are optional
ports for extending monitoring to other specialized signals
such as oximetry and blood pressure. Although the authors
describe preliminar y findings regarding the use of the shirt
for sleep staging, no validation data is currently available.
6.5. Magic Vest (Foundation Don Gnocchi). Thevestisa
form-fitting combination of cotton and lyrica to facilitate
a close connection between the skin and nonadhesive ECG
electrodes. The ECG signal tracking was reliable for detecting
normal and abnormal rhythms compared with standard
6 ISRN Neurology
ECG [30]. The removable transmitter attaches to the vest and
is about the size of a cell phone.
6.6. Radiofrequency Monitor. Radiofrequency monitors have
been developed to detect movements of the sleeping subject
as a means for unobtrusive monitoring. Li et al. studied the
use of microwave band signals (26.5–40 GHz) [31], which are
often used in radar detection. An antenna underneath the
bed transmits r adiofrequencies to a nearby receiver (i.e.,
a laptop) and allows long-term, overnight respiration and
heart rate monitoring. The accuracy of the device in measur-
ing heart rate was about 80% compared to finger tip pulse
sensor, depending on body position. Of note, the accuracy
of heart rate determination was compromised by nonsupine
body position. The system is proposed as a way to monitor
patients at home for sleep apnea, but such algorithms have
yet to b e developed.
6.7. SenseWear Armband (BodyMedia). This device uses a
combination of sensors including an accelerometer as well
as sensors for heat flux, temperature, and galvanic skin
response. Heart rate variability, body temperature, and other
recoded measurements are used to determine wake, sleep
onset, and total sleep time. In a small study on self-identified
normal sleepers available on their web site, the armband
agreed 85.3% of the time with PSG data in determining sleep
versus wake states.
6.8. SmartShirt (Sensatex). The SmartShirt is a cotton tee-
shirt which uses sensors inside the fabric to measure and
transmit real-time data of heart rate, body temperature, and
6.9. Shirt Monitor (Universidad Carlos III de Madrid). This
smart shirt contains embedded electrodes to measure cardiac
and pulmonary physiology and also includes a global-
positioning system. Body temperature and position as well
as geographic location can be monitored, the latter feature
having high enough spatial resolution for application to
hospital patients. An extended version of the device uses
imbedded electrodes to monitor and wirelessly transmit vital
signs as well as body position and temperature. The shirt is
machine washable.
6.10. Smart Shirt (Numetrex). This line of clothing contains
sensing fibers knitted into the fabric. Sensors record heart
rate for viewing on the accompanying watch receiver.
6.11. V-Patch and Aingeal Devices (Intelesens). Intelesens
develops wearable vital sign monitors for home and hospital
use. The V-patch is an adhesive device that records 3-lead
EKG for up to 7 days. The Aingeal is an adhesive de vice
that records cardiac, respiratory, actigraphy, and temperature
metrics for up to 48 hours of monitoring in hospitalized
patients via a nearby bay station.
6.12. Wealthy ( This device is
a tight-fitting garment which uses impedance pneumog-
raphy to determine respiration, piezoresistive sensors and
accelerometers to determine movement and position, as well
as sensors to track body temperature and heart rate. The shirt
can store data locally or transmit data via bluetooth.
6.13. WristCare (Vivago). This wrist device has a four-day
subject-specific activity adaptation period after which it
tracks movement, skin temperature, and skin conductivity as
well as the location of a patient in the hospital and remotely
transmits the data. The device was designed a s an automatic
alarm device for the elderly and chronically ill. In a study
of 28 adults, WristCare overestimated total sleep time by 59
minutes (whereas wrist actigraphy only overestimated total
sleep time by 41 minutes) in comparison to PSG data [32].
6.14. Wrist Device (AMON). This device is a wristband
which remotely transmits heart rate, blood pressure, oxygen
saturation, and skin temperature. In addition, it has an
The device was tested in 33 healthy adult volunteers during
wakefulness; compared to standard laboratory devices, it had
varying degrees of accuracy when measuring blood pressure,
oxygen saturation, and heart rate [33].
6.15. Zio ( The Zio consists of
a small patch with two electrodes worn on the chest that
enables cardiac monitoring for up to fourteen days. The
disposable patch is intended for clinical use in patients
with cardiac complaints, similar to a Holter monitor. Zio is
designed to record cardiac arrhythmias and heart rate.
7. Discussion
Wearable monitors and passive o-body sensors are grow-
ing in popularity as novel strategies for recording and
transmitting various physiological signals [6, 34, 35]. Their
medical and wellness applications are vast and include the
measurement of sleep patterns and potentially sleep quality.
However, research must parallel this expanding arena to
ensure appropriate validation studies and understanding of
each devices limitations in order to maximize the utility of
home monitoring. Validation is a costly and time-consuming
process, and we discuss here various considerations for the
design and testing of home sleep monitors, in hopes of
providing a research agenda going forward. Peer reviewed
studies remain the gold standard in the biomedical commu-
nity, yet the claims of health and wellness devices have not
been universally held to similar standards. This is crucial in
the field of sleep monitors, since disturbed or nonrefreshing
sleep may be associated with a number of medical and psy-
chiatric disorders and may warrant physician consultation.
We propose topics here to consider as a framework for a
research agenda in this expanding field.
7.1. Hardware Considerations in Sleep-Monitoring. The two
major considerations are cost and comfort, as these may
be the main factors limiting the scope of implementation.
From a cost perspective, both up-front costs and ongoing
costs (disposable parts or software/web site access) should be
ISRN Neurology 7
considered. The part of the body involved in contact-based
devices (headband, wristband, shirt, etc.) may influence fac-
tors such as comfort and the integrity of recording. Whether
the device can fall oor be influenced by subject placement
should be considered—importantly, these factors may dier
from person to person. Battery life and method/frequency
of recharging may also play a role in consumer a cceptance,
especially for devices designed for long-term repeated mon-
itoring. Finally, it is critical to consider the resilience to
various factors present in the sleep environment, such as
body movements, sweating, temperature, humidit y, and the
bed partner (e.g., the bed partner’s movements, sounds, or
sleep disorders may influence the data collected).
7.2. Software Considerations in Sleep-Monitoring. The main
consideration in sleep-monitoring software is ease of use.
The manner of data access may influence user acceptance.
In some cases, the data is processed for output analytics
but not stored in its primary form. This has advantages of
minimizing data storage needs and may be optimal for real-
time analysis and use in the field. Although storing the raw
data for o-line analysis has the advantage of facilitating
algorithm improvements, the need for storage space and/or
frequent uploading to a server may be cumbersome. Given
the wide variety of sleep problems and comorbidities, it may
be dicult if not impossible to have a single algorithm that
is applicable across diverse populations, and thus it seems
highly useful to have the raw data available for ongoing
7.3. Utility of Objective, Longitudinal Monitoring. The clin-
ical paradigm of monitoring in sleep patients involves two
main strategies: the laboratory PSG and the home diary. The
PSG is information rich but has serious limitations of the
unnatural environment and the single-night snapshot. The
diary approach captures a persons experience in their home
environment in a longitudinal manner but lacks objectivity.
Home sleep monitors oer the potential to bridge these
two extremes by providing some objective measures over
time, ideally in parallel with subjective diary reports. In this
way, patterns may be revealed in sleep and symptoms at
the individual level. The longitudinal aspect allows analysis
over multiple time scales, as certain people may have
fluctuations in their sleep or symptoms over days, weeks,
months, seasons, menstrual cycles, and so forth. To the extent
that variability can be found, it then becomes possible to
link this variability to behaviors such as caeine, alcohol,
medications, stress, and exercise—in principle any factor the
individual may care to measure in hopes of finding sleep
correlations. If correlations can be identified, this opens the
opportunity to implement personalized behavioral modifi-
cation plans to optimize sleep. For some individuals, certain
medical or psychiatric treatment may be undertaken to
improve sleep, and a home monitor may provide an adjunc-
tive outcome measure in parallel with subjective response.
There is even data to suggest that simply providing feedback
to individuals with sleep problems, through object ive sleep
measurements, can improve subjective sleep complaints [36].
Finally, from a research and progress standpoint, having the
capacity to add objec tive sleep measurements holds promise
for improving the ability to phenotype sleep disorders such as
insomnia that currently have purely subjective criteria. Such
improvements could theoretically contribute to improved
understanding of which types of treatments (prescriptions
or alternative therapies or behavioral interventions) may be
most beneficial.
7.4. A Comment on Device Validation. Validating a home
sleep monitor involves comparing the performance to some
other measurement. When this comparison involves the
gold standard laboratory PSG, w hich is scored manually
by experienced technicians, it is important to recognize
that this gold standard is itself imprecise. Depending on
the study, the interscorer reliability may be approximately
85%. This sets an upper limit on what can be expected
of an automated algorithm (e.g., see the validation study
of the Zeo device, in which two human scorers were used
[9]). Furthermore, scoring reliability, and by extension,
automated device scoring, may be influenced by the presence
of sleep disorders such as sleep apnea or of factors that
influence the aspect of sleep physiology measured by a home
monitor. For example, an actigraphy device would need
to be separately validated in patients with versus without
intrinsic movement disorders (such as Parkinsons disease).
Ideally, validations should include a spectrum of subject
characteristics (age, sex, BMI, and health status), to improve
the generality of use.
The American Academy of Sleep Medicine scoring
criteria utilizes a time interval of 30 seconds to define an
epoch of sleep, with a “majority rules” approach to assigning
a stage to an epoch that contains features of more than
one stage [37]. Thus, if a device utilizes a time interval
that is either shorter or longer than the AASM criteria, the
validation results may dier. For example, shorter epochs
may capture nuances of sleep architecture, while longer
epochs or smoothing algorithms may yield a dierent image
of sleep physiology.
It is worth mentioning that the term “accuracy” may
carry several meanings. The standard manner of reporting
diagnostic tests in medicine involves the sensitivity and speci-
ficity when tested against a gold standard. In diagnostic tests,
typically one considers a disease to be either present or absent
and a test result to be either positive or negative. In that
setting, sensitivity refers to the portion of patients with the
disease who test positive and specificity refers to the portion
of patients without the disease who test negative. Accuracy
is a term that incorporates sensitivity and specificity but is
strongly dependent on the actual number of disease versus
healthy individuals being tested (i.e., the prevalence or prior
probability of disease). Specifically, accuracy refers to the
sum of true positives and true negatives divided by all tested
subjects. In the measurement of sleep, one can consider the
analysis framework as follows: instead of disease presence
versus absence, the diagnostic device indicates the presence
or absence of sleep. For example, the portion of true sleep
epochs (defined, e.g., by PSG) that are correctly classified by
8 ISRN Neurology
a device as sleep can be called the de vice sensitivit y, while the
portion of true wake epochs correctly classified by a device
as wake can be called the device specificity. If the recording
time were split evenly between wake and sleep (50% each),
one could interpret the accuracy because the evenly divided
time avoids a prevalence bias in the expected number of true
positives and true negatives. However, for most individuals,
wake is such a small part of time in bed that the composite
accuracy metric may be dominated by the sensitivity of
the device, especially if sensitivity and specificity values are
dissimilar. Put another way, if test subjects sleep >95% of
time in bed, a device can report high accuracy if it can
correctly identify sleep epochs most of the time even if it
labels most wake periods as sleep (but the opposite is not
7.5. A Comment on “Normal Sleep”. Although it is com-
monly stated that the average number of hours of sleep
needed by an adult is 8 hours, sleep duration requirements
depend on many factors, and there may be a wide spectrum
of acceptable sleep physiology in humans. Normative sleep
stage data has been published from large data sets [38], but
these studies typically focus on what is called summary statis-
tics, such as the percentage of time spent in various sleep-
wake stages in the night. This coarse view does not capture
much of the rich physiology of sleep. For example, sleep
apnea is known to fragment or inter rupt REM sleep in many
individuals (whether this finding relates to clinical symptoms
remains unproven). If one measures the percentage of the
night spent in REM sleep in people with severe sleep apnea
versus no sleep apnea, there is little or no dierence; however,
if the time spent in REM sleep is measured through more
appropriate methods, called transition analysis, there is clear
evidence of fr agmentation [3941]. This concept applies to
time spent in any sleep-wake stage, and there is growing
evidence that alternative metrics for quantifying sleep-wake
stage architecture provide unique insights and may prove
more relevant for subjective and medical endpoints than the
traditional summary statistics.
Despite the attractive idea that certain aspects of sleep are
more or less important than others for us to feel refreshed
and performing optimally, many challenges remain. For
example, the “sleep cycle length of approximately 90 min-
utes of alternation between REM and NREM sleep is variable
from night to night, and thus that pattern may only be
evident upon averaging across multiple nights. The amount
of time spent in REM sleep may vary depending on disease
states (like sleep apnea), medications (like antidepressants),
or alcohol ingestion. Many medications used for sleep have
been show n to suppress REM and slow wave sleep and yet
may improve the subjective impression of sleep in some
individuals. It is important to recognize that much remains
unknown in terms of what is normal or optimal, a nd the
answers (if they can be surmised) may even dier from
individual to individual. Perhaps the most striking example
of individual variability in sleep involves the symptoms of
sleep apnea, the best described and most dramatic source
of sleep disturbance in the field of sleep medicine. Only
half of individuals with severe sleep apnea have daytime
sleepiness, whether assessed by subjective report or by
objective measurement [42]. The use of home monitoring
may allow individuals to attempt to identify patterns of
interest that correlate with their own subjective sense.
However, caution should be exercised when the output of
sleep monitors overlaps with widely held concepts that have
little clinical basis (e.g., “The device says I’m not getting
enough REM sleep,”) and thus may introduce distractions
from self-discovery.
7.6. A Comment on Smart Alarms. There is clearly a sense of
“face validity” for the concept of an alarm clock allowing one
to wake up at the optimal time, that is, when sleep is already
lightest. Face validity refers to situations in which a concept
is so obvious as to obviate the need for validation data.
Unfortunately, the history of biomedical research teaches
us that most ideas initially felt to have face validity do not
withstand the test of rigorous experimentation. Smart alarm
claims are not new, and patents based on the idea that one
should ideally awaken at a time of light sleep date back over
20 years. The lack of data is concerning, given that this feature
is no doubt an important attraction to potential consumers.
In fact, how the stage of sleep from which one awakens
impacts subjective alertness remains largely unknown. It may
be the case that waking from deep, N3 sleep is more dicult
and some people may experience sleep inertia when aroused
from this stage, and thus alarms that tend to detect periods
when one is less likely to be in this stage may be beneficial
in terms of avoiding sleep inertia. Testing this would be
fairly straightforward. For example, a trial could involve 1-2
weeks of monitoring, in which each morning is randomized
to either alarm at the supposedly optimal” time or at a
nonoptimal time. The subject would be blinded to this and
would only report their level of alertness upon awakening
or how refreshing their sleep was. In this way, one could
determine whether the smart-alarm feature was ac tually
serving some benefit at the individual level. Unfortunately,
such a trial has not been done, despite multiple devices
claiming a smart alarm feature.
7.7. Concluding Comments on a Research Agenda. The most
common measurement technique in the wellness arena is
limb movement. Most commercial devices using actigraphy
do not have available validation studies. This is a critical
limitation given the widespread use of these devices in
the wellness arena due to their ease of measurement and
simple graphical display of data. It is insucient to refer,
as some devices do, to the rich literature of research-
grade wrist actigraphy for two reasons: one is that each
device has nuances of movement detection and analysis that
cannot be assumed to generalize, and the other is that wrist
actigr aphy has enjoyed only limited clinical use. For example,
actigraphy is mainly used to assess gross sleep-wake patterns
over long periods of time in the assessment of circadian
rhythm disorders, and in research studies to ensure certain
sleep-wake schedules are being adhered to for experimental
validity. It is not used to determine sleep stages and is
ISRN Neurology 9
rarely used as a measure of sleep quality outside of research
studies. Thus, while the idea that movement-based analysis
might prove useful for individuals performing longitudinal
tracking is interesting, the limitations of this method should
be appreciated.
The pace of research studies seems to be lagging behind
the pace of advertising in the field of home sleep monitors.
The M1 device has undergone research in medical contexts
and has clearance for use by physicians. Of the products
targeting the consumer wel lness market, the Zeo headband
has published validation data but has not been validated
in those with sleep problems (such as insomnia), medical
illness, or exposures (medications, alcohol, caeine), any of
which might alter the measured signals and thus confound
the built-in analysis. For example, many antidepressant
medications are known to alter several aspects of sleep-
physiology, including muscle tone, eye movements, and EEG
rhythms. This is a common problem in general in medical
trials: the conclusions may only be relevant for the specific
population under the specific conditions of the study. In
other words, generalizing the findings of any study to a
broader population should be undertaken with caution. It
is possible that future developments using analysis of raw
data will clarify the strengths and weaknesses of various
devices w ith respect to particular populations. We suggest
that the diversity of methods being marketed for sleep
monitoring should be subjected to formal validation studies
across a spectrum of populations most likely to benefit, as
the algorithms and validity may dier by population. This
will be crucial to understand monitoring limitations as well
as to maximize the utilit y of the time and money invested in
Dr. M. Bianchi receives funding from the Department of
Neurology, Massachusetts Gener al Hospital, a Young Clini-
cian Aw ard from the Center for Integration of Medicine and
Innovative Technology, and a Harvard Catalyst KL2 Medical
Research Investigator Fellowship.
[1] N. A. Collop, S. L. Tracy, V. Kapur et al., “Obstructive sleep
apnea devices for out-of-center (OOC) testing: technology
evaluation, Journal of Clinical Sleep Medicine, vol. 7, pp. 531–
548, 2011.
[2] N. A. Collop, W. M. Anderson, B. Boehlecke et al., “Clinical
guidelines for the use of unattended portable monitors in the
diagnosis of obstructive sleep apnea in adult patients, Journal
of Clinical Sleep Medicine, vol. 3, no. 7, pp. 737–747, 2007.
[3] M. Littner, C. A. Kushida, W. M. Anderson et al., “Practice
parameters for the role of actigraphy in the study of sleep and
circadian rhythms: an update for 2002, Sleep, vol. 26, no. 3,
pp. 337–341, 2003.
[4] A. Sadeh, The role and validity of actigraphy in sleep medi-
cine: an update, Sleep Medicine Reviews,vol.15,no.4,pp.
259–267, 2011.
[5] S. Ancoli-Israel, R. Cole, C. Alessi, M. Chambers, W. Moor-
croft, and C. P. Pollak, The role of actigraphy in the study of
sleep and circadian rhythms, Sleep, vol. 26, no. 3, pp. 342–392,
[6] A. Dittmar, F. Axisa, G. D elhomme, and C. Gehin, “New
concepts and technologies in home care and ambulatory
monitoring, Studies in Health Technology and Informatics, vol.
108, pp. 9–35, 2004.
[7] A.T.M.VandeWater,A.Holmes,andD.A.Hurley,“Objec-
tive measurements of sleep for non-laboratory settings as
alternatives to polysomnography—a systematic review, Jour-
nal of Sleep Research, vol. 20, no. 1, pp. 183–200, 2011.
[8] P. S. Low, S. S. Shank, T. J. Sejnowski, and D. Margoliash,
“Mammalian-like features of sleep str ucture in zebra finches,
Proceedings of the National Academy of Sciences of the United
States of America, vol. 105, no. 26, pp. 9081–9086, 2008.
[9] J.R.Shambroom,S.E.Fabregas,andJ.Johnstone,“Validation
of an automated wireless system to monitor sleep in healthy
adults, Journal of Sleep Research, vol. 21, pp. 221–230, 2012.
[10] W. Karlen, C. Mattiussi, and D. Floreano, “Sleep and wake
classification with ECG and respiratory eort signals, IEEE
Transactions on Biomedical Circuits and Systems,vol.3,no.2,
pp. 71–78, 2009.
[11] R. J. Thomas, J. E. Mietus, C. K. Peng, and A. L. Goldberger,
An electrocardiogram-based technique to assess cardiopul-
monary coupling during sleep, Sleep, vol. 28, no. 9, pp. 1151–
1161, 2005.
[12] M. G. Terzano and L. Parrino, “Origin and sig nificance of the
cyclic alternating pattern (CAP), Sleep Medicine Reviews, vol.
4, no. 1, pp. 101–123, 2000.
[13] R. J. Thomas, J. E. Mietus, C. K. Peng et al., “Dierentiating
obstructive from central and complex sleep apnea using an
automated electrocardiogram-based method, Sleep, vol. 30,
no. 12, pp. 1756–1769, 2007.
Goldberger, and D. J. Gottlieb, “Prevalent hypertension and
stroke in the sleep heart health study: association with
an ECG-derived spectrographic marker of cardiopulmonary
coupling, Sleep, vol. 32, no. 7, pp. 897–904, 2009.
J. Croord, and R. D. Chervin, “Impaired sleep quality in
fibromyalgia: detection and quantification with ECG-based
cardiopulmonary coupling spectrograms, Sleep Medicine, vol.
11, no. 5, pp. 497–498, 2010.
[16] A. C. Yang, C. H. Yang, C. J. Hong et al., “Sleep state insta-
bilities in major depressive disorder: detection and quantifica-
tion with electrocardiogram-based cardiopulmonary coupling
analysis, Psychophysiology, vol. 48, no. 2, pp. 285–291, 2011.
[17] T. Watanabe and K. Watanabe, “Noncontact method for
sleep stage estimation, IEEE Transactions on Biomedical
Engineering, vol. 51, no. 10, pp. 1735–1748, 2004.
[18] M. Migliorini, A. M. Bianchi, D. Nistic
o et al., Automatic
sleep staging based on ballistocardiographic signals recorded
through bed sensors, Conference Proceedings IEEE Engineering
in Medicine and Biology Society, vol. 2010, pp. 3273–3276,
[19] M. O. Mendez, M. Migliorini, J. M. Kor telainen et al., “Eval-
uation of the sleep quality based on bed sensor signals: time-
variant analysis, Conference Proceedings IEEE Engineering in
Medicine and Biology Society, vol. 2010, pp. 3994–3997, 2010.
[20] J. M. Kortelainen, M. O. Mendez, A. M. Bianchi, M. Matteucci,
and S. Cerutti, “Sleep staging based on signals acquired
through bed sensor, IEEE Transactions on Information Tech-
nology in Biomedicine, vol. 14, no. 3, pp. 776–785, 2010.
10 ISRN Neurology
[21] M. T. Tuomisto, T. Terho, I. Korhonen et al., “Diurnal and
weekly rhythms of health-related variables in home recordings
for two months, Physiology and Behavior,vol.87,no.4,pp.
650–658, 2006.
[22] B. H. Jansen and K. Shankar, “Sleep staging with movement-
related signals, International Journal of Bio-Medical Comput-
ing, vol. 32, no. 3-4, pp. 289–297, 1993.
[23] T. Salmi and L. Leinonen, Automatic analysis of sleep records
with static charge sensitive bed, Electroencephalography and
Clinical Neurophysiology, vol. 64, no. 1, pp. 84–87, 1986.
[24] T. Salmi, M. Partinen, M. Hyyppa, and E. Kronholm, Auto-
matic analysis of static charge sensitive bed (SCSB) recordings
in the evaluation of sleep-related apneas, Acta Neurologica
Scandinavica, vol. 74, no. 5, pp. 360–364, 1986.
[25] S. Devot, A. M. Bianchi, E. Naujokat, M. O. Mendez, A. Braurs,
and S. Cerutti, “Sleep monitor ing through a textile recording
system, Conference Proceedings IEEE Engineering in Medicine
and Biology Society, vol. 2007, pp. 2560–2563, 2007.
[26] P. De Chazal, N. Fox, E. O’Hare et al., “Sleep/wake measure-
ment using a non-contact biomotion sensor, Journal of Sleep
Research, vol. 20, no. 2, pp. 356–366, 2011.
[27] A. Zaaroni, P. de Chazal, C. Heneghan, P. Boyle, P. R. Mppm,
and W. T. McNicholas, “SleepMinder: an innovative contact-
free device for the estimation of the apnoea-hypopnoea
index, Conference Proceedings IEEE Engineering in Medicine
and Biology Society, vol. 2009, pp. 7091–7094, 2009.
[28] J. L. Weber, D. Blanc, A. Dittmar et al., “Telemonitoring of vital
parameters with newly designed biomedical clothing, Studies
in Health Technology and Informatics, vol. 108, pp. 260–265,
[29] F. H. Wilhelm, W. T. Roth, and M. A. Sackner, The LifeShirt:
an advanced system for ambulatory measurement of respira-
tory and cardiac function, Behavior Modification, vol. 27, no.
5, pp. 671–691, 2003.
[30] M. Di Rienzo, F. Rizzo, G. Parati et al., “MagIC System: a new
textile-based wearable device for biological signal monitoring.
Applicability in daily life and clinical setting, Conference
Proceedings Enginee ring in Medicine and Biology Socie ty, vol.
7, pp. 7167–7169, 2005.
[31] C. Li, J. Lin, and Y. Xiao, “Robust overnight monitoring of
human vital signs by a non-contact respiration and heartb eat
detector, Conference Proceedings IEEE Engineering in Medicine
and Biology Society, vol. 1, pp. 2235–2238, 2006.
[32] J. L
onen, I. Korhonen, K. Hirvonen, S. Eskelinen, M.
aki, and M. Partinen, Automatic sleep-wake and nap
analysis with a new wrist worn online activity monitoring
device vivago WristCare, Sleep, vol. 26, no. 1, pp. 86–90, 2003.
[33] U. Anliker, J. A. Ward, P. Lukowicz et al., AMON: a wearable
multiparameter medical monitoring and alert system, IEEE
Transactions on Information Technology in Biomedicine, vol. 8,
no. 4, pp. 415–427, 2004.
[34] M. Pacelli, L. Caldani, and R. Paradiso, Textile piezoresistive
sensors for biomechanical variables monitoring, Conference
Proceedings IEEE Engineering in Medicine and Biology Society,
vol. 1, pp. 5358–5361, 2006.
[35] A. Mazzoldi, D. De Rossi, F. Lorussi, E. P. Scilingo, and R. Par-
adiso, “Smart textiles for wearable motion capture systems,
Autex Research Journal, vol. 2, no. 4, pp. 199–203, 2002.
[36] N. K. Y. Tang and A . G. Harvey, Altering misperception of
sleep in insomnia: behavioral experiment versus verbal feed-
back, Journal of Consulting and Clinical Psychology, vol. 74,
no. 4, pp. 767–776, 2006.
[37] M. H. Silber, S. Ancoli-Israel, M. H. Bonnet et al., The v i sual
scoring of sleep in adults, Journal of Clinical Sleep Medicine,
vol. 3, no. 2, pp. 121–131, 2007.
[38] M. M. Ohayon, M. A. Carskadon, C. Guilleminault, and M.
V. Vitiello, “Meta-analysis of quantitative sleep parameters
from childhood to old age in healthy individuals: developing
normative sleep values across the human lifespan, Sleep, vol.
27, no. 7, pp. 1255–1273, 2004.
D. M. Rapoport, “Sleep continuity measured by surv ival curve
analysis, Sleep, vol. 29, no. 12, pp. 1625–1631, 2006.
[40] B.J.Swihart,B.Cao, K. Bandeen-Roche, and N. M. Punjabi,
“Characterizing sleep structure using the hypnogram, Journal
of Clinical Sleep Medicine, vol. 4, no. 4, pp. 349–355, 2008.
[41] M. T. Bianchi, S. S. Cash, J. Mietus, C. K. Peng, and R.
Thomas, “Obstructive sleep apnea alters sleep stage transition
dynamics, PLoS ONE, vol. 5, no. 6, Article ID e11356, 2010.
[42] R. D. Cher vin, “The multiple sleep latency test and Epworth
sleepiness scale in the assessment of daytime sleepiness,
Journal of Sleep Research, vol. 9, no. 4, pp. 399–401, 2000.
... The telemonitoring strategy has been shown to be effective in increasing adherence to PAP therapy when compared to usual care 41 . The benefits of telemonitoring are early adjustments to PAP device parameters with the potential for increased adherence, reduced time spent visiting the health professional, reduced absenteeism at work, solves problems with distance from urban centers and traffic jam, and reduction costs with face-to-face visits 42,43,[45][46][47] . ...
... The current position of the American Academy of Sleep Medicine (AASM) stresses the importance of telemonitoring in promoting a model of care in which specialized physicians and qualified sleep professionals can work together to improve the provision of healthcare for patients with sleep disorders in general. However, the most recent studies have focused on sleeprelated breathing disorders 43,44 . ...
Full-text available
This document “Proposed management model for the use of telemonitoring to positive airway pressure adherence” was prepared by a special commission of the Brazilian Association of Sleep Medicine, with the objective of recommending a follow-up model for patients undergoing positive airway pressure therapy using telemonitoring. This proposal was prepared based on a survey and analysis of the most up-to-date national and international literature and uses the best available evidence to facilitate the standardization of care by Sleep Science specialists with potential benefit for patients. Among the conclusions of the document, it is emphasized that telemonitoring is an important tool that allows health professionals trained in sleep-disordered breathing to remotely monitor PAP therapy, allowing prompt and, when necessary, daily adjustments to be made in order to increase adherence to treatment. The authors also conclude that the privacy of the data received and shared during the provision of telemonitoring must be respected by the physician or health professional trained in sleep, with the authorization of the patient and/or person responsible, who should be made aware of the short-, medium- and long-term provision of the service.
... Occasionally such studies are accompanied by home-based sleep monitoring devices, which allow for collection of various sleep parameters or even set awakenings. Such devices are often either of low validity in assessing the parameter or hard to come-by (Kelly et al., 2012). As a result, most of the reports derive from morning awakenings and thus only represent a subset of dream experiences (Domhoff, 1996). ...
... In sleep research and clinical sleep assessment the sleep-specific EEG-based assessment combined with other variables, such as eye movements (electro-oculography) and muscle activation (electromyography) is called polysomnography (PSG), often augmented with heart rate, breathing or blood oxygenation measures (Deak & Epstein, 2009). While there are some home-based sleep monitoring devices for consumer use that arguably use electrodes to pick up electric activity in the brain they are only rarely rigorously validated (Kelly et al., 2012), and thus the use of EEG for sleep analysis is mostly constrained to sleep laboratory settings (see section 3.2.3). 29 ...
Full-text available
Every night during sleep we experience an immersive world of dreams, woven together by our sleeping brain unbound by external stimulation. Despite considerable effort the question of why we dream has eluded a conclusive answer. Understanding dreams also arguably makes progress toward answering the broader question of consciousness: why do we experience anything at all? I attempt to illuminate these questions by concentrating on the quintessentially social nature of dreams. First, in Study I a novel theoretical accountthe Social Simulation Theory of dreaming (SST)is proposed, together with the first outlines of a research program for its empirical study. SST suggests the world simulation form of dreams provides clues for its function by preferentially simulating certain kinds of scenariosnamely social interactions. Second, in Studies II and III specific hypotheses derived from the SST in Study I are empirically evaluated. These provide evidence for dreams to contain more social content than corresponding waking life and to remain so even when social interactions are removed from waking life (Sociality Bias). Furthermore, the Strengthening Hypothesis that suggests dreams serve to maintain and/or increase social bonding with close others gains partial support. The Practise and Preparation Hypothesis gained support as dreams simulated positive interactions in one fifth of dream interactions and overall simulate complex social behaviours. The Compensation Hypothesis suggests dreams simulations to increase when waking social contacts are abolished, but this was not supported in the data as dream sociality remained stable despite social seclusion. When excluded from others our dreams reconfigure to decrease simulations of interactions with strangers. However, dreams during normal day-to-day life do not preferentially simulate bond-strengthening interactions with close others. In opposition to previous findings, Study II found no differences in social dream contents between either stage of sleep or time of night. In Study III a short social seclusion showed not only differences in dream content, but also in sleep structure, with an increase in REM sleep. Third, methodological development was undertaken by, both, developing a content analysis method for extracting social episodes in narrative reports (Social Content Scale, SCS; Study II), and by assessing the validity of a novel home sleep monitor device, the Beddit Sleep Tracker (BST). While the SCS proved useful for categorizing the social features in both studies II and III, BST failed to provide accurate sleep data as measured against a polysomnogram. Overall, the development of SST and the initial empirical evidence for some of its hypotheses brings us closer to understanding the twin problems of dreaming and consciousness.
... CSTs+AI could overcome some limitations of PSGs, such as a foreign sleep environment and data captured only from one night of sleep. Having only one night of information can limit the translatability of PSG studies to real life [18]. Sleep changes from night to night and longitudinal data can provide valuable information on trends. ...
Full-text available
This study aims to assess the perspectives and usability of different consumer sleep technologies (CSTs) that leverage artificial intelligence (AI). We answer the following research questions: (1) what are user perceptions and ideations of CSTs (phase 1), (2) what are the users’ actual experiences with CSTs (phase 2), (3) and what are the design recommendations from participants (phases 1 and 2)? In this two-phase qualitative study, we conducted focus groups and usability testing to describe user ideations of desires and experiences with different AI sleep technologies and identify ways to improve the technologies. Results showed that focus group participants prioritized comfort, actionable feedback, and ease of use. Participants desired customized suggestions about their habitual sleeping environments and were interested in CSTs+AI that could integrate with tools and CSTs they already use. Usability study participants felt CSTs+AI provided an accurate picture of the quantity and quality of sleep. Participants identified room for improvement in usability, accuracy, and design of the technologies. We conclude that CSTs can be a valuable, affordable, and convenient tool for people who have issues or concerns with sleep and want more information. They provide objective data that can be discussed with clinicians.
... There are several surveys that summarize the works done in sleep monitoring and focus on different aspects of sleep monitoring. Kelly et al. [45] presented a review article summarizing the recent developments for in-home sleep-monitoring devices. This study mainly focuses on the commercially available devices and categorizes them in brain signal-based, autonomic signal-based, movement-based, and bed-based systems. ...
Quality sleep is very important for a healthy life. Nowadays, many people around the world are not getting enough sleep, which has negative impacts on their lifestyles. Studies are being conducted for sleep monitoring and better understanding sleep behaviors. The gold standard method for sleep analysis is polysomnography conducted in a clinical environment, but this method is both expensive and complex for long-term use. With the advancements in the field of sensors and the introduction of off-the-shelf technologies, unobtrusive solutions are becoming common as alternatives for in-home sleep monitoring. Various solutions have been proposed using both wearable and non-wearable methods, which are cheap and easy to use for in-home sleep monitoring. In this article, we present a comprehensive survey of the latest research works (2015 and after) conducted in various categories of sleep monitoring, including sleep stage classification, sleep posture recognition, sleep disorders detection, and vital signs monitoring. We review the latest research efforts using the non-invasive approach and cover both wearable and non-wearable methods. We discuss the design approaches and key attributes of the work presented and provide an extensive analysis based on ten key factors, with the goal to give a comprehensive overview of the recent developments and trends in all four categories of sleep monitoring. We also collect publicly available datasets for different categories of sleep monitoring. We finally discuss several open issues and future research directions in the area of sleep monitoring.
... The validation data presented in this study suggest that the smart bed may provide reliable, longitudinal estimation of sleep quality in an ecologically valid environment, enabling access to sleep metrics in a large population for longer periods of time than is currently possible with PSG [44,45]. ...
Full-text available
The Sleep Number smart bed uses embedded ballistocardiography, together with network connectivity, signal processing, and machine learning, to detect heart rate (HR), breathing rate (BR), and sleep vs. wake states. This study evaluated the performance of the smart bed relative to polysomnography (PSG) in estimating epoch-by-epoch HR, BR, sleep vs. wake, mean overnight HR and BR, and summary sleep variables. Forty-five participants (aged 22–64 years; 55% women) slept one night on the smart bed with standard PSG. Smart bed data were compared to PSG by Bland–Altman analysis and Pearson correlation for epoch-by-epoch HR and epoch-by-epoch BR. Agreement in sleep vs. wake classification was quantified using Cohen’s kappa, ROC analysis, sensitivity, specificity, accuracy, and precision. Epoch-by-epoch HR and BR were highly correlated with PSG (HR: r = 0.81, |bias| = 0.23 beats/min; BR: r = 0.71, |bias| = 0.08 breaths/min), as were estimations of mean overnight HR and BR (HR: r = 0.94, |bias| = 0.15 beats/min; BR: r = 0.96, |bias| = 0.09 breaths/min). Calculated agreement for sleep vs. wake detection included kappa (prevalence and bias-adjusted) = 0.74 ± 0.11, AUC = 0.86, sensitivity = 0.94 ± 0.05, specificity = 0.48 ± 0.18, accuracy = 0.86 ± 0.11, and precision = 0.90 ± 0.06. For all-night summary variables, agreement was moderate to strong. Overall, the findings suggest that the Sleep Number smart bed may provide reliable metrics to unobtrusively characterize human sleep under real life-conditions.
... Clinical home testing is currently targeting sleep disordered breathing, and the data supporting the use of home sleep apnea devices has been reviewed recently [1,2]. The personal wellness goal of sleep-monitoring in order to optimize health also stands to be achieved through longitudinal monitoring and self-tracking [3]. Sleep monitoring system is a system to track the sleep cycle of humans. ...
Full-text available
Health monitoring systems demand a huge data transaction, especially if it was set on a multi-regional architecture. With a huge data transaction, there is a higher probability of overloaded number of requests. Conventional load balancer requires tunnelling , which means it takes longer time to reach the designated system that handles request and go back to the request source. This paper proposes the enhancement of the current multi-regional architecture in health monitoring systems by applying Domain Name System (DNS) load balancing as the better alternative to conventional load bal-ancer. The proposed architecture significantly increases the throughput and decreases the response time. The proposed architecture performance is compared with the traditional load balancing with a weighted round robin and another sleep monitoring system architecture with Kubernetes but without load balancing. When receiving the request, the data center has to allocate these requests efficiently so that the response time should be minimized to avoid overloading or congestions. Response time, throughput, completion time, Central Processing Unit (CPU) usage and error rate are the metrics that we use here. The proposed architecture achieves the lowest average response time, the highest average throughput, the lowest error rate and the lowest completion time for 50,000 requests hit from each region (Oregon and Singapore).
... This term is widely used by the sleep research community to describe wearables or non-wearable sleep tracking technology [1]. The authors noted in 2015 that there was "scant literature discussing these technologies", providing only two examples of previous studies [4][5][6]. Since its publication in 2015, however, "Consumer Sleep Technologies: A Review of the Landscape" has been cited 167 times according to Google Scholar. The literature is no longer scant. ...
Global demand for sleep-tracking wearables, or consumer sleep technologies (CSTs), is steadily increasing. CST marketing campaigns often feature a scientific component, but the scientific relevancy and monetary value of CST features within the sleep research community remains unquantified. Sleep medicine experts were recruited through social media and nonprobability sampling techniques to complete a survey identifying sleep metrics and device features that are most desirable to the scientific community. A hypothetical purchase task (HPT) estimated economic valuation for devices with different features by price. Forty-six (N=46) respondents with an average of 10±6 years’ experience conducting research in real-world settings completed the online survey. Total sleep time was ranked as the most important measure of sleep followed by objective sleep quality while sleep architecture/depth and diagnostic information were ranked as least important. Experts preferred wrist-worn devices that could reliably determine sleep episodes as short as 20 minutes. Economic value was greater for hypothetical devices with longer battery life. These data set a precedent to determine how scientific relevance of a product impacts the potential market value of a CST device. This is the first known attempt to establish consensus opinion or economic valuation for scientifically-desirable CST features and metrics using expert elicitation.
Objective The objective of this study is to evaluate the validity of an under-mattress monitoring device (Fullpower Technologies) in estimating sleep continuity and architecture, as well as estimating obstructive sleep apnea in an adult population. Methods Adult volunteers (n=102, 55% male and 45% female, aged 40.6 ± 13.7 years with a mean body mass index of 26.8 ± 5.8 kg/m²) each participated in a one-night unattended in-lab study conducted by Fullpower Technologies. Each participant slept on a queen-sized bed with Sleeptracker-AI Monitor sensors placed underneath the mattress. Standard polysomnography (PSG) was simultaneously recorded on the same night. Researchers (FD and CK) were provided de-identified sleep studies and datasets by Fullpower Technologies for analysis. Sleep continuity measures, 30-s epoch-by-epoch sleep stages, and apnea and hypopnea events estimated by an automated algorithm from the Sleeptracker-AI Monitor were compared with the PSG recordings, with the PSG recordings serving as the reference. Results Overall, the Sleeptracker-AI Monitor estimated similar sleep continuity measures compared with PSG. The Sleeptracker-AI Monitor overestimated total sleep time (TST) by an average of 6.3 min and underestimated wake after sleep onset (WASO) by 10.2 min. Sleep efficiency (SE) was similar between the Sleeptracker-AI Monitor and PSG (87.6% and 86.3%, respectively). The epoch-by-epoch accuracy of Sleeptracker-AI Monitor to distinguish 4-stage sleep (wake, light, deep, and REM sleep) was 79.0% (95% CI: 77.8%, 80.2%) with a Cohen's kappa of 0.676 (95% CI: 0.656, 0.697). Thirty-five participants (34.3%) were diagnosed with obstructive sleep apnea (OSA) with an apnea-hypopnea index (AHI) ≥ 5 based on PSG. Accuracy, sensitivity, and specificity for the Sleeptracker-AI Monitor to estimate OSA (an AHI ≥5) were 87.3% (95% CI: 80.8%, 93.7%), 85.7% (95% CI: 74.1%, 97.3%), and 88.1% (95% CI: 80.3%, 95.8%) respectively. The positive likelihood ratio (LR+) for AHI ≥5 was 7.18 (95% CI: 3.69, 14.0), and the negative likelihood ratio (LR-) for AHI ≥5 was 0.16 (95% CI: 0.072, 0.368). Conclusion The Sleeptracker-AI Monitor had high accuracy, sensitivity, and specificity in estimating sleep continuity measures and sleep architecture, as well as in estimating apnea and hypopnea events. These findings indicate that Sleeptracker-AI Monitor is a valid device to monitor sleep quantity and quality among adults. Sleeptracker-AI Monitor may also be a reliable complementary tool to PSG for OSA screening in clinical practice.
The goal of Internet of Medical Things (IoMT) and digital healthcare systems is to provide people with the ease of receiving quality healthcare at the comfort of their homes. Hence, the aim of IoMT is the ubiquitous deployment of home-based healthcare systems. Making such systems intelligent and efficient for timely prediction of critical diseases can save millions of lives while simultaneously reducing the burden on the traditional healthcare systems e.g., hospitals. The advancement in IoT has enabled both patients and doctors to access real time data. This advancement has reduced the cost and energy consumption of digital healthcare systems by using efficient sensors and communication technologies. This paper provides a comprehensive review of various studies conducted for the development and improvement of IoMT. It analyses different sensors used for measurement of various parameters ranging from physiological to emotional signals. It also provides a detailed investigation of different communication technologies being used, their advantages, and limitations. Moreover, digital healthcare systems are now deploying machine learning technology for the prediction of health status of patients. These techniques and algorithms are also discussed. Data security and prediction accuracy are the main concerns in the development of this area. In conclusion, this paper reviews the various digital system designs in the context of healthcare, their methodology, limitations, and the present challenges faced by the e-health sector.
Full-text available
Actigraphy is a method used to study sleep-wake patterns and circadian rhythms by assessing movement, most commonly of the wrist. These evidence-based practice parameters are an update to the Practice Parameters for the Use of Actigraphy in the Clinical Assessment of Sleep Disorders, published in 1995. These practice parameters were developed by the Standards of Practice Committee and reviewed and approved by the Board of Directors of the American Academy of Sleep Medicine. Recommendations are based on the accompanying comprehensive review of the medical literature regarding the role of actigraphy, which was developed by a task force commissioned by the American Academy of Sleep Medicine. The following recommendations serve as a guide to the appropriate use of actigraphy. Actigraphy is reliable and valid for detecting sleep in normal, healthy populations, but less reliable for detecting disturbed sleep. Although actigraphy is not indicated for the routine diagnosis, assessment, or management of any of the sleep disorders, it may serve as a useful adjunct to routine clinical evaluation of insomnia, circadian-rhythm disorders, and excessive sleepiness, and may be helpful in the assessment of specific aspects of some disorders, such as insomnia and restless legs syndrome/periodic limb movement disorder. The assessment of daytime sleepiness, the demonstration of multiday human-rest activity patterns, and the estimation of sleep-wake patterns are potential uses of actigraphy in clinical situations where other techniques cannot provide similar information (e.g., psychiatric ward patients). Superiority of actigraphy placement on different parts of the body is not currently established. Actigraphy may be useful in characterizing and monitoring circadian rhythm patterns or disturbances in certain special populations (e.g., children, demented individuals), and appears useful as an outcome measure in certain applications and populations. Although actigraphy may be a useful adjunct to portable sleep apnea testing, the use of actigraphy alone in the detection of sleep apnea is not currently established. Specific technical recommendations are discussed, such as using concomitant completion of a sleep log for artifact rejection and timing of lights out and on; conducting actigraphy studies for a minimum of three consecutive 24-hour periods; requiring raw data inspection; permitting some preprocessing of movement counts; stating that epoch lengths up to 1 minute are usually sufficient, except for circadian rhythm assessment; requiring interpretation to be performed manually by visual inspection; and allowing automatic scoring in addition to manual scoring methods.
Full-text available
We describe a method for the online classification of sleep/wake states based on cardiorespiratory signals produced by wearable sensors. The method was conceived in view of its applicability to a wearable sleepiness monitoring device. The method uses a fast Fourier transform as the main feature extraction tool and a feedforward artificial neural network as a classifier. We show that when the method is applied to data collected from a single young male adult, the system can correctly classify, on average, 95.4% of unseen data from the same user. When the method is applied to classify data from multiple users with the same age and gender, its accuracy is reduced to 85.3%. However, receiver operating characteristic analysis shows that compared to actigraphy, the proposed method produces a more balanced correct classification of sleep and wake periods. Additionally, by adjusting the classification threshold of the neural classifier, 86.7% of correct classification is obtained.
Full-text available
Guidance is needed to help clinicians decide which out-of-center (OOC) testing devices are appropriate for diagnosing obstructive sleep apnea (OSA). A new classification system that details the type of signals measured by these devices is presented. This proposed system categorizes OOC devices based on measurements of Sleep, Cardiovascular, Oximetry, Position, Effort, and Respiratory (SCOPER) parameters.Criteria for evaluating the devices are also presented, which were generated from chosen pre-test and post-test probabilities. These criteria state that in patients with a high pretest probability of having OSA, the OOC testing device has a positive likelihood ratio (LR+) of 5 or greater coinciding with an in-lab-polysomnography (PSG)-generated apnea hypopnea index (AHI) ≥ 5, and an adequate sensitivity (at least 0.825).Since oximetry is a mandatory signal for scoring AHI using PSG, devices that do not incorporate oximetry were excluded. English peer-reviewed literature on FDA-approved devices utilizing more than 1 signal was reviewed according to the above criteria for 6 questions. These questions specifically addressed the adequacy of different respiratory and effort sensors and combinations thereof to diagnose OSA. In summary, the literature is currently inadequate to state with confidence that a thermistor alone without any effort sensor is adequate to diagnose OSA; if a thermal sensing device is used as the only measure of respiration, 2 effort belts are required as part of the montage and piezoelectric belts are acceptable in this context; nasal pressure can be an adequate measurement of respiration with no effort measure with the caveat that this may be device specific; nasal pressure may be used in combination with either 2 piezoelectric or respiratory inductance plethysmographic (RIP) belts (but not 1 piezoelectric belt); and there is insufficient evidence to state that both nasal pressure and thermistor are required to adequately diagnose OSA. With respect to alternative devices for diagnosing OSA, the data indicate that peripheral arterial tonometry (PAT) devices are adequate for the proposed use; the device based on cardiac signals shows promise, but more study is required as it has not been tested in the home setting; for the device based on end-tidal CO(2) (ETCO(2)), it appears to be adequate for a hospital population; and for devices utilizing acoustic signals, the data are insufficient to determine whether the use of acoustic signals with other signals as a substitute for airflow is adequate to diagnose OSA.Standardized research is needed on OOC devices that report LR+ at the appropriate AHI (≥ 5) and scored according to the recommended definitions, while using appropriate research reporting and methodology to minimize bias.
Full-text available
The availability of a reliable system to record sleep stage measures easily and automatically in ambulatory settings could be of utility for research and clinical work. The aim of this study was to evaluate a novel wireless system (WS) that does not require skilled preparation for the automatic collection and scoring of human sleep. Twenty-nine healthy adults underwent concurrent sleep measurement via the WS, polysomnography (PSG) and an actigraph (ACT) in a sleep laboratory for one assessment night preceded by an acclimation night. The PSG recordings were scored by two experienced trained technicians from separate laboratories. Each recording was scored by both technicians to Rechtschaffen and Kales (R&K) criteria. The WS and ACT were compared with each of the PSG scores and a consensus PSG score, and the PSG scores were compared with each other. Inter-rater agreement was assessed for each pair over all pooled epochs by percentage agreement, Cohen's kappa and intraclass correlation coefficient. The WS agreement with each of the two PSG scores for sleep stages was 75.8 and 74.7%, respectively. WS agreement with each of the two PSG scores for sleep/wakefulness was 92.6 and 91.1%, ACT agreement with PSG was 86.3 and 85.7%. The PSG scorers' agreement with each other for sleep stages was 83.2%, and for sleep/wakefulness was 95.8%. The findings from the current study indicate that the WS may provide an easy to use and accurate complement to other established technologies for measuring sleep in healthy adults.
The implementation of truly wearable instrumented garments capable of recording biomechanical variables is crucial to several fields of application, from multi-media to physical rehabilitation, from sporting to artistic fields. Here we report on wearable devices which are able to read and record the posture and movements of a subject wearing the system. The sensory function of the garments is achieved by fabric strain sensors, based on threads coated with polypyrrole or carbon-loaded rubbers. The presence of conductive elements gives these materials piezoresistive properties, enabling the detection of local strain on the fabric. Strips of strain fabrics are applied together with conductive tracks at strategic points in a shirt and a glove in order to detect the movements of the principal joints. The 'smart shirt'-sensing architecture can be divided into two parts: a textile platform, where a wearable device acquires biomechanical signals, and a hardware/software platform, to which a wireless communication system sends the acquired data after electrical conditioning. 2002
Several additional problems with the ESS are not addressed in Johns’ article. Increasing evidence suggests that in the assessment of sleepiness, the ESS is subject to undesirable confounding variables, including gender (Chervin and Aldrich 1999), psychological influences (Olson et al. 1998), and subjective perception of fatigue, tiredness, and lack of energy (Chervin, 2000a). Although Johns repeatedly argues, based on face validity, that the ESS measures sleep propensity in eight specific situations rather than just one (like the MSLT) (Johns 1991, Johns 1993; Johns 1994; Johns 1998; Johns 2000) he has provided no criterion validity to substantiate this argument. In one study that did test his hypothesis, subjective responses to the ESS item that asks about ‘lying down to rest in the afternoon when circumstances permit’ failed to show any robust association with objective measures in this specific situation, namely the afternoon naps of MSLTs (Chervin et al. 1997).