ResearchPDF Available

Technological considerations for sensor-assisted chronic pain assessment in natural environments

Technological considerations for sensor-assisted chronic pain assessment in natural
Aditya Ponnada
Personal Health Informatics
Northeastern University, Boston, MA
Author’s Note
Aditya Ponnada, College of Computer and Information Science and Bouve College of
Health Sciences, Northeastern University, Boston, MA.
Submitted in partial fulfilment of the qualifying examination requirement for
the degree of Doctor of Philosophy in Personal Health Informatics. Contact:, +1508-904-7389.
Chronic pain is a prevalent public health problem affecting millions of Americans
every year, but assessing it in naturalistic settings is a challenging task due to its
unpredictable occurrence and fleeting nature (e.g., neck or lower back pain) throughout the
day. Self-report questionnaires are the traditional method for measuring chronic pain,
although recall bias and high response burden limit their utility. Consumer-grade mobile and
wearable technologies bring new opportunities for measuring health behaviour in naturalistic
settings. For example, in-situ self-report methods such as ecological momentary assessments
have been used in research studies wherein multiple chronic pain episodes are captured
throughout the day. Similarly, passive sensing methods for continuously gathering
information on chronic pain episodes are also being explored. While promising, both active
and passive mobile sensing approaches pose challenges for behavioral assessments.
This review explores potential challenges and opportunities when using consumer-
grade mobile and wearable devices for sensor-assisted assessment of chronic pain in
naturalistic settings. In particular, I discuss the technological challenges of cardiovascular,
electrodermal, electromyographic, and active sensing (self-report) approaches using mobile
and wearable devices suitable for chronic pain assessment.
Keywords: Chronic Pain, Pain Assessment, Psychophysiological Sensing, Mobile and
Wearable computing
Technological considerations for sensor-assisted chronic pain assessment in natural
Chronic pain is defined as a distressing experience associated with actual or
potential tissue damage with sensory, emotional, cognitive, and social components
(Williams & Craig, 2016). As many as 1.5 billion people suffer from chronic pain (Analytics,
2015), and it is estimated that 11.2% of the adult population in the United States (US) alone
suffers from chronic pain (Dowell, Haegerich, & Chou, 2016). According to the Institute of
Medicine of the National Academies, 100 million Americans are diagnosed with chronic
pain, more than Americans affected by diabetes (25.8 million), coronary heart disease (16.3
million), and cancer (11.9 million) (Steglitz, Buscemi, & Ferguson, 2012). Chronic pain
accounts for loss of more than 18 hours of productivity per week for each affected individual
in the US (Stewart, Ricci, Chee, Morganstein, & Lipton, 2003). It is known to affect 20%
Europeans (Häuser et al., 2014), 13% Indians (Dureja et al., 2014), and 15.4% Japanese
(Nakamura, Nishiwaki, Ushida, & Toyama, 2011). Therefore, it is reasonable to consider
chronic pain as a global public health problem (Goldberg & McGee, 2011).
Chronic pain assessment is of interest to epidemiologists as well as interventionists.
Epidemiologists seek to study and summarize chronic pain trends in populations, and
interventionists seek to design and evaluate personalized pain management programs. Across
domains, chronic pain assessments traditionally consist of self-report questionnaires and
observation-guides, which have limitations due to recall bias (Gorin & Stone, 2001), response
burden (e.g., (Fuller-Tyszkiewicz et al., 2013), and inability to capture within-subject
variability throughout the day (e.g., (Shiffman, Paty, Gnys, Kassel, & Hickcox, 1996)).
Mobile and wearable technology can enable data collection of health behaviour
(Goodwin, Velicer, & Intille, 2008; A. A. Stone & Broderick, 2007). These devices can
reliably prompt users at different times throughout the day to capture momentary chronic
pain experience (e.g., (Jamison, Mei, & Ross, 2016)). In fact, momentary self-report has been
found to be less prone to recall bias, enabling researchers to capture more accurate
information in situ (Gorin & Stone, 2001). Mobile technologies are more commonplace today
than 10 years ago (IDC, 2016, 2018; Smith, 2017; Statistica, 2018), enabling pain
researchers to gather unprecedented health related data from millions of Americans using
their personal devices (e.g., Precision Medicine Initiative, (F. S. Collins & Varmus, 2015)).
These devices enable repeated assessments in close time intervals, oftentimes continuous
(using passive sensors), potentially capturing more sensitive variability of perceived pain
experience. Finally, mobile and wearable devices are increasingly equipped with sensors
capable of collecting peripheral physiological data from the body passively, enabling robust
data-driven health interventions (Goodwin et al., 2008) such as just-in-time adaptive
interventions (e.g., (S.S. Intille, 2003; S.S. Intille & Ho, 2004; Nahum-Shani et al., 2016)).
However, like any data collection platform, mobile technologies also have limitations.
Sensing methods using mobile devices are prone to inaccuracies due to hardware limitations,
small changes in measurement site on the body, effect of body movement, and device form
factors. In fact, the placement of device, its displacement, and changes in orientation alone
can damage the behavior measurement algorithms, such as physical activity recognition (e.g.,
(Banos, Toth, Damas, Pomares, & Rojas, 2014; Kunze & Lukowicz, 2008; Yurtman &
Barshan, 2017)). Similarly, for self-report assessments, prompt interruptions, complexity of
the mobile interface, and touchscreen sensitivities can influence usability and response rates
in longitudinal studies (Scollon, Kim-Prieto, & Diener, 2003). It is important to consider
these challenges when conducting chronic pain assessments using consumer-grade
mobile/wearable devices in naturalistic environments. This review presents prevalent
challenges, limitations, and opportunities when using mobile/wearable technologies for
chronic pain assessment.
Chronic pain episodes are associated with several peripheral physiological responses
(i.e. pertaining to peripheral nervous system) such as heart rate (and heart rate variability),
galvanic skin responses, and muscle tensions (and movements) (Herta Flor & Turk, 1989).
The physiological responses observed with chronic pain are often referred to as both
antecedents or consequences of chronic pain that increase stress (Herta Flor & Turk, 1989),
highlighting the need for more research studying associations between chronic pain
experiences and physiological responses (Keefe, 1982). Chronic pain is different from
nociceptive pain. Nociceptive pain, also referred to as acute pain, is caused by damage to the
tissues due to an injury, burn, or cut, and it lasts as long as the tissues are in damaged state
(Dray, Urban, & Dickenson, 1994; Merskey, 2007; Nicholson, 2000). On the other hand,
chronic pain is defined as pain that is fleeting, unpredictable, and lasts for six months or
longer (Dray et al., 1994).
In general, peripheral physiological responses found to correlate with chronic pain
include elevated cardiovascular, electrodermal, and electromyography (EMG) responses.
Table 1 summarizes some key studies that have observed statistically significant relationships
between peripheral physiological responses such as muscle tension using electromyography
(EMG), heart rate (HR), and electrodermal activity (EDA) and chronic pain in lab-based pain
studies. These studies were selected because: (1) chronic pain was the primary variable
measured; (2) at least one psychophysiological sensor was used as an objective measure; and
(3) a significant relationship was found between the objective sensing and chronic pain self-
report questionnaires/surveys. The purpose of this paper is to report on the challenges of
using mobile and wearable technologies for chronic pain assessment. The list of studies in
Table 1 is a sample of studies that demonstrate the relationship between different
psychophysiological measures and pain experiences. A detailed review of several studies
examining physiological response to pain can be found in the works of Herta Flor and Turk
(1989), Lin et al. (2018); Peters and Schmidt (1991).
Table 1. Studies examining psychophysiological responses associated with chronic pain
Sensing location
Key findings
Yemm (1969)
Left and right rectus abdominis; left
and right lumbar erector spine
Muscle tension responses increased with
increases in induced pain
Kravitz, Moore,
and Glaros (1981)
Middle lumbar spine
Lower back pain patients exhibited higher
muscle spasms than non-pain patients
G. Collins, Cohen,
Naliboff, and
Schandler (1981)
EMG: bilateral frontalis, bilateral
lumbar erector, HR: Chest, EDA:
Finger tips
Chronic lower back pain patients
exhibited higher frontal EMG response
and arousal (EDA and HRV) compared to
control group.
Cram and Steger
Frontalis, temporalis, masseter,
sternocleidomastoid, cervical
paraspinals, trapezius
Chronic and tension-type headache
patients exhibited higher EMG activity
Soderberg and Barr
Right erector spine, right rectus
Chronic pain patients showed increased
levels of muscle tensions compared to
control group.
Fischer and Chang
Bilateral paraspinal spasm, unilateral at
the painful site
Lower back pain patients did not show
elevation in EMG observed during sleep.
H. Flor, Turk, and
Birbaumer (1985)
Right and left lumbar erector spine
Chronic pain patients showed higher
muscular tension and stress (HRV
Cohen, Swanson,
Schandler, and
McArthur (1986)
EMG: right and left lumbar paraspinal,
abdominal, hamstring
No significant difference between pain
and non-pain groups for muscle tension.
Muscle tensions differed for different
Hermosillo, Rosas,
and Soto (1998)
No information
Individuals with fibromyalgia show
diminished 24-hour heart rate variability.
Hallman, Olsson,
von Scheele, Melin,
and Lyskov (2011)
No information
Pain group participants showed a
decreased resting heart rate variability
during the cold presser tests.
Storella et al.
No information
Increased heart rate variability observed
with reducing chronic pain intensity
All studies listed in Table 1, and many like them not included in this review, used
research-grade equipment to measure physiological responses in a controlled lab setting.
Such equipment is expensive and obtrusive, and thus not scalable, when conducting studies
with large sample sizes longitudinally in naturalistic settings. A traditional and low-cost pain
assessment method is to use recall questionnaires, which are often long with several items
and are complex to respond (high recall burden). However, chronic pain has been shown to
occur frequently throughout the day with changing intensities (Harris et al., 2005), calling
into question whether one-time, recall-based self-report assessments are suitable to accurately
capture chronic pain experiences.
All the studies in Table 1 used self-report methods as chronic pain ground truth
including the Pain Visual Analogue Scales (VAS) (Ohnhaus & Adler, 1975), the McGill Pain
Questionnaire (Melzack, 1975), and The Brief Pain Inventory (Cleeland & Ryan, 1994). It is
assumed that subjective self-report of pain using these questions is trustworthy to be used as
gold standard for in capturing the intended behavior. However, just like objective
measurements, self-report has its own limitations. Most self-report pain inventories use
weekly recall, resulting in biases in the form of “peak effects” or “duration neglect. Peak
effects are caused when respondents are able to recall only the most painful episodes and
ignore any fluctuations in their pain throughout the day (Schneider, Stone, Schwartz, &
Broderick, 2011), whereas duration neglect is caused when responses do not account for
duration between two pain episodes (Fredrickson & Kahneman, 1993). Single item
questionnaires such the pain-VAS only capture one construct of chronic pain, i.e., pain
intensity, which is insufficient because it does not measure other factors such as mood, affect,
and perceived disability, which more comprehensive pain assessment surveys are capable of
measuring (Ballantyne & Sullivan, 2015).
Modern mobile and wearable technology can be useful in capturing chronic pain-
related behaviours in naturalistic settings. These technologies range from commodity
smartphones and smartwatches to smart home sensors and interactive devices (e.g., Amazon
Echo, Amazon Dash Buttons, and Google Home). Native capabilities of these devices
measure motion (e.g., accelerometery), cardiovascular (optical HR sensors), and
electrodermal activities. However, these peripheral physiological measures do not directly
measure chronic pain episodes. They are capable of measuring our primary biological
responses such as arousal. However, we rely on self-report interfaces to gather information
on subjective feelings in the naturalistic settings. Smartphone and smartwatch touchscreen
displays are suitable for gathering self-report information (such as using such as ecological
momentary assessments (EMA)) to measure behaviors, states, and environmental contexts
that cannot be measured using sensors (A.A. Stone & Shiffman, 1994). Therefore, it may be
possible to achieve less burdensome and more comprehensive assessments of chronic pain
symptoms in natural environments by combining the objective sensing capabilities of passive
sensors with active user-input.
However, consumer-grade mobile and wearable technologies have their own
challenges when used for research purposes. Smartphone sensors often have lower sampling
rates than research-designed devices, resulting in less measurement data that can lead to
inaccurate assessment of behaviour (Bayat, Pomplun, & Tran, 2014; Miller; San-Segundo,
Blunck, Moreno-Pimentel, Stisen, & Gil-Martín, 2018; Tryon & Williams, 1996). Similarly,
most pain research studies collect data in controlled laboratory settings, which do not always
represent natural settings lack of ecological validity. Controlled laboratory settings do not
have the same kind of unpredictable behaviors exhibited in the real world (Le Masurier &
Tudor-Locke, 2003; Tudor-Locke, Ainsworth, Thompson, & Matthews, 2002). Therefore, the
findings of the laboratory-based studies cannot be generalized to everyday life. Reliability
and validity of passive sensing can also vary depending on body location of sensing, motion
artefacts, as well as environmental context including temperature, time of the day, and device
battery performance to name a few.
In the following sections, considerations and challenges of using physiological
sensors and self-report interfaces on mobile/wearable devices are reviewed. In particular, we
focus on cardiovascular, electrodermal, electromyographic, and active sensing (self-report)
technologies using consumer-grade mobile/wearable devices.
Cardiovascular sensing
Severe chronic pain episodes are known to increase cardiovascular responses and
emotional distress (Davydov & Perlo, 2015). In fact, clinical studies have shown that an
increase in severity of pain increases hypertension and systolic blood pressure (Bruehl &
Chung, 2004; Drouin & McGrath, 2013; Olsen et al., 2013). Several studies have also found
that chronic pain results in an increase in pain-related distress, which is measured as a
decrease in heart rate variability (Appelhans & Luecken, 2008; Hallman et al., 2011;
Martínez-Lavín et al., 1998; Storella et al., 1999). In fact, Koenig, Loerbroks, Jarczok,
Fischer, and Thayer (2016) found that vagal nerve activity, measured using heart rate
variability, has a negative correlation with increased severity of chronic pain. Even though
heart rate variability cannot measure chronic pain directly, the associated variables such as
stress, anxiety, and vagal nerve activity are widely used proxies for chronic pain severity.
Measuring HR using PPG. Most consumer-grade mobile/wearable devices measure
cardiovascular activities using pulse oximetry. Pulse oximetry works based on the principle
that oxygenated and de-oxygenated haemoglobin have different optical properties (Millikan,
1942), therefore, pulse oximetry measures the change in blood flow under the skin to
measure HR. A typical photoplethysmogram (PPG) sensor has two components an LED
(optical emitter) and a receiver (Figure 1). The LED emits light on the skin and receiver
captures the reflection. Each cardiac cycle results in a periodic change in absorption
properties recorded by the PPG receiver. This way beat-to-beat interval of HR is measured.
Fig. 1 Smartphone and smartwatch PPG sensors. The smartphone sensor uses red PPG light to
measure HR at rest. The smartwatch PPG generally uses green-light PPG, preferable to measure
HR while in motion.
Blood profusion recording from the skin depends on the wavelength of light emitted
from the PPG sensor (Cui, Ostrander, & Lee, 1990). PPG sensors used in mobile and
wearable devices consist of red or green or both light emitting diodes. Most mobile and
wearable devices have a green-light PPG sensor suitable for continuous HR monitoring
during daily activities. Some devices also include red-light PPG sensors that measure HR
while a person is stationary (such as at the end of a physical activity). The red-light PPG
emitter (using near infra-red spectroscopy) has the ability to penetrate light ~10 X deeper
inside tissue layers as compared to green-light PPG. This also allows red-light emissions to
capture information on hydration, muscle saturation, and total haemoglobin. However, a red-
light PPG is prone to disturbance from ambient light. On the other hand, green light is
absorbed less by the skin, thereby reducing the effect of ambient light noise on the HR signal.
Thus, it is more common to find wearable devices using green light instead of red-light PPG.
Effect of Motion Artefacts. PPG is a non-invasive method, which relies on skin to
measure HR, and therefore, measurements tend to be affected by motion artefacts (e.g.,
(Karlen, Lim, Ansermino, Dumont, & Scheffer, 2012)). Matsumura, Rolfe, Lee, and
Yamakoshi (2014) found that green-light and blue-light PPG had higher signal-to-noise ratios
under the influence of motion than the red-light PPG. Similarly, in another study, Lee et al.
(2013) found that a 530nm- 600nm (i.e., green visible light) wavelength was more suitable
for capturing HR in naturalistic situations compared to red light. Furthermore, it has been
found that chest strap HR monitors (typically ECG) are more accurate than wrist-worn HR
monitors when people are moving, and wrist-worn monitors are more accurate than chest
strap HR monitors when people are still (Wang et al., 2017).
Effect of Body Location and Type. PPG sensors depend on recording blood flow to
capture the cardiac cycle, therefore, thickness of the skin, skin colour, amount of blood flow,
and sensor displacements can affect accurate HR measurement. PPG using wearable sensors
commonly take place on the wrist (using smart wrist watches and fitness bands), finger tips
(using specialized PPG measurement device), and forehead (common in electronic-helmets).
Giardino, Lehrer, and Edelberg (2002) have compared fingertip-based PPG measurement
with ECG and found a significant correlation. However, they argue that fingertip-based PPG
is best suited for measuring HR at rest as compared to when in motion. In fact, many
smartphone-based PPG sensors (e.g., Figure 1) are red-light PPG, which require the users to
stay as stationary as possible to even begin the HR recording.
The wrist is the most common place to measure HR continuously. This is perhaps
driven by the rapid rise of wrist-based activity trackers and smartwatches. Typically, PPG
sensors are placed on the back of the wrist-band dials, to measure HR from the medial side of
the wrist. The emitter and recorder are placed next to each other. The number of light
emitters can vary from one to four as shown in Figure 1. However, there has been no formal
study evaluating advantages of multiple light emitters in PPG sensors.
Wrist-based devices are worn similar to wristwatches, and therefore likely to have
higher compliance given their popularity. However, the wrist as a body location does not
have good blood flow compared to other regions of the body. Most wrist-bands are worn
closer to the wrist bone, where blood flow is minimal. Therefore, consumer-gradely-available
wrist-based HR measurements tend to be less accurate research grade electrical measures
(ECG) (Parak & Korhonen, 2014). In addition, green-light PPG gets absorbed more easily by
darker skin tones (Maeda, Sekine, & Tamura, 2011). Furthermore, cosmetic add-ons such as
enamel nail polish have not been found to affect PPG measurement from the fingertip (Brand,
Brand, & Jay, 2002).
A less conventional yet well-studied location for acquiring HR is the forehead. Just
like the wrist and finger-tips, PPG is the most popular method of measuring HR from the
forehead (Wendelken, McGrath, Blike, & Akay, 2004). However, unlike wrist-based HR
measurement where green light is found to be more suitable, Vizbara (2013) found that red
light PPG is more suitable for the forehead. This could be because of the thin layers of tissues
on the forehead as compared to other parts of the body. However, as mentioned previously, it
is important to counter-balance the effect of motion when using red-light PPG, and head
movements are common in everyday life (e.g., (Kim, Ryoo, & Bae, 2007)). The forehead also
limits the assessments of pulse oximetry due to vasodilatation in the region, which is decrease
in blood pressure due to dilation of blood vessels (Jorgensen et al., 1995). Wearable
computing on the forehead is also still ergonomically uncomfortable and socially awkward
for use in real-world settings (Ekandem, Davis, Alvarez, James, & Gilbert, 2012).
Another common location for ambulatory HR measurement is the ear-lobe, given its
soft skin and high blood flow. A sensor attached to this location tends to move/displace less
against the skin compared to wrist (Patterson, McIlwraith, & Yang, 2009). In addition,
Maeda et al. (2011) have found that green-light PPG recordings from the upper arm were less
affected by motion artefacts than HR measured on the wrist, forearm, and finger. However,
research on consumer-grade ear-lobe sensors to measure heart rate is limited.
Miscellaneous effects. In addition to motion, LED color, and body site, HR
measurement is also influenced by external factors. In particular, caffine consumption can
affect heart rate measurements. Nurminen, Niittynen, Korpela, and Vapaatalo (1999) and
Green, Kirby, and Suls (1996) found that caffeine consumption results in elevated measures
of heart rate and blood pressure, which is further exacerbated for patients with hypertension
(also found in chronic pain). Similarly, medications (including pain-killers) taken by patients
can affect heart rate variability (Jandackova, Scholes, Britton, & Steptoe, 2016). Therefore, it
is important to gather contextual information on a user’s context (e.g., consumption practices)
before interpreting that person’s heart rate and related data sensed from the body.
Electrodermal sensing
Electrodermal activity (EDA), also known as galvanic skin response (GSR), or skin
conductance response (SCR), reflects continuous changes in electrical signals from the skin
(Wolfram Boucsein, 2012). EDA is a commonly used measure of peripheral physiological
arousal controlled by the sympathetic nervous system. G. Collins et al. (1981) have found
that patients with chronic lower back had significantly higher galvanic skin response than
control patients during the pain inducing episodes. In fact, Moseley, Nicholas, and Hodges
(2004) found that in a controlled lab setting anticipatory back pain also resulted in increased
levels of electrodermal activity perhaps because of arousal caused by anxiety. In a similar
study, Bonnet and Naveteur (2004) found that patients diagnosed with depression as well as
chronic back pain had significantly higher levels of skin conductance during pain episodes
than non-depression patients. These are just some of the studies that indicate the importance
of measuring electrodermal activity for the assessment of the severity of chronic pain.
Technology to measure EDA. EDA measurement devices include two electrodes
placed on the skin with a gel matching skin salinity to measure continuous changes in tonic
(slowly changing skin conductance level) and phasic (rapidly changing skin conductance
response) electrodermal activity. Phasic skin conductance is caused by sudden/short term
events in the environment such as smell, sound, sight or the anticipation of an event such as
chronic pain. Tonic skin conductance is a slow change in skin conductance depending upon a
participant’s psychological state, skin dryness, or hydration levels. Emotional distress
associated with the chronic pain may result in elevated levels of tonic skin conductance.
Peaks per minute (i.e., total number of elevations within a minute) in skin conductance are
used to identify if the response is tonic (fewer peaks) or phasic (many peaks) (Ermis,
Krakow, & Voss, 2010; Lim et al., 1997).
Fig. 2 (Left) A typical EDA signal (Reproduced from W. Boucsein et al. (2012)). (Right) Measurement
sites for EDA on upper arm.
EDA signals provide four types of information about phasic response: latency, rise
time, amplitude, and recovery time. Latency is the the delay in skin conductance response
(SCR) start time after an arousing stimulus is experienced. Rise time is the amount of time it
takes an SCR to reach its peak in a given time interval. Amplitude is the highest peak value
of the SCR. Recovery time (also known as half recovery time) is the amount of time it takes
for an SCR to return to baseline post peak amplitude. EDA is measured in microSiemens as:
𝐶𝑜𝑛𝑑𝑢𝑐𝑡𝑎𝑛𝑐𝑒 = 1/𝑅𝑒𝑠𝑖𝑠𝑡𝑎𝑛𝑐𝑒
Therefore, during episodes leading to chronic pain, users have been found to
experience increased and SCRs over time. However, this increased SCRs could be due to
increased stress, anxiety or affective changes due to pain and chronic pain itself.
Measurement site and environmental effects. Figure 2 (right) shows the most
suitable locations to capture EDA from the body. The most common site to measure EDA in
laboratory settings is the medial phalanx (Figure 3). However, Scerbo, Freedman, Raine,
Dawson, and Venables (1992) identified that distal phalanx recorded 3.5 times more EDA
amplitude than medial phalanx, suggesting that distal sites could be more sensitive to external
stimuli. Poh, Swenson, and Picard (2010) demonstrated that the distal forearm is a viable
alternative to measuring EDA, which resulted in the development of wrist-based EDA
measurement sensor (Fletcher et al., 2010; Hedman et al., 2009) to capture EDA in natural
settings outside of the laboratory. van Dooren, de Vries, and Janssen (2012), have found that
foot is the most responsive body site for EDA measurement in naturalistic settings, when
compared with the gold standard of finger-tip-based EDA, followed by the shoulders.
Researchers have shown that EDA measurement at the foot is less affected by motion
artefacts than the wrist (Freidman, Martin, Reis, Lambert, & Wilson, 2001; Gravenhorst et
al., 2012). In a 24-hour free-living study where EDA was measured on the fingers, Doberenz,
Roth, Wollburg, Maslowski, and Kim (2011) did not find any systematic influence of
physical activity, ambient temperature, and body-mass-index on EDA.
Fig. 3 (Left) EDA electrodes on the finger. (Right) Wrist-based Q-sensor for EDA measurement
Electromyographic (EMG) sensing
EMG records the electrical voltage originating in nerves that stimulate muscle
contractions at a given location on skin ((Long & Brown, 1964), (Prahlow & Buschbacher,
2003)). Ahern, Follick, Council, Laser-Wolston, and Litchman (1988) found higher muscle
tensions recorded in EMG for chronic pain patients than non-pain patients. However, they did
not find any significant difference in muscle tensions in stationary or quiet standing positions
in lab settings. In fact, Van Damme et al. (2014) found that EMG measurement from the
surface of the skin can be used to differentiate chronic lower back patients from non-pain
patients. Just like EDA, EMG is typically measured using a three-electrode system (positive,
negative, and ground). Muscle tension is measured between the two electrodes (positive and
negative) on the measurement site. When muscle tension is higher, amplitude of the voltage
signal is higher (Figure 4). Often, episodes of chronic pain result in muscle tension at the site
of pain (Cram & Steger, 1983). Therefore, measuring EMG in real time may help localise
chronic pain intensity, provided EMG sensors are placed at the right location. However, this
demands a prior information of the pain location such as neck, lower back, or knees before
placing the sensor a challenge that requires expert intervention.
Fig. 4 An example of a typical EMG signal with muscle tension peaks shown in red lines
(Reproduced from, accessed on 01/14/2018).
The two kinds of EMG measurement methods are surface EMG and intramuscular
EMG. Surface EMG assesses muscle function by recording muscle activity from the surface
above the muscle on the skin. For instance, consumer devices such as Myo-band measure
EMG from the skin of the upper arm (Labs, 2016). However, surface electrodes provide only
limited assessment of underlying muscle activity. Similarly, for subcutaneous measures,
tissue depth at the recording site can be an additional measurement to assess the intensity of
chronic pain (Cifrek, Medved, Tonkovic, & Ostojic, 2009).
Intramuscular EMG is performed using a monopolar needle electrode, which is a fine
wire inserted into a muscle with a surface electrode as a reference; or two fine wires inserted
into muscle referenced to each other (positive and negative). Fine wire recordings are used
for research in kinesiology studies in lab settings. Diagnostic monopolar EMG electrodes are
typically insulated and stiff enough to penetrate the skin, with only the tip exposed using a
surface electrode for reference. Intramuscular EMG is used in severe acute pain assessments
such as spinal cord injuries (Sandoval, 2010). However, intramuscular EMG is an invasive
and complex procedure for in situ assessment of chronic pain, which requires a trained
expert. Therefore, surface EMG is a more pragmatic approach for chronic pain assessment.
Measurement sites. The majority of pain related studies listed in Table 1 (and similar
other studies) have measured pain-related muscle tensions on the wrist and upper arm region.
However, the sensations of chronic pain can also be felt at the lower back, neck, shoulder,
and even knees (e.g., fibromyalgia).
Inhyuk, Myungjoon, Junuk, and Museong (2005) designed an EMG data collection
system for the levator scapulae muscles around the neck (Figure 5). This system can measure
muscle tension during chronic neck pain. However, it is bulky and thus, not suitable for
longitudinal data collection in naturalistic settings. Wrist-based EMG measurements have
also been developed but are affected by motion artefacts similar to EDA and HR
measurements. Therefore, Niijima, Isezaki, Aoki, Watanabe, and Yamada (2017) proposed a
wrist-EMG measurement device that filters typical wrist motion artefacts during different
activities such as typing and eating. However, wrist-based EMG will only be able to measure
muscle tensions on the wrist, which is not a common chronic pain location. There has been a
dearth of research on consumer-grade EMG devices for other locations on the body. When
designing an EMG system, it is important to consider skin dryness, skin thickness, and gel
infections (if present in some participants). An advantage of measuring surface EMG is that it
can work between two electrodes on any part of the skin (where it is safe to do so).
Therefore, challenges of EMG measurement are more associated with form factor than the
measurement site itself, in particular bulky devices and stretching wires that are not
comfortable in the naturalistic settings.
Fig. 5 Wearable EMG-system for neck muscles (Reproduced from (Inhyuk et al., 2005))
Active sensing (self-report)
A common limitation across the previously reviewed peripheral physiological
measures is that they are deployed in laboratory settings under controlled conditions. In many
of the studies cited in this review, standard cold-presser tests were used to induce pain. In a
typical cold-presser test, participants hands are dipped in a temperature controlled cold-water
container. The temperature is reduced in stages to induce pain. This approach has been used
to collect ground truth data to validate pain detection algorithms and sensors (e.g., (Lin et al.,
2018; D. Liu, Peng, Shea, & Picard, 2017)). However, it is impossible to deploy cold-presser
tests in naturalistic settings, especially for longitudinal studies.
Consumer-grade-device sensors suffer from motion artefacts due to the unpredictable
and random movements of everyday human behaviour, which is often controlled for in the
lab experiments. Moreover, chronic pain is a subjective experience and can occur at any time
throughout the day. Therefore, unlike an induced pain in the lab, chronic pain experiences are
unpredictable and can occur any time, depending on a person’s context (Davis, Affleck,
Zautra, & Tennen, 2006; Dray et al., 1994). This sets an expectation in a patients mind of
when the pain can occur, which can bias the measurements. Moreover, uncertainty in chronic
pain occurrences can result in increased levels of stress throughout the day (e.g., (Deyo,
Mirza, Turner, & Martin, 2009)). Finally, consumer-grade phone, smartwatch, and fitness
tracking devices with reliable psychophysiological sensing are at nascent stages.. In fact,
rarely is a consumer-grade device equipped with complete a suite of sensors suitable for
research such as Verily study watch or E4 by Affectiva. These devices are either expensive
for a large scale longitudinal research study or are designed for specific research purposes,
making their adoption harder for everyday use.
One alternative method of measuring subjective experiences using mobile
technologies is “ecological momentary assessment” or EMA. EMA, also known as
experience sampling methods are design to quantify our everyday feelings, states, and
experiences (Mehl & Conner, 2011). by beeping/prompting a user’s smartphone 6-7 times (or
more) a day with a set of multiple-choice questions related to a research construct of interest.
EMA allows repeated measurements of behaviour in multiweek longitudinal studies.
However, in this review, the focus is limited to technological challenges of using EMA. A
detailed review of self-report surveys used in chronic pain literature and their use in EMA
interfaces can be found in the works by May, Junghaenel, Ono, Stone, and Schneider (2018)
,Salaffi, Sarzi-Puttini, and Atzeni (2015), and Waterman et al. (2010).
Technology used. May et al. (2018) examined 62 EMA studies measuring chronic
pain intensity as the primary outcome variable. A large number of studies used smartphones
(from 2010) or PDAs (before 2009) to measure chronic pain and prompted participants 3 6
times a day to measure chronic pain intensity in addition to daily activities, mood, and
medication adherence. Likewise, Lalloo, Kumbhare, Stinson, and Henry (2014) used a
smartphone optimized web-interface to measure pain intensity using a modified McGill Pain
Questionnaire. A native smartphone application is preferred over a web-application due to
lesser reliance on internet connectivity for prompting, better memory management (to store
data locally on the phone), and processing data in real time for just-in-time interventions. In
addition to smartphones, watch-type computers have been used to measure tension type
headaches; however, these interfaces have been extended to measure chronic pain intensity as
well (e.g., (Kikuchi et al., 2006)). In these prototype-computers, participants use a joy-stick
like modality attached to the watch to adjust their perceived pain intensity on a numeric
rating scale (0 for no pain at all to 10 for excruciating pain).
Technology used for EMA has evolved from traditional smartphones to emerging
wearable devices and situated pervasive displays. For instance, Hernandez et al. (2016)
compared the performance of smartwatches and google glass as EMA devices with
smartphone-based EMA. Even though they did not find any significant difference in response
rates using these devices compare to smartphones, the shorter access time on the wearable
devices was favoured over smartphone by their participants. S. Intille, Haynes, Maniar,
Ponnada, and Manjourides (2016) proposed microinteractions-based ecological momentary
assessments (μEMA), where all EMA prompts are reduced to single questions with Yes/No
type answers. Responding to each prompt in μEMA is a quick glanceable tap (a
microinteraction) taking hardly 2s to respond. In fact, in a follow-up study, Ponnada, Haynes,
Maniar, Manjourides, and Intille (2017) demonstrated high response rates in μEMA are
indeed related to quick micro-interactions and not the novelty of smartwatches. Strategies
like μEMA pave the way for temporally dense self-report measures for chronic pain. These
methods provide in-situ assessments with improved compliance and can allow researchers to
capture temporal patterns of pain experiences and related psychophysiological variables
without burdening the users.
Limitations of active sensing. The primary limitation of EMA is response burden
requiring participants to be prompted several times a day. Each prompt requires participants
to reach out to their phones, unlock them, and answer a set of questions. As a result, EMA
could suffer from low compliance (e.g.,(R. L. Collins, Kashdan, & Gollnisch, 2003;
Courvoisier, Eid, & Lischetzke, 2012; E. Dzubur, Huh, Maher, Intille, & Dunton, 2018; E.
Dzubur, Intille, & Dunton, 2018; Fuller-Tyszkiewicz et al., 2013; Morren, van Dulmen,
Ouwerkerk, & Bensing, 2009; Alexander W Sokolovsky, Mermelstein, & Hedeker, 2013;
Wen, Schneider, Stone, & Spruijt-Metz, 2017)). Lower compliance is also exacerbated
because smartphones are not always in close proximity to participants (K. Dey et al., 2011).
Poor compliance in EMA can result in missing values affecting accuracy of the information
gathered (e.g., (Hedeker, Mermelstein, & Demirtas, 2008; B. Liu, Yu, Graubard, Troiano, &
Schenker, 2016)). Researchers have employed various strategies to increase compliance in
EMA such as adding them to lock screens (Truong, Shihipar, & Wigdor, 2014), compliance
visualization (Hsieh, Li, Dey, Forlizzi, & Hudson, 2008), gamification (Berkel, Goncalves,
Hosio, & Kostakos, 2017) or displaying compliance notification to remind users to maintain
high response rates (Alexander W. Sokolovsky, Mermelstein, & Hedeker, 2014).
Context sensitive EMA. Stephen S Intille (2007) proposed the design of context-
sensitive ecological momentary assessment (CS-EMA) that use smartphone sensors to
determine user’s context and then prompt relevant self-report surveys. Due to the technical
challenges of collecting and processing high quality sensor data from mobile devices in real
time, there has been a dearth of context sensitive EMA developed for consumer-grade mobile
devices. For instance, S.S. Intille, Rondoni, Kukla, Anacona, and Bao (2003) and S.S. Intille,
Larson, and Kukla (2002) developed a context-sensitive EMA tool that prompts users only
when a user engages in a specific activity (e.g., leaving the house). Dunton et al. (2014)
developed a smartphone application for sensor-assisted measurement of physical activity and
related behaviors. Likewise, Bachmann et al. (2015) have performed a feasibility test of
sensing sources to enable context-sensitive EMA for ambulatory assessment. They found that
the motion sensors such as accelerometers, magnetometers, and gyroscope to be robust to
enable context-sensitive EMA. However, the common sensing technologies discussed here
(i.e., heart rate, galvanic skin response, and electromyography) are still in nascent stages of
development for consumer-grade devices. van Wel et al. (2017) developed a context-sensitive
EMA to measure health behaviors based on changes in exposimeter that measures changes in
the environmental pollution-levels. More research is required to use sensors such as heart rate
monitor (PPG) on consumer-grade device to trigger relevant survey prompts.
Situated/pervasive EMA. Assessment of self-reported pain in natural settings require
users to be in close proximity to the assessment device. While smartphones are pervasive,
they are within reach only 50% of the time (K. Dey et al., 2011) despite being in the same
room 90% of the time, and smartwatch wear behaviour is sporadic in indoor settings because
people can take off the watches at the end of the day or on weekends (e.g., (Chun, Dey, Lee,
& Kim, 2018; Jeong, Kim, Kim, Lee, & Jeong, 2017; Khakurel, Knutas, Immonen, & Porras,
2017; McMillan et al., 2017; Mujahid, Sierra, Abdalkareem, Shihab, & Shang, 2018; Pizza,
Brown, McMillan, & Lampinen, 2016)). The smartwatch non-wear can result in missingness
in μEMA data. The effect of the non-wear behaviour and its contextual factors, on the data
loss in μEMA has not been studied. Nevertheless, non-wear detection algorithms are being
developed to enable researchers to filter out meaningful wear time data for accurate
behavioral assessment (Cain et al., 2017; Choi, Liu, Matthews, & Buchowski, 2011; Evenson
& Terry, 2009; King, Li, Leishear, Mitchell, & Belle, 2011).
However, advances in smart home technologies (Aipperspach, Woodruff, Anderson,
& Hooker, 2005; Stephen S. Intille, 2002; Munguia Tapia, 2003; Wan, O’Grady, & O’Hare,
2014) have broadened the possibilities of situated self-report technologies. For instance,
HEEDs (Paruthi et al., 2018; Paruthi et al., 2017) are small self-report devices that can be
installed at different locations in a home that allow researchers to gather self-reported
assessment of chronic pain and its symptoms at different locations within a home setting.
HEEDs comprise of tiny displays with pre-programmed survey question. These devices can
be placed at different locations throughout the home. When users are around these devices,
they get user’s attention using the LED lights. User can then tap on a response on the device,
which is sent to user’s smartphone and a remote data storage. The questions on these devices
can be customized depending on the location at home (e.g., diet related questions in the
kitchen area). Situated EMA can also allow researchers to gather other contextual factors
such as prolonged sitting postures, dietary practices, or any medication related information
that could result in elevated chronic pain levels, which was not possible using other self-
report measures as well as physiological sensing methods used in isolation. In particular,
approaches such as μEMA and HEEDs can make use of object usage sensing technologies
(e.g., (Tapia, Intille, & Larson, 2007)) to trigger relevant surveys.
This review discussed different technological challenges and considerations for in-situ
assessment of chronic pain. In particular, mobile and wearable technologies able to capture
chronic pain episodes either objectively (physiologically) or subjectively (through self-report)
in naturalistic settings were highlighted.
The common method of measuring cardiovascular activity using consumer-grade
devices, PPG, is prone to errors due to motion artefacts. In fact, most PPG devices are worn
on the wrist, making HR measurement prone to inaccuracy. Researchers should account for
this potential inaccuracy when using HR as a variable in chronic pain research in naturalistic
settings. One opportunity is to monitor motion using motion sensors such as accelerometers
and gyroscopes to evaluate the extent to which the heart rate data might be noised. Similarly
for electrodermal activity, studies discussed in table 1 have used finger tips to validate EDA
data for pain assessment in lab settings. Recent technological advances in EDA form factors
have focused on obtaining measures from the wrist. However, EDA primarily measures
arousal, which could result from experiences other than pain as well such as frustration and
excitement. Therefore, similar to heart rate, computational models to predict chronic episodes
based on EDA signals are yet to be developed.
Prior work has also found significant relationships between muscle tensions measured
using EMG and chronic pain episodes. However, EMG requires a prior knowledge of the
pain site to attach sensors at the right location. Fewer consumer-grade devices have built-in
EMG sensors for large scale use, making it an expensive measure for in situ assessment of
chronic pain. Current EMG systems are also bulgy for everyday use. Nevertheless, EMG in
consumer-grade devices is a recent phenomenon, and the form factors of the devices are
expected to improve moving forward.
Studies assessing chronic pain in naturalistic settings acknowledge that tools to gather
self-report data used to train models on real-world data are insufficient (e.g., (D. Liu et al.,
2017)). Self-report methods such as EMA often suffer from interruption burden, response
burden, and missing data. This suggests a need for improvements in gathering chronic pain
related self-report data. However, it is worth noticing that different measurement methods
(physiological and self-report interfaces) are capable of measuring different aspects of pain.
Figure 6. Sensor-assisted chronic pain assessment on emerging mobile/wearable technologies
Therefore, a robust combination of these measures in naturalistic settings could allow future
researchers to study and predict the temporal patterns of chronic pain experiences.
For accurate context sensitive pain assessment, researchers could utilize available
physiological sensors on consumer-grade devices coupled with well-timed self-report. In a
sensor-assisted chronic pain assessment, HR sensors could detect moments of reduced
variability (using PPG) and galvanic skin response sensors detect sympathetic arousal (using
EDA). Ether of these two measures or a combination of both can trigger a μEMA (discussed
above) prompt checking for a chronic pain episode with a yes/no answer (like a
microinteraction). This would not only facilitate higher response rates, but also allow
researchers to personalize pain related questions once patterns of pain experiences are
detected. For higher accuracy, EMG sensors could be added to locations of the body where
chronic pain is felt more frequently to gauge localized effects of pain. The information on the
location of pain can be gathered using μEMA. Similarly, HEED-like devices could be used
throughout the house to gather information on different variables that could enhance our
understanding of chronic pain experiences such as mood, affect, and sedentary behavior that
can be helpful in modelling chronic pain in naturalistic settings. Chronic pain researchers are
also interested in studying other experiences such as mood, affect, and perceived exertion at
the end of a physical activity during pain episodes ((Ballantyne & Sullivan, 2015)).
Therefore, context-sensitive EMA prompts on phone based on the μEMA responses could
allow researchers to gather more details occasionally throughout the day. This will enable
EMA self-report prompts to be tailored according to the dense measurements from μEMA
and passive psychophysiological sensing.
The research community interested in chronic pain assessment can develop user-in-
the loop chronic pain detection algorithms, where user-inputs on μEMA or end of day recall
(e.g., day reconstruction method by Kahneman, Krueger, Schkade, Schwarz, and Stone
(2004)) teach algorithms in real time. Involving users to provide information on their
subjective experience could enable development of personalized health technologies to
measure and manage chronic pain in everyday life.
Chronic pain is a prevalent and stressful experience with unpredictable episodes in
everyday life. The disability induced by chronic pain can affect overall quality of life.
Assessing chronic pain in the real-world can enable better pain management interventions
that current methods (especially long surveys) may not be helpful in. It is this space where a
novel combination of mobile and wearable device’ native capabilities (sensing and interactive
self-report interface) can collaborate with traditional approaches of chronic pain assessment
and allow researchers to design robust computational models of chronic pain experiences.
I sincerely thank Prof. Matthew Goodwin (examiner) and Prof. Stephen Intille (advisor) for
their valuable comments in shaping this report.
Ahern, D. K., Follick, M. J., Council, J. R., Laser-Wolston, N., & Litchman, H. (1988).
Comparison of lumbar paravertebral EMG patterns in chronic low back pain patients
and non-patient controls. Pain, 34(2), 153-160.
Aipperspach, R. J., Woodruff, A., Anderson, K., & Hooker, B. (2005). Maps of our lives:
Sensing people and objects together in the home. Retrieved from
Analytics, G. I. (2015). Pain management: A global strategic business report. Global
Industries Analytics, Inc. Retrieved from
Appelhans, B. M., & Luecken, L. J. (2008). Heart rate variability and pain: associations of two
interrelated homeostatic processes. Biological psychology, 77(2), 174-182.
Bachmann, A., Zetzsche, R., Riedel, T., Beigl, M., Reichert, M., Santangelo, P., & Ebner-
Priemer, U. (2015). Identification of relevant sensor sources for context-aware ESM
apps in ambulatory assessment. Paper presented at the Adjunct Proceedings of the
2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing
and Proceedings of the 2015 ACM International Symposium on Wearable
Computers, Osaka, Japan.
Ballantyne, J. C., & Sullivan, M. D. (2015). Intensity of Chronic Pain--The Wrong Metric? N
Engl J Med, 373(22), 2098-2099. doi:10.1056/NEJMp1507136
Banos, O., Toth, A. M., Damas, M., Pomares, H., & Rojas, I. (2014). Dealing with the effects
of sensor displacement in wearable activity recognition. Sensors, 14(6).
Bayat, A., Pomplun, M., & Tran, D. A. (2014). A Study on Human Activity Recognition Using
Accelerometer Data from Smartphones. Procedia Computer Science, 34, 450-457.
Berkel, N. V., Goncalves, J., Hosio, S., & Kostakos, V. (2017). Gamification of mobile
experience sampling improves data quality and quantity. Proc. ACM Interact. Mob.
Wearable Ubiquitous Technol., 1(3), 1-21. doi:10.1145/3130972
Bonnet, A., & Naveteur, J. (2004). Electrodermal activity in low back pain patients with and
without co-morbid depression. International Journal of Psychophysiology, 53(1), 37-
Boucsein, W. (2012). Electrodermal activity: Springer Science & Business Media.
Boucsein, W., Fowles, D. C., Grimnes, S., Ben-Shakhar, G., roth, W. T., Dawson, M. E., . . .
Society for Psychophysiological Research Ad Hoc Committee on Electrodermal, M.
(2012). Publication recommendations for electrodermal measurements.
Psychophysiology, 49(8), 1017-1034. doi:10.1111/j.1469-8986.2012.01384.x
Brand, T. M., Brand, M. E., & Jay, G. D. (2002). Enamel Nail Polish Does Not Interfere With
Pulse Oximetry Among Normoxic Volunteers. Journal of Clinical Monitoring and
Computing, 17(2), 93-96. doi:10.1023/a:1016385222568
Bruehl, S., & Chung, O. Y. (2004). Interactions between the cardiovascular and pain
regulatory systems: an updated review of mechanisms and possible alterations in
chronic pain. Neuroscience & Biobehavioral Reviews, 28(4), 395-414.
Cain, K. L., Bonilla, E., Conway, T. L., Schipperijn, J., Geremia, C. M., Mignano, A., . . . Sallis, J.
F. (2017). Defining accelerometer nonwear time to maximize detection of sedentary
time in youth. Pediatric Exercise Science, 1-8. doi:10.1123/pes.2017-0132
Choi, L., Liu, Z., Matthews, C. E., & Buchowski, M. S. (2011). Validation of accelerometer
wear and nonwear time classification algorithm. Med Sci Sports Exerc, 43(2), 357-
364. doi:10.1249/MSS.0b013e3181ed61a3
Chun, J., Dey, A., Lee, K., & Kim, S. (2018). A qualitative sutdy of smartwatch usage and its
usability. Human Factors and Ergonomics in Manufacturing & Service Industries.
Cifrek, M., Medved, V., Tonkovic, S., & Ostojic, S. (2009). Surface EMG based muscle fatigue
evaluation in biomechanics. Clin Biomech (Bristol, Avon), 24(4), 327-340.
Cleeland, C. S., & Ryan, K. (1994). The Brief Pain Inventory. Available onli ne: https://www.
mdanderson. org/education-andresearch/departments-programs-and-
tools/BPI_UserGuide. pdf (accessed on 1 March 2016).
Cohen, M. J., Swanson, G. A., Naliboff, B. D., Schandler, S. L., & McArthur, D. L. (1986).
Comparison of electromyographic response patterns during posture and stress tasks
in chronic low back pain patterns and control. J Psychosom Res, 30(2), 135-141.
Collins, F. S., & Varmus, H. (2015). A new initiative on precision medicine. New England
Journal of Medicine, 372(9), 793-795. doi:doi:10.1056/NEJMp1500523
Collins, G., Cohen, M., Naliboff, B., & Schandler, S. (1981). Comparative analysis of
paraspinal and frontalis EMG, heart rate and skin conductance in chronic low back
pain patients and normals to various postures and stress. Scandinavian Journal of
Rehabilitation Medicine, 14(1), 39-46.
Collins, R. L., Kashdan, T. B., & Gollnisch, G. (2003). The feasibility of using cellular phones to
collect ecological momentary assessment data: Application to alcohol consumption.
Experimental and Clinical Psychopharmacology, 11(1), 73-78.
Courvoisier, D. S., Eid, M., & Lischetzke, T. (2012). Compliance to a cell phone-based
ecological momentary assessment study: The effect of time and personality
characteristics. Psychol Assess, 24(3), 713-720. doi:10.1037/a0026733
Cram, J. R., & Steger, J. C. (1983). EMG scanning in the diagnosis of chronic pain.
Biofeedback and Self-Regulation, 8(2), 229-241.
Cui, W., Ostrander, L. E., & Lee, B. Y. (1990). In vivo reflectance of blood and tissue as a
function of light wavelength. IEEE Transactions on Biomedical Engineering, 37(6),
632-639. doi:10.1109/10.55667
Davis, M. C., Affleck, G., Zautra, A. J., & Tennen, H. (2006). Daily interpersonal events in pain
patients: applying action theory to chronic illness. J Clin Psychol, 62(9), 1097-1113.
Davydov, D. M., & Perlo, S. (2015). Cardiovascular activity and chronic pain severity.
Physiology & behavior, 152(Pt A), 203-216. doi:10.1016/j.physbeh.2015.09.029
Deyo, R. A., Mirza, S. K., Turner, J. A., & Martin, B. I. (2009). Overtreating chronic back pain:
time to back off? J Am Board Fam Med, 22(1), 62-68.
Doberenz, S., Roth, W. T., Wollburg, E., Maslowski, N. I., & Kim, S. (2011). Methodological
considerations in ambulatory skin conductance monitoring. International Journal of
Psychophysiology, 80(2), 87-95. doi:10.1016/j.ijpsycho.2011.02.002
Dowell, D., Haegerich, T. M., & Chou, R. (2016). Cdc guideline for prescribing opioids for
chronic painunited states, 2016. JAMA, 315(15), 1624-1645.
Dray, A., Urban, L., & Dickenson, A. (1994). Pharmacology of chronic pain. Trends in
Pharmacological Sciences, 15(6), 190-197. doi:
Drouin, S., & McGrath, J. J. (2013). Blood pressure and pain sensitivity in children and
adolescents. Psychophysiology, 50(6), 513-520. doi:10.1111/psyp.12036
Dunton, G. F., Dzubur, E., Kawabata, K., Yanez, B., Bo, B., & Intille, S. (2014). Development of
a smartphone application to measure physical activity using sensor-assisted self-
report. Frontiers in Public Health, 2.
Dureja, G. P., Jain, P. N., Shetty, N., Mandal, S. P., Prabhoo, R., Joshi, M., . . . Phansalkar, A.
A. (2014). Prevalence of chronic pain, impact on daily life, and treatment practices in
India. Pain Pract, 14(2), E51-62. doi:10.1111/papr.12132
Dzubur, E., Huh, J., Maher, J. P., Intille, S. S., & Dunton, G. F. (2018). Response patterns and
intra-dyadic factors related to compliance with ecological momentary assessment
among mothers and children. Transl Behav Med. doi:10.1093/tbm/ibx002
Dzubur, E., Intille, S., & Dunton, G. F. (2018). Reactivity to a longitudinal smartphone-based
time-intensive physical activity assessment. Paper presented at the 39th Annual
Meeting and Scientific Sessions of the Society of Behavioral Medicine (SBM), New
Orleans, LA.
Ekandem, J. I., Davis, T. A., Alvarez, I., James, M. T., & Gilbert, J. E. (2012). Evaluating the
ergonomics of BCI devices for research and experimentation. Ergonomics, 55(5),
592-598. doi:10.1080/00140139.2012.662527
Ermis, U., Krakow, K., & Voss, U. (2010). Arousal thresholds during human tonic and phasic
REM sleep. J Sleep Res, 19(3), 400-406. doi:10.1111/j.1365-2869.2010.00831.x
Evenson, K. R., & Terry, J. W., Jr. (2009). Assessment of differing definitions of
accelerometer nonwear time. Res Q Exerc Sport, 80(2), 355-362.
Fischer, A. A., & Chang, C. H. (1985). Electromyographic Evidence of Paraspinal Muscle
Spasm During Sleep in Patients With Low Back Pain. Clin J Pain, 1(3), 147-154.
Fletcher, R. R., Dobson, K., Goodwin, M. S., Eydgahi, H., Wilder-Smith, O., Fernholz, D., . . .
Picard, R. W. (2010). iCalm: wearable sensor and network architecture for wirelessly
communicating and logging autonomic activity. IEEE Trans Inf Technol Biomed, 14(2),
215-223. doi:10.1109/TITB.2009.2038692
Flor, H., & Turk, D. C. (1989). Psychophysiology of chronic pain: do chronic pain patients
exhibit symptom-specific psychophysiological responses? : American Psychological
Flor, H., Turk, D. C., & Birbaumer, N. (1985). Assessment of stress-related
psychophysiological reactions in chronic back pain patients. J Consult Clin Psychol,
53(3), 354-364.
Fredrickson, B. L., & Kahneman, D. (1993). Duration neglect in retrospective evaluations of
affective episodes. J Pers Soc Psychol, 65(1), 45-55.
Freidman, R. N., Martin, J., Reis, G., Lambert, J., & Wilson, G. F. (2001). Evaluation of flexible
cloth electrodes for electrodermal activity recording. Aviat Space Environ Med, 72(8),
Fuller-Tyszkiewicz, M., Skouteris, H., Richardson, B., Blore, J., Holmes, M., & Mills, J. (2013).
Does the burden of the experience sampling method undermine data quality in state
body image research? Body Image, 10(4), 607-613.
Giardino, N. D., Lehrer, P. M., & Edelberg, R. (2002). Comparison of finger plethysmograph
to ECG in the measurement of heart rate variability. Psychophysiology, 39(2), 246-
253. doi:10.1111/1469-8986.3920246
Goldberg, D. S., & McGee, S. J. (2011). Pain as a global public health priority. BMC public
health, 11, 770-770. doi:10.1186/1471-2458-11-770
Goodwin, M. S., Velicer, W. F., & Intille, S. S. (2008). Telemetric monitoring in the behavior
sciences. Behavior Research Methods, 40(1), 328-341.
Gorin, A. A., & Stone, A. A. (2001). Recall biases and cognitive errors in retrospective self
reports: A call for momentary assessments. In A. Baum, T. Revenson, & J. Singer
(Eds.), Handbook of Health Psychology (pp. 405-413). Mahwah, NJ: Erlbaum.
Gravenhorst, F., Tessendorf, B., Muaremi, A., Kappeler-Setz, C., Arnrich, B., & Tröster, G.
(2012). Mobile system for unobtrusive monitoring of electrodermal activity in daily
life. Paper presented at the 6th International Workshop on Ubiquitous Health and
Wellness (UbiHealth 2012).
Green, P. J., Kirby, R., & Suls, J. (1996). The effects of caffeine on blood pressure and heart
rate: A review. Ann Behav Med, 18(3), 201-216. doi:10.1007/BF02883398
Hallman, D. M., Olsson, E. M., von Scheele, B., Melin, L., & Lyskov, E. (2011). Effects of heart
rate variability biofeedback in subjects with stress-related chronic neck pain: a pilot
study. Appl Psychophysiol Biofeedback, 36(2), 71-80. doi:10.1007/s10484-011-9147-
Harris, R. E., Williams, D. A., McLean, S. A., Sen, A., Hufford, M., Gendreau, R. M., . . . Clauw,
D. J. (2005). Characterization and consequences of pain variability in individuals with
fibromyalgia. Arthritis Rheum, 52(11), 3670-3674. doi:10.1002/art.21407
Häuser, W., Wolfe, F., Henningsen, P., Schmutzer, G., Brähler, E., & Hinz, A. (2014). Untying
chronic pain: prevalence and societal burden of chronic pain stages in the general
population - a cross-sectional survey. BMC public health, 14(1), 352.
Hedeker, D., Mermelstein, R. J., & Demirtas, H. (2008). An application of a mixed-effects
location scale model for analysis of Ecological Momentary Assessment (EMA) data.
Biometrics, 64(2), 627-634. doi:10.1111/j.1541-0420.2007.00924.x
Hedman, E., Wilder-Smith, O., Goodwin, M. S., Poh, M. Z., Fletcher, R., & Picard, R. (2009,
10-12 Sept. 2009). iCalm: Measuring electrodermal activity in almost any setting.
Paper presented at the 2009 3rd International Conference on Affective Computing
and Intelligent Interaction and Workshops.
Hernandez, J., McDuff, D., Infante, C., Maes, P., Quigley, K., & Picard, R. (2016). Wearable
ESM: differences in the experience sampling method across wearable devices. Paper
presented at the Proceedings of the 18th International Conference on Human-
Computer Interaction with Mobile Devices and Services, Florence, Italy.
Hsieh, G., Li, I., Dey, A., Forlizzi, J., & Hudson, S. E. (2008). Using visualizations to increase
compliance in experience sampling Proceedings of UbiComp '08 (pp. 164-167): ACM.
IDC. (2016). The Worldwide Wearables Market Leaps 126.9% in the Fourth Quarter and
171.6% in 2015, According to IDC
IDC. (2018). IDC Forecasts Shipments of Wearable Devices to Nearly Double by 2021 as
Smart Watches and New Product Categories Gain Traction. Retrieved from
Inhyuk, M., Myungjoon, L., Junuk, C., & Museong, M. (2005, 18-22 April 2005). Wearable
EMG-based HCI for Electric-Powered Wheelchair Users with Motor Disabilities. Paper
presented at the Proceedings of the 2005 IEEE International Conference on Robotics
and Automation.
Intille, S., Haynes, C., Maniar, D., Ponnada, A., & Manjourides, J. (2016). μEMA:
Microinteraction-based ecological momentary assessment (EMA) using a
smartwatch Proceedings of the 2016 ACM International Joint Conference on
Pervasive and Ubiquitous Computing (pp. 1124-1128): ACM.
Intille, S. S. (2002). Designing a home of the future. IEEE Pervasive Computing, 1, 80-86.
Intille, S. S. (2003). Ubiquitous computing technology for just-in-time motivation of behavior
change Proceedings of the UbiHealth Workshop.
Intille, S. S. (2007). Technological innovations enabling automatic, context-sensitive
ecological momentary assessment. In A. Stone, S. Shiffman, A. Atienza, & L. Nebeling
(Eds.), The Science of Real-Time Data Capture: Self-Report in Health Research (pp.
308-337). New York, NY: Oxford University Press.
Intille, S. S., & Ho, J. (2004). A just-in-time persuasive interface for healthy aging. Retrieved
from Cambridge, MA:
Intille, S. S., Larson, K., & Kukla, C. (2002). Just-in-time context-sensitive questioning for
preventative health care Proceedings of the AAAI 2002 Workshop on Automation as
Caregiver: The Role of Intelligent Technology in Elder Care. Menlo Park, CA: AAAI
Intille, S. S., Rondoni, J., Kukla, C., Anacona, I., & Bao, L. (2003). A context-aware experience
sampling tool Proceedings of CHI '03 Extended Abstracts on Human Factors in
Computing Systems (pp. 972-973). New York, NY: ACM Press.
Jamison, R. N., Mei, A., & Ross, E. L. (2016). Longitudinal trial of a smartphone pain
application for chronic pain patients: Predictors of compliance and satisfaction. J
Telemed Telecare. doi:10.1177/1357633X16679049
Jandackova, V. K., Scholes, S., Britton, A., & Steptoe, A. (2016). Are Changes in Heart Rate
Variability in Middle-Aged and Older People Normative or Caused by Pathological
Conditions? Findings From a Large Population-Based Longitudinal Cohort Study. J Am
Heart Assoc, 5(2). doi:10.1161/JAHA.115.002365
Jeong, H., Kim, H., Kim, R., Lee, U., & Jeong, Y. (2017). Smartwatch Wearing Behavior
Analysis: A Longitudinal Study. Proc. ACM Interact. Mob. Wearable Ubiquitous
Technol., 1(3), 1-31. doi:10.1145/3131892
Jorgensen, J. S., Schmid, E. R., Konig, V., Faisst, K., Huch, A., & Huch, R. (1995). Limitations of
forehead pulse oximetry. J Clin Monit, 11(4), 253-256.
K. Dey, A. K., Wac, K., Ferreira, D., Tassini, K., Hong, J.-K., & Ramos, R. (2011). Getting closer:
An empirical investigation of the proximity of user to their smart phones. Paper
presented at the Proceedings of the 13th International Conference on Ubiquitous
Computing, Beijing, China.
Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., & Stone, A. A. (2004). A survey
method for characterizing daily life experience: The day reconstruction method.
Science, 306(5702), 1776-1780. doi:10.1126/science.1103572
Karlen, W., Lim, J., Ansermino, J. M., Dumont, G., & Scheffer, C. (2012). Design challenges
for camera oximetry on a mobile phone. Conf Proc IEEE Eng Med Biol Soc, 2012,
2448-2451. doi:10.1109/EMBC.2012.6346459
Keefe, F. J. (1982). Behavioral assessment and treatment of chronic pain: current status and
future directions. J Consult Clin Psychol, 50(6), 896-911.
Khakurel, J., Knutas, A., Immonen, M., & Porras, J. (2017). Intended use of smartwatches and
pedometers in the university environment: an empirical analysis. Paper presented at
the Proceedings of the 2017 ACM International Joint Conference on Pervasive and
Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium
on Wearable Computers, Maui, Hawaii.
Kikuchi, H., Yoshiuchi, K., Miyasaka, N., Ohashi, K., Yamamoto, Y., Kumano, H., . . .
Akabayashi, A. (2006). Reliability of recalled self-report on headache intensity:
Investigation using ecological momentary assessment technique. Cephalalgia,
26(11), 1335-1343. doi:10.1111/j.1468-2982.2006.01221.x
Kim, S. H., Ryoo, D. W., & Bae, C. (2007, 22-26 Aug. 2007). Adaptive Noise Cancellation Using
Accelerometers for the PPG Signal from Forehead. Paper presented at the 2007 29th
Annual International Conference of the IEEE Engineering in Medicine and Biology
King, W. C., Li, J., Leishear, K., Mitchell, J. E., & Belle, S. H. (2011). Determining activity
monitor wear time: An influential decision rule. J Phys Act Health, 8(4), 566-580.
Koenig, J., Loerbroks, A., Jarczok, M. N., Fischer, J. E., & Thayer, J. F. (2016). Chronic Pain and
Heart Rate Variability in a Cross-Sectional Occupational Sample: Evidence for
Impaired Vagal Control. Clin J Pain, 32(3), 218-225.
Kravitz, E., Moore, M. E., & Glaros, A. (1981). Paralumbar muscle activity in chronic low back
pain. Arch Phys Med Rehabil, 62(4), 172-176.
Kunze, K., & Lukowicz, P. (2008). Dealing with sensor displacement in motion-based onbody
activity recognition systems Proc. of the 10th Int'l Conf. on Ubiquitous Computing
(pp. 20-29): ACM Press.
Labs, T. (2016). Myo Armband. Retrieved from
Lalloo, C., Kumbhare, D., Stinson, J. N., & Henry, J. L. (2014). Pain-QuILT: clinical feasibility of
a web-based visual pain assessment tool in adults with chronic pain. J Med Internet
Res, 16(5), e127. doi:10.2196/jmir.3292
Le Masurier, G. C., & Tudor-Locke, C. (2003). Comparison of pedometer and accelerometer
accuracy under controlled conditions. Medicine & Science in Sports & Exercise, 35(5),
Lee, J., Matsumura, K., Yamakoshi, K. i., Rolfe, P., Tanaka, S., & Yamakoshi, T. (2013, 3-7 July
2013). Comparison between red, green and blue light reflection
photoplethysmography for heart rate monitoring during motion. Paper presented at
the 2013 35th Annual International Conference of the IEEE Engineering in Medicine
and Biology Society (EMBC).
Lim, C. L., Rennie, C., Barry, R. J., Bahramali, H., Lazzaro, I., Manor, B., & Gordon, E. (1997).
Decomposing skin conductance into tonic and phasic components. International
Journal of Psychophysiology, 25(2), 97-109. doi:
Lin, Y., Wang, L., Xiao, Y., Urman, R. D., Dutton, R., & Ramsay, M. (2018). Objective Pain
Measurement based on Physiological Signals. Proceedings of the International
Symposium on Human Factors and Ergonomics in Health Care, 7(1), 240-247.
Liu, B., Yu, M., Graubard, B. I., Troiano, R. P., & Schenker, N. (2016). Multiple imputation of
completely missing repeated measures data within person from a complex sample:
application to accelerometer data in the National Health and Nutrition Examination
Survey. Stat Med, 35(28), 5170-5188. doi:10.1002/sim.7049
Liu, D., Peng, F., Shea, A., & Picard, R. (2017). DeepFaceLIFT: Interpretable Personalized
Models for Automatic Estimation of Self-Reported Pain. arXiv preprint
Long, C., & Brown, M. E. (1964). Electromyographic Kinesiology of the Hand: Muscles
Moving the Long Finger. J Bone Joint Surg Am, 46, 1683-1706.
Maeda, Y., Sekine, M., & Tamura, T. (2011). Relationship Between Measurement Site and
Motion Artifacts in Wearable Reflected Photoplethysmography. Journal of Medical
Systems, 35(5), 969-976. doi:10.1007/s10916-010-9505-0
Martínez-Lavín, M., Hermosillo, A. G., Rosas, M., & Soto, M.-E. (1998). Circadian studies of
autonomic nervous balance in patients with fibromyalgia: A heart rate variability
analysis. Arthritis & Rheumatism, 41(11), 1966-1971. doi:10.1002/1529-
Matsumura, K., Rolfe, P., Lee, J., & Yamakoshi, T. (2014). iPhone 4s Photoplethysmography:
Which Light Color Yields the Most Accurate Heart Rate and Normalized Pulse Volume
Using the iPhysioMeter Application in the Presence of Motion Artifact? PloS one,
9(3), e91205. doi:10.1371/journal.pone.0091205
May, M., Junghaenel, D. U., Ono, M., Stone, A. A., & Schneider, S. (2018). Ecological
momentary assessment methodology in chronic pain research: A systematic review.
J Pain. doi:10.1016/j.jpain.2018.01.006
McMillan, D., Brown, B., Lampinen, A., McGregor, M., Hoggan, E., & Pizza, S. (2017).
Situating Wearables: Smartwatch Use in Context. Paper presented at the
Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems,
Denver, Colorado, USA.
Mehl, M. R., & Conner, T. S. (Eds.). (2011). Handbook of Research Methods for Studying
Daily Life. New York, NY: Guilford Publications, Inc.
Melzack, R. (1975). The McGill Pain Questionnaire: major properties and scoring methods.
Pain, 1(3), 277-299.
Merskey, H. (2007). The taxonomy of pain. Med Clin North Am, 91(1), 13-20, vii.
Miller, J. Accelerometer technologies, specfications, and limitations. Retrieved from
Millikan, G. A. (1942). The Oximeter, an Instrument for Measuring Continuously the Oxygen
Saturation of Arterial Blood in Man. Review of Scientific Instruments, 13(10), 434-
444. doi:10.1063/1.1769941
Morren, M., van Dulmen, S., Ouwerkerk, J., & Bensing, J. (2009). Compliance with
momentary pain measurement using electronic diaries: a systematic review. Eur J
Pain, 13(4), 354-365. doi:10.1016/j.ejpain.2008.05.010
Moseley, G. L., Nicholas, M. K., & Hodges, P. W. (2004). Does anticipation of back pain
predispose to back trouble? Brain, 127(10), 2339-2347. doi:10.1093/brain/awh248
Mujahid, S., Sierra, G., Abdalkareem, R., Shihab, E., & Shang, W. (2018). An empirical study
of Android Wear user complaints. Empirical Software Engineering.
Munguia Tapia, E. (2003). Activity Recognition in the Home Setting Using Simple and
Ubiquitous Sensors. (S.M. Thesis), Massachusetts Institute of Technology,
Nahum-Shani, I., Smith, S. N., Spring, B. J., Collins, L. M., Witkiewitz, K., Tewari, A., &
Murphy, S. A. (2016). Just-in-time adaptive interventions (JITAIs) in mobile health:
Key components and design principles for ongoing health behavior support. Ann
Behav Med(1-17). doi:10.1007/s12160-016-9830-8
Nakamura, M., Nishiwaki, Y., Ushida, T., & Toyama, Y. (2011). Prevalence and characteristics
of chronic musculoskeletal pain in Japan. J Orthop Sci, 16(4), 424-432.
Nicholson, B. (2000). Taxonomy of pain. Clin J Pain, 16(3 Suppl), S114-117.
Niijima, A., Isezaki, T., Aoki, R., Watanabe, T., & Yamada, T. (2017). hitoeCap: wearable EMG
sensor for monitoring masticatory muscles with PEDOT-PSS textile electrodes. Paper
presented at the Proceedings of the 2017 ACM International Symposium on
Wearable Computers, Maui, Hawaii.
Nurminen, M. L., Niittynen, L., Korpela, R., & Vapaatalo, H. (1999). Coffee, caffeine and
blood pressure: a critical review. Eur J Clin Nutr, 53(11), 831-839.
Ohnhaus, E. E., & Adler, R. (1975). Methodological problems in the measurement of pain: A
comparison between the verbal rating scale and the visual analogue scale. Pain, 1(4),
379-384. doi:
Olsen, R. B., Bruehl, S., Nielsen, C. S., Rosseland, L. A., Eggen, A. E., & Stubhaug, A. (2013).
Hypertension prevalence and diminished blood pressure-related hypoalgesia in
individuals reporting chronic pain in a general population: the Tromso study. Pain,
154(2), 257-262. doi:10.1016/j.pain.2012.10.020
Parak, J., & Korhonen, I. (2014, 26-30 Aug. 2014). Evaluation of wearable consumer heart
rate monitors based on photopletysmography. Paper presented at the Engineering in
Medicine and Biology Society (EMBC), 2014 36th Annual International Conference of
the IEEE.
Paruthi, G., Raj, S., Baek, S., Wang, C., Huang, C.-c., Chang, Y.-J., & Newman, M. W. (2018).
Heed: Exploring the Design of Situated Self-Reporting Devices. Proc. ACM Interact.
Mob. Wearable Ubiquitous Technol., 2(3), 1-21. doi:10.1145/3264942
Paruthi, G., Raj, S., Gupta, A., Huang, C.-C., Chang, Y.-J., & Newman, M. W. (2017). HEED:
situated and distributed interactive devices for self-reporting. Paper presented at the
Proceedings of the 2017 ACM International Joint Conference on Pervasive and
Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium
on Wearable Computers, Maui, Hawaii.
Patterson, J. A. C., McIlwraith, D. C., & Yang, G. Z. (2009, 3-5 June 2009). A Flexible, Low
Noise Reflective PPG Sensor Platform for Ear-Worn Heart Rate Monitoring. Paper
presented at the 2009 Sixth International Workshop on Wearable and Implantable
Body Sensor Networks.
Peters, M. L., & Schmidt, A. J. (1991). Psychophysiological responses to repeated acute pain
stimulation in chronic low back pain patients. J Psychosom Res, 35(1), 59-74.
Pizza, S., Brown, B., McMillan, D., & Lampinen, A. (2016). Smartwatch in vivo. Paper
presented at the Proceedings of the 2016 CHI Conference on Human Factors in
Computing Systems, San Jose, California, USA.
Poh, M. Z., Swenson, N. C., & Picard, R. W. (2010). A wearable sensor for unobtrusive, long-
term assessment of electrodermal activity. IEEE Transactions on Bio-Medical
Engineering, 57(5), 1243-1252. doi:10.1109/TBME.2009.2038487
Ponnada, A., Haynes, C., Maniar, D., Manjourides, J., & Intille, S. (2017). Microinteraction
ecological momentary assessment response rates: Effect of microinteractions or the
smartwatch? Proc. of the ACM Journal on Interactive, Mobile, Wearable, and
Ubiquitous Technology, 1(3).
Prahlow, N. D., & Buschbacher, R. M. (2003). An introduction to electromyography: an
invited review. J Long Term Eff Med Implants, 13(4), 289-307.
Salaffi, F., Sarzi-Puttini, P., & Atzeni, F. (2015). How to measure chronic pain: New concepts.
Best Pract Res Clin Rheumatol, 29(1), 164-186. doi:10.1016/j.berh.2015.04.023
San-Segundo, R., Blunck, H., Moreno-Pimentel, J., Stisen, A., & Gil-Martín, M. (2018). Robust
human activity recognition using smartwatches and smartphones. Engineering
Applications of Artificial Intelligence, 72, 190-202.
Sandoval, A. E. (2010). Electrodiagnostics for low back pain. Phys Med Rehabil Clin N Am,
21(4), 767-776. doi:10.1016/j.pmr.2010.06.007
Scerbo, A. S., Freedman, L. W., Raine, A., Dawson, M. E., & Venables, P. H. (1992). A major
effect of recording site on measurement of electrodermal activity. Psychophysiology,
29(2), 241-246.
Schneider, S., Stone, A. A., Schwartz, J. E., & Broderick, J. E. (2011). Peak and end effects in
patients' daily recall of pain and fatigue: a within-subjects analysis. J Pain, 12(2), 228-
235. doi:10.1016/j.jpain.2010.07.001
Scollon, C. N., Kim-Prieto, C., & Diener, E. (2003). Experience sampling: Promises and pitfalls,
strengths and weaknesses. Journal of Happiness Studies, 4, 5-34.
Shiffman, S., Paty, J. a., Gnys, M., Kassel, J. a., & Hickcox, M. (1996). First lapses to smoking:
within-subjects analysis of real-time reports. Journal of consulting and clinical
psychology, 64, 366-379.
Smith, A. (2017). Record shares of Americans now own smartphones, have home
broadband. Retrieved from
Soderberg, G. L., & Barr, J. O. (1983). Muscular function in chronic low-back dysfunction.
Spine (Phila Pa 1976), 8(1), 79-85.
Sokolovsky, A. W., Mermelstein, R. J., & Hedeker, D. (2013). Factors predicting compliance
to ecological momentary assessment among adolescent smokers. Nicotine &
Tobacco Research, 16(3), 351-358.
Sokolovsky, A. W., Mermelstein, R. J., & Hedeker, D. (2014). Factors Predicting Compliance
to Ecological Momentary Assessment Among Adolescent Smokers. Nicotine &
Tobacco Research, 16(3), 351-358. doi:10.1093/ntr/ntt154
Statistica. (2018). Smartphone user penetration as percentage of total global population
from 2014 to 2021. Retrieved from
Steglitz, J., Buscemi, J., & Ferguson, M. J. (2012). The future of pain research, education, and
treatment: a summary of the IOM report "Relieving pain in America: a blueprint for
transforming prevention, care, education, and research". Transl Behav Med, 2(1), 6-
8. doi:10.1007/s13142-012-0110-2
Stewart, W. F., Ricci, J. A., Chee, E., Morganstein, D., & Lipton, R. (2003). Lost productive
time and cost due to common pain conditions in the us workforce. JAMA, 290(18),
2443-2454. doi:10.1001/jama.290.18.2443
Stone, A. A., & Broderick, J. E. (2007). Real-time data collection for pain: appraisal and
current status. Pain Med, 8 Suppl 3, S85-93. doi:10.1111/j.1526-4637.2007.00372.x
Stone, A. A., & Shiffman, S. (1994). Ecological momentary assessment: Measuring real world
processes in behavioral medicine. Annals of Behavioral Medicine, 16, 199-202.
Storella, R. J., Shi, Y., O'Connor, D. M., Pharo, G. H., Abrams, J. T., & Levitt, J. (1999). Relief of
Chronic Pain May Be Accompanied by an Increase in a Measure of Heart Rate
Variability. Anesthesia & Analgesia, 89(2), 448-450. doi:10.1213/00000539-
Tapia, E. M., Intille, S. S., & Larson, K. (2007). Portable wireless sensors for object usage
sensing in the home: Challenges and practicalities Proceedings of the European
Ambient Intelligence Conference 2007 (Vol. LNCS 4794, pp. 19-37). Berlin Heidelberg:
Treede, R.-D., Rief, W., Barke, A., Aziz, Q., Bennett, M. I., Benoliel, R., . . . Wang, S.-J. (2015).
A classification of chronic pain for ICD-11. Pain, 156(6), 1003-1007.
Truong, K. N., Shihipar, T., & Wigdor, D. J. (2014). Slide to X: unlocking the potential of
smartphone unlocking. Paper presented at the Proceedings of the SIGCHI Conference
on Human Factors in Computing Systems, Toronto, Ontario, Canada.
Tryon, W. W., & Williams, R. (1996). Fully proportional actigraphy: A new instrument.
Behavior Research Methods, Instruments, & Computers, 28(3), 392-403.
Tudor-Locke, C., Ainsworth, B. E., Thompson, R. W., & Matthews, C. E. (2002). Comparison
of pedometer and accelerometer measures of free-living physical activity. Medicine
& Science in Sports & Exercise, 34(12), 2045-2051.
Van Damme, B., Stevens, V., Perneel, C., Van Tiggelen, D., Neyens, E., Duvigneaud, N., . . .
Danneels, L. (2014). A surface electromyography based objective method to identify
patients with nonspecific chronic low back pain, presenting a flexion related
movement control impairment. J Electromyogr Kinesiol, 24(6), 954-964.
van Dooren, M., de Vries, J. J., & Janssen, J. H. (2012). Emotional sweating across the body:
Comparing 16 different skin conductance measurement locations. Physiology &
behavior, 106(2), 298-304. doi:10.1016/j.physbeh.2012.01.020
van Wel, L., Huss, A., Bachmann, P., Zahner, M., Kromhout, H., Frohlich, J., & Vermeulen, R.
(2017). Context-sensitive ecological momentary assessments; integrating real-time
exposure measurements, data-analytics and health assessment using a smartphone
application. Environ Int, 103, 8-12. doi:10.1016/j.envint.2017.03.016
Vizbara, V. (2013). Comparison of green, blue and infrared light in wrist and forehead
photoplethysmography. BIOMEDICAL ENGINEERING 2016, 17(1).
Wan, J., O’Grady, M. J., & O’Hare, G. M. P. (2014). Dynamic sensor event segmentation for
real-time activity recognition in a smart home context. Pers. Ubiquit. Comput.
Wang, R., Blackburn, G., Desai, M., Phelan, D., Gillinov, L., Houghtaling, P., & Gillinov, M.
(2017). Accuracy of Wrist-Worn Heart Rate Monitors. JAMA Cardiol, 2(1), 104-106.
Waterman, C., Victor, T. W., Jensen, M. P., Gould, E. M., Gammaitoni, A. R., & Galer, B. S.
(2010). The assessment of pain quality: an item response theory analysis. J Pain,
11(3), 273-279. doi:10.1016/j.jpain.2009.07.014
Wen, C. K. F., Schneider, S., Stone, A. A., & Spruijt-Metz, D. (2017). Compliance with mobile
Ecological Momentary Assessment protocols in children and adolescents: A
systematic review and meta-analysis. J Med Internet Res, 19(4), e132.
Wendelken, S., McGrath, S., Blike, G., & Akay, M. (2004, 17-18 April 2004). The feasibility of
using a forehead reflectance pulse oximeter for automated remote triage. Paper
presented at the IEEE 30th Annual Northeast Bioengineering Conference, 2004.
Proceedings of the.
Williams, A. C. d. C., & Craig, K. D. (2016). Updating the definition of pain. Pain, 157(11),
2420-2423. doi:10.1097/j.pain.0000000000000613
Yemm, R. (1969). Masseter muscle activity in stress: adaptation of response to a repeated
stimulus in man. Arch Oral Biol, 14(12), 1437-1439.
Yurtman, A., & Barshan, B. (2017). Activity Recognition Invariant to Sensor Orientation with
Wearable Motion Sensors. Sensors, 17(8). doi:10.3390/s17081838
Table 2.
Classification of chronic pain (Treede et al., 2015)
Pain classification
Chronic primary pain
Chronic pain in one or more regions of the body, persistent for 3
or more months. It is often associated with emotional distress and
functional disability in everyday life. Common chronic pain
examples include back pain, fibromyalgia, and irritable bowel
syndrome, which cannot be explained using other categories and
little is known about their causes/roots.
Chronic cancer pain
Chronic pain caused by cancer or cancer treatments (e.g., surgery
and chemotherapy).
Chronic post-traumatic pain
Chronic pain that persists even after 3 months of a treatment of an
injury or surgery.
Chronic neuropathic pain
Chronic pain caused due to a lesion or disease of the
somatosensory nervous system. Neuropathic pain can be
spontaneous or evoked and an intense response to a painful
stimulus or a painful response to a normal/unpainful stimulus.
Chronic headache/orofacial
Chronic pain experienced as headache or orofacial pain (e.g.,
temporomandibular disorders) at least 50% of the time for three
months or longer.
Chronic visceral pain
Chronic pain originating from the internal organs of head, neck
and thoracic, abdominal, and pelvic cavities. This pain is felt on
the somatic tissues of skin and muscles that receive the same
sensory input as internal organs.
Chronic musculoskeletal pain
Chronic pain resulting from a disease and directly affecting joints,
bones, muscles, soft tissues. It includes chronic pain resulting
from rheumatoid arthritis or structural changes affecting joints and
muscles such as symptomatic osteoarthrosis.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
The upcoming 11th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD) of the World Health Organization (WHO) offers a unique opportunity to improve the representation of painful disorders. For this purpose, the International Association for the Study of Pain (IASP) has convened an interdisciplinary task force of pain specialists. Here, we present the case for a reclassification of nervous system lesions or diseases associated with persistent or recurrent pain for ≥3 months. The new classification lists the most common conditions of peripheral neuropathic pain: trigeminal neuralgia, peripheral nerve injury, painful polyneuropathy, postherpetic neuralgia, and painful radiculopathy. Conditions of central neuropathic pain include pain caused by spinal cord or brain injury, poststroke pain, and pain associated with multiple sclerosis. Diseases not explicitly mentioned in the classification are captured in residual categories of ICD-11. Conditions of chronic neuropathic pain are either insufficiently defined or missing in the current version of the ICD, despite their prevalence and clinical importance. We provide the short definitions of diagnostic entities for which we submitted more detailed content models to the WHO. Definitions and content models were established in collaboration with the Classification Committee of the IASP's Neuropathic Pain Special Interest Group (NeuPSIG). Up to 10% of the general population experience neuropathic pain. The majority of these patients do not receive satisfactory relief with existing treatments. A precise classification of chronic neuropathic pain in ICD-11 is necessary to document this public health need and the therapeutic challenges related to chronic neuropathic pain.
Full-text available
Wearable apps are becoming increasingly popular in recent years. Nevertheless, to date, very few studies have examined the issues that wearable apps face. Prior studies showed that user reviews contain a plethora of insights that can be used to understand quality issues and help developers build better quality mobile apps. Therefore, in this paper, we mine user reviews in order to understand the user complaints about wearable apps. We manually sample and categorize 2,667 reviews from 19 Android wearable apps. Additionally, we examine the replies posted by developers in response to user complaints. This allows us to determine the type of complaints that developers care about the most, and to identify problems that despite being important to users, do not receive a proper response from developers. Our findings indicate that the most frequent complaints are related to Functional Errors, Cost, and Lack of Functionality, whereas the most negatively impacting complaints are related to Installation Problems, Device Compatibility, and Privacy & Ethical Issues. We also find that developers mostly reply to complaints related to Privacy & Ethical Issues, Performance Issues, and notification related issues. Furthermore, we observe that when developers reply, they tend to provide a solution, request more details, or let the user know that they are working on a solution. Lastly, we compare our findings on wearable apps with the study done by Khalid et al. (2015) on handheld devices. From this, we find that some complaint types that appear in handheld apps also appear in wearable apps; though wearable apps have unique issues related to Lack of Functionality, Installation Problems, Connection & Sync, Spam Notifications, and Missing Notifications. Our results highlight the issues that users of wearable apps face the most, and the issues to which developers should pay additional attention to due to their negative impact.
In-situ self-reporting is a widely used data collection technique for understanding people's behavior in context. Characteristics of smartphones such as their high proliferation, close proximity to their users, and heavy use have made them a popular choice for applications that require frequent self-reporting. Newer device categories such as wearables and voice assistants offer their own advantages, providing an opportunity to explore a wider range of self-reporting approaches. In this paper, we focus on exploring the design space of Situated Self-Reporting (SSR) devices. We present the Heed system, consisting of simple, low-cost, and low-power SSR devices that are distributed in the environment of the user and can be appropriated for reporting measures such as stress, sleepiness, and activities. In two real-world studies with 10 and 7 users, we compared and analyzed the use of smartphone and Heed devices to uncover differences in their use due to the influence of factors such as situational and social context, notification types, and physical design. Our findings show that Heed devices complemented smartphones in the coverage of activities, locations and interaction preferences. While the advantage of Heed was its single-purpose and dedicated location, smartphones provided mobility and flexibility of use.
In an attempt to understand human physiological signals when an individual is subjected to pain, we set up a tonic pain experiment in a laboratory setting. The subjects’ physiological signals were recorded, timestamped, and compared to an initial 30 second baseline measurement. Subjects were also asked to verbally state their level of pain based on a visual analog scale in order to compare reported pain levels with physiological signals. The physiological signals measured were: Electroencephalography (EEG), Pupillary Unrest Under Ambient Light (PUAL), Skin Conductance (SC), Electromyography (EMG), Respiration Rate (RR), Blood Volume Pulse (BVP), Skin Temperature (ST), Blood Pressure (BP), and Facial Expression (FE). ANOVA and frequency domain analyses were conducted on the data in order to determine whether there was a significant difference between the ‘pain’ and ‘no pain’ (baseline) states of an individual. Based on our results, skin conductance, PUAL, facial expression, and EEG signals were theorized to be good signals for the classification of tonic pain, or any pain applied directly to an individual.
Smart user devices are becoming increasingly ubiquitous and useful for detecting the user’s context and his/her current activity. This work analyzes and proposes several techniques to improve the robustness of a Human Activity Recognition (HAR) system that uses accelerometer signals from different smartwatches and smartphones. This analysis reveals some of the challenges associated with both device heterogeneity and the different use of smartwatches compared to smartphones. When using smartwatches to recognize whole body activities, the arm movements introduce additional variability giving rise to a significant degradation in HAR. In this analysis, we describe and evaluate several techniques which successfully address these challenges when using smartwatches and when training and testing with different devices and/or users.
This study aims to explore usability issues of watch-type wearable devices and to suggest guidelines for improved operation of smartwatches. To do so, we conducted a series of surveys, interviews, and task performance experiments. Thirty smartwatch users from ages 20 to 43 years were recruited. Users’ experiences of smartwatches were collected via a weeklong, online-based diary study, which consisted of various tasks to be completed while smartwatches were in use. Our study assessed usability problems associated with those tasks, concurrent tasks conducted while interacting with smartwatches, pain points/discomfort that users had while interacting with their devices, and requirements/requests of the smartwatch users. During the week of tracking, participants were asked to complete the usability evaluation three times a day using usability principles we designed for the study: information display, control, learnability, interoperability, and preference. In addition, task performance tests were conducted for the tasks most frequently conducted on touch-based displays: number entry, swiping, and scrolling. Specific usability issues of smartwatches were identified and summarized for each usability principle by triangulating survey, interview, and task performance evaluation results. Based on the insights from the results of the study, we conclude by suggesting guidelines for further enhancing users’ experience of future smartwatches.
Self-reported pain intensity assessments are central to chronic pain research. Ecological Momentary Assessment (EMA) methodologies are uniquely positioned to collect these data, and are indeed being utilized in the field. However, EMA protocols are complex, and many decisions are necessary in the design of EMA research studies. A systematic literature review identified 105 articles drawing from 62 quantitative EMA research projects examining pain intensity in adult chronic pain patients. Study characteristics were tabulated in order to summarize and describe the use of EMA, with an emphasis placed on various dimensions of decision-making involved in executing EMA methodologies. Most identified studies considered within-person relationships between pain and other variables, and a few examined interventions on chronic pain. There was a trend toward the use of smartphones as EMA data collection devices more recently, and completion rates were not reported in nearly one-third of studies. Pain intensity items varied widely with respect to number of scale points, anchor labels, and length of reporting period; most used numeric rating scales. Recommendations are provided for reporting to improve reproducibility, comparability, and interpretation of results, and for opportunities to clarify the importance of design decisions. Perspective Studies that utilize Ecological Momentary Assessment methodologies to assess pain intensity are heterogeneous. Aspects of protocol design, including data input modality and pain item construction, have the potential to influence the data collected. Thorough reporting on design features and completion rates therefore facilitates reproducibility, comparability, and interpretation of study results.
Purpose: The present study examined various accelerometer nonwear definitions and their impact on detection of sedentary time using different ActiGraph models, filters, and axes. Methods: In total, 61 youth (34 children and 27 adolescents; aged 5-17 y) wore a 7164 and GT3X+ ActiGraph on a hip-worn belt during a 90-minute structured sedentary activity. Data from GT3X+ were downloaded using the Normal filter (N) and low-frequency extension (LFE), and vertical axis (V) and vector magnitude (VM) counts were examined. Nine nonwear definitions were applied to the 7164 model (V), GT3X+LFE (V and VM), and GT3X+N (V and VM), and sedentary estimates were computed. Results: The GT3X+LFE-VM was most sensitive to movement and could accurately detect observed sedentary time with the shortest nonwear definition of 20 minutes of consecutive "0" counts for children and 40 minutes for adolescents. The GT3X+N-V was least sensitive to movement and required longer definitions to detect observed sedentary time (40 min for children and 90 min for adolescents). VM definitions were 10 minutes shorter than V definitions. LFE definitions were 40 minutes shorter than N definitions in adolescents. Conclusion: Different nonwear definitions are needed for children and adolescents and for different model-filter-axis types. Authors need to consider nonwear definitions when comparing prevalence rates of sedentary behavior across studies.
Mobile-based ecological-momentary-assessment (EMA) is an in-situ measurement methodology where an electronic device prompts a person to answer questions of research interest. EMA has a key limitation: interruption burden. Microinteraction-EMA(µEMA) may reduce burden without sacrificing high temporal density of measurement. In µEMA, all EMA prompts can be answered with ‘at a glance' microinteractions. In a prior 4-week pilot study comparing standard EMA delivered on a phone (phone-EMA) vs. µEMA delivered on a smartwatch (watch-µEMA), watch-µEMA demonstrated higher response rates and lower perceived burden than phone-EMA, even when the watch-µEMA interruption rate was 8 times more than phone-EMA. A new 4-week dataset was gathered on smartwatch-based EMA (i.e., watch-EMA with 6 back-to-back, multiple-choice questions on a watch) to compare whether the high response rates of watch-µEMA previously observed were a result of using microinteractions, or due to the novelty and accessibility of the smartwatch. No statistically significant differences in compliance, completion, and first-prompt response rates were observed between phone-EMA and watch-EMA. However, watch-µEMA response rates were significantly higher than watch-EMA. This pilot suggests that (1) the high compliance and low burden previously observed in watch-µEMA is likely due to the microinteraction question technique, not simply the use of the watch versus the phone, and that (2) compliance with traditional EMA (with long surveys) may not improve simply by moving survey delivery from the phone to a smartwatch.
Smartwaches are the representative wearable or body-worn devices that provide convenient and easy information access. There is a growing body of research work on enabling novel interaction techniques and understanding user experiences of smartwatches. However, there is still lack of user experience research on wearing behaviors of smartwatches, which is critical for wearable device and service design. In this work, we investigate how college students wear smartwatches and what factors affect wearing behaviors by analyzing a longitudinal activity dataset collected from 50 smartwatch users for 203 days. Our results show that there are several temporal usage patterns and distinct groups of usage patterns. The factors affecting wearing behaviors are contextual, nuanced, and multifaceted. Our findings provide diverse design implications for improving wearability of smartwatches and leveraging smartwatches for behavioral changes.