Wearable technology for monitoring behavioral and physiological responses in children with autism spectrum disorder: A literature review

3.1.1 Head-mounted devices

A variety of head-mounted devices were described, including glasses or devices worn in the form of a headband. These WTs were used to gather data such as head orientation, gaze patterns, facial engagement, emotion recognition, and direct interaction.

Smart or assistive reality glasses were discussed by six studies (see Table 1). These features of glasses were, an outward-facing camera capturing the faces of others, and an inward-facing tracker monitoring the wearer’s eye gaze patterns and was capable of providing youth with real-time feedback on their social function through providing audio or visual feedback. Several systems utilizing smart glass technology were presented to support and improve social communication and interaction in children with ASD.

A study used superpower glass intervention, a wearable social learning aid for children with ASD to encourage facial engagement and provide feedback to the child during social interactions at home. A computer vision system runs on Google Glass that is wirelessly connected to a smartphone app to provide the intervention to the child wearing the glasses [13]. Another pilot study evaluated the potential of the Superpower Glass prototype as a wearable therapy intervention that increases social skills, facial affect recognition, and eye contact for children with autism between the ages of 3 and 17 years [10]. Daniels J et al created software capable of running on Google Glass and designed to give preprogrammed emotion-feedback cues to the wearer [11]. A study introduced the wearable device with infrared (IR) communication, called FaceLooks, which quantifies face-to-face behavior. This smart sensor device was worn on the forehead like a headband and emits light to visually represent real-time face-to-face behavior [28]. A similar concept was employed by Hurwitz S. et al. by using Video Recording Glasses. During the conversations, both conversational partners (i.e., participants and interventionists) wore special glasses that contained a small camera in the bridge. These glasses captured video and audio using direct first-person point-of-view (PoV) wearable and wireless video cameras (Pivothead, Denver, CO). During conversations, the glasses recorded what each person saw and said from the viewpoint of the person wearing the glasses (i.e., two videos were recorded during each conversation, one from the perspective of the participant and one from the interventionist) [27]. Another study used the Empowered Brain, an assistive technology tool designed to help improve social communication in children and adults with ASD. The user wore the smart glasses and participates in a semi-structured interaction with a facilitator. Smart glasses provide users with pro-social cues and guidance by visual and auditory feedback delivered through the optical display and speaker [20].

Overall, studies investigating smart glasses suggested that they were generally tolerated by autistic youth, and were perceived as useable [10, 11, 13] and enjoyable [20, 28]. Some studies, however, reported that smart glasses may become uncomfortable if worn for long periods, or become too warm [27]. When examining the potential for negative effects, evidence was found suggesting that a small number of participants experienced mild adverse effects including dizziness, nose-bridge discomfort, and eye strain [27].

3.1.2 Accessory or body-worn devices

Accessory included wrist-worn devices [12, 15, 16, 17, 21, 23] and sensorised clothes [25]. Wrist-worn devices were examined in six studies (Table 1). These devices measured physiological signals and functional parameters including touch [23], heart rate [12, 23], and electrodermal activity (EDA) [16, 17].

One study used a smart watch for directive-following and anticipatory directive-following responses during the intervention [23]. Another study used a smartwatch to evaluate the heart rate as high values of heart rate in these children with ASD are more likely to be caused by a stress or temper episode. Moreover, as these episodes are emotional, changes in heart rate are remarkable quantitatively [12]. A study by Goodwin in 2019 recorded peripheral physiological (cardiovascular and electrodermal activity) and motion (accelerometry) signals from a biosensor worn on the wrist [17]. Participants wore a bracelet for wireless EDA recording for the measurement of anxiety levels [16]. Another study utilized wearable sensors for leg movement data collection from five high-risk infants at 3, 6, 9, and 12 months of age. The protocol included visitation of infants each morning in their home environment intending to collect full-day movement data. The wearable sensors were placed on the infants’ ankles and were attached using legwarmers. Each ankle legwarmer contains a pocket to hold the wearable sensor in place [25].

Another study by Suarez used electrodermal activity data collection during laboratory food presentation protocol using Biopac System software. Electrodes were placed on the thenar, and hypothenar eminences of the participant’s non-dominant hand and data was collected with a wireless transmitter using a constant voltage (0.5V) technique [18]. Another study by Sasikumar also used photoplethysmography (PPG) and the galvanic skin response (GSR) signals during mental tasks through clue cards with visuals like alphabets, numbers, fruits, colors, and animals. Physiological signals processed, and dicrotic notch, heart rate, and EDA response feature were extracted from the acquired signals [30].

Body-worn technologies included sensors placed on the body, such as accelerometers, electromyography sensors, and chest belts or sensors. Four articles (Table 1) investigated the use of accelerometers to measure and classify stereotypical motor movements such as body rocking and flapping [14, 15, 17, 21] with two studies also investigating accelerometers to detect emotional self-regulation behavior [12, 21]. Few studies also used accelerometers to detect stereotypical movement and motion during the task [14, 15].

Wearable sensors to record ECG signals and to provide markers for ANS response were used by three studies [14, 19, 29]. Electrocardiography (ECG) signals were recorded during an eye-tracking joint attention task in toddlers and attention to the stimuli was assured by the simultaneous acquisition of eye-tracking and ECG data. The ECG sensor was modified by adding two pins to allow its interfacing with the common PolarTM cardio-fitness chest strap, featuring two dry electrodes in the inner side of the strap, interfaced to the skin surface of the subject for a single-lead acquisition [29]. Another wearable sensor device used was a low-cost system that integrated a three-lead electrocardiograph (ECG) sensor for cardiac activity monitoring and feasibility testing suggested that the use of chest-belt to measure ECG in a semi naturalistic context may enable real-time tracking of engagement and autonomic nervous system activity [14, 19]. Despite some limitations, including children’s tolerability of the device and comfort, the device effectively measured proximity, was useable in everyday life, and resulted in autistic children modifying their proximity to others appropriately [14, 29].

3.2.1 Eye trackers

Individuals diagnosed with ASD have been observed to exhibit different gaze patterns that are evident as they look at socially salient stimuli such as faces [10, 11, 13, 29]. This is a consequence of impaired social communication and interactive skills, where making eye contact is particularly challenging. It must be noted that the studies reported in this section were conducted in a naturalistic classroom or home setting [10, 13, 28] and clinical settings [11, 20, 27] which tends to provide more obvious results. Most of these studies were designed for early ASD screening, while some were aimed to facilitate intervention and improvement in social and emotional behaviors. Driven by the availability of general use eye gaze trackers either head-mounted or eye-tracking smart glasses, there are more contributions in this category as compared to others [6].

3.2.2 Motion detectors

Children with ASD are known to engage in stereotypical behaviors, which inhibit the development of appropriate social and adaptive behaviors and can turn into self-injurious activities [6]. A mechanism for detecting such behavior gives caregivers the opportunity for timely intervention. In most of the studies, the movement was tracked through an accelerometer [14, 15, 17, 25]. Accelerometers are devices that measure the acceleration of motion of a structure. With acceleration data being collected in the x, y, and z coordinates, these can be processed further to obtain velocity and displacement for time. Furthermore, the raw signals can be converted into machine learning methods to extract meaningful measures. As wearable devices, the sensors are required to be in physical contact with the subjects’ bodies either worn in a bodysuit or as a smart watch.

A study Utilized the wearable sensors, leg movement data collected from five HR infants at 3, 6, 9, and 12 months of age. The protocol included visitation of infants each morning in their home environment to collect full-day movement data. The wearable sensors were placed on the infants’ ankles and were attached using legwarmers. Each ankle legwarmer contains a pocket to hold the wearable sensor in place. Using the full-day movement data derived from the sensor, a novel motion complexity measure was constructed. A minimally complex movement pattern is one fully described by a singular periodic function [25].

3.2.3 Heart rate (HR) and Heart rate variability (HRV)

Heart rate (HR) is a measure of the number of beats per minute and can be used for monitoring the cardiovascular activity of the heart. However, heart rate gives limited information about heart activity. A more useful measure is called heart rate variability (HRV) which calculates the time intervals between two consecutive R peaks in the ECG signal is measured from an ECG sensor. The gold standard for measuring and calculating HRV is typically used is which is widely recognized and used by physiology researchers and clinicians [34]. However, in the recent developments of wearable devices, a new class of devices that measure heart rate and heart rate variability has emerged, using a photoplethysmography (PPG) sensor on wristbands. PPG sensor relies on a different principle the measurement of Blood Volume Pulse (BVP) to estimate the heart rate and approximated heart rate variability value. Although this seems to provide fairly accurate heart rate measurements, its heart rate variability recording is not as accurate as an ECG sensor and the decision to trust HRV analysis using PPG depends on the parameter one intends to use [30].

The increased HR found in ASD children seems to confirm that they experience stress during all interactive activities, with a low capability to modulate the emotional response process. HR was more elevated and less modulated in children with ASD as compared to typically developing children [19].

3.2.4 Respiratory sinus arrhythmia

Another parameter that is closely related to Heart Rate Variability is Respiratory Sinus Arrhythmia (RSA). RSA can be defined as the heart rate variability in synchrony with respiration, by which the R-R interval on an ECG signal is reduced during inhalation and increased during exhalation [35]. RSA reacts to stress or other emotional experiences and thus can be used to characterize anxiety disorder, by revealing information pertinent to the emotion regulation process. In addition, RSA changes as a function of the stressor. For instance, it has been reported that high resting RSA amplitude is associated with a greater withdrawal of stressors and subsequent recovery could represent a flexible and adaptive physiological response system to a challenge [29]. On the other hand, low resting RSA accompanied by an inadequate reactivity to stress could indicate an inappropriate regulatory mechanism. Along with this Respiratory rate can also be a stress indicator.

3.2.5 Electrodermal activity (EDA)

Electrodermal activity (EDA) is also called Galvanic Skin Response (GSR) measures the conductance of the skin. The utility of wearable devices is now extending beyond the goal to detect and assess complex behaviors in the autistic population. Wearable, unobtrusive sensors allow the measurement of physiological data over extended periods for continuous health monitoring. Electrodermal activity monitors may be able to estimate a child’s internal state through physiological signals that are related to sweat rates, blood volume pulse, heart rate, and skin temperature in real-time. It took the form of a chest strap, wristband, ankle band, and baby sock. The sensor data was made available across devices with a wireless communication network [6].

EDA is used as a measure of stress levels in humans. When a person experiences stress, moisture collects under the skin as a sympathetic nervous system response [12, 16, 18, 30]. The increased moisture level increases the electrical conductivity of the skin, which the skin conductance sensors utilize to determine the stress levels in a subject. The same EDA results may be obtained when a person experiences excitement, highlighting the need for context evaluation. These sensors allow others to get an insight into the internal states of subjects with ASD, who might otherwise find it difficult to communicate their feelings. Skin temperature measurement allows the elimination of environmental factors that may cause an increase in body temperature, which can improve the integrity of the result. Tracking the motion of a subject can be used to detect, analyze and preempt repetitive, stereotypical behaviors, allowing the therapist to reinforce functional replacement behaviors. This can prevent the child from engaging in these behaviors to the exclusion of more adaptive behaviors and decreases the likelihood that certain behaviors escalate to self-injury [6].

3.2.6 Blood volume pulse

Blood Volume Pulse (BVP) is measured using photoplethysmography (PPG) and indicates dynamic changes in blood volume underneath the sensor. The BVP signal indicates relative changes in the vascular bed due to vasodilation or vasoconstriction (increase or decrease in blood flow) as well as to changes in the elasticity of the vascular walls. BVP is commonly used to detect heart rate, heart rate variability, and relative blood volume. In the context of stress detection, research has shown features extracted from blood volume pulse can be used to detect stress [17, 33].