Concussion Assessment With Smartglasses: Validation Study of Balance Measurement Toward a Lightweight, Multimodal, Field-Ready Platform

Abstract

BackgroundLightweight and portable devices that objectively measure concussion-related impairments could improve injury detection and critical decision-making in contact sports and the military, where brain injuries commonly occur but remain underreported. Current standard assessments often rely heavily on subjective methods such as symptom self-reporting. Head-mounted wearables, such as smartglasses, provide an emerging platform for consideration that could deliver the range of assessments necessary to develop a rapid and objective screen for brain injury. Standing balance assessment, one parameter that may inform a concussion diagnosis, could theoretically be performed quantitatively using current off-the-shelf smartglasses with an internal accelerometer. However, the validity of balance measurement using smartglasses has not been investigated.ObjectiveThis study aimed to perform preliminary validation of a smartglasses-based balance accelerometer measure (BAM) compared with the well-described and characterized waist-based BAM.MethodsForty-two healthy individuals (26 male, 16 female; mean age 23.8 [SD 5.2] years) participated in the study. Following the BAM protocol, each subject performed 2 trials of 6 balance stances while accelerometer and gyroscope data were recorded from smartglasses (Glass Explorer Edition). Test-retest reliability and correlation were determined relative to waist-based BAM as used in the National Institutes of Health’s Standing Balance Toolbox.ResultsBalance measurements obtained using a head-mounted wearable were highly correlated with those obtained through a waist-mounted accelerometer (Spearman rho, ρ=.85). Test-retest reliability was high (intraclass correlation coefficient, ICC2,1=0.85, 95% CI 0.81-0.88) and in good agreement with waist balance measurements (ICC2,1=0.84, 95% CI 0.80-0.88). Considering the normalized path length magnitude across all 3 axes improved interdevice correlation (ρ=.90) while maintaining test-retest reliability (ICC2,1=0.87, 95% CI 0.83-0.90). All subjects successfully completed the study, demonstrating the feasibility of using a head-mounted wearable to assess balance in a healthy population.ConclusionsBalance measurements derived from the smartglasses-based accelerometer were consistent with those obtained using a waist-mounted accelerometer. Additional research is necessary to determine to what extent smartglasses-based accelerometry measures can detect balance dysfunction associated with concussion. However, given the potential for smartglasses to perform additional concussion-related assessments in an integrated, wearable platform, continued development and validation of a smartglasses-based balance assessment is warranted. This approach could lead to a wearable platform for real-time assessment of concussion-related impairments that could be further augmented with telemedicine capabilities to integrate professional clinical guidance. Smartglasses may be superior to fully immersive virtual reality headsets for this application, given their lighter wei...

Longitudinal Socio-Emotional Learning Intervention for Autism via Smartglasses: Qualitative School Teacher Descriptions of Practicality, Usability, and Efficacy in General and Special Education Classroom Settings

“…Smartglasses, such as for example Google Glass, contain a complex array of sensors, and are able to collect video and audio data from the user and the environment. They are also able to monitor movement and velocity through an inbuilt gyroscope and accelerometer [35]. Smartglasses can provide users with video and audio guidance through an inbuilt display and bone conduction/audio speakers.…”

Reliability, Validity and Utility of Inertial Sensor Systems for Postural Control Assessment in Sport Science and Medicine Applications: A Systematic Review

“…Six studies investigated the convergence validity of inertial sensor-based PC assessments [12,13,16,[40][41][42]. Two studies were excluded from the validity synthesis as they investigated the convergence validity by comparing the measure to a second inertial sensor measurement [40,41], thus not comparing the method of measurement to another accepted standard.…”

“…Six studies investigated the convergence validity of inertial sensor-based PC assessments [12,13,16,[40][41][42]. Two studies were excluded from the validity synthesis as they investigated the convergence validity by comparing the measure to a second inertial sensor measurement [40,41], thus not comparing the method of measurement to another accepted standard. The four remaining studies used a clinical standard comparator (ie BESS/mBESS errors) and reported varying correlation coefficients ranging from r = 0.01 to -0.77, rho = 0.16 to 0.94 and ICC = 0.68 to 0.94 [12,13,16,42].…”

“…Twelve studies assessed discriminant validity in healthy cohorts, assessing the ability of inertial sensor systems to discriminate between known stance, visual, hearing, fatigue or taping conditions [25,28,32,41,[43][44][45][46][47][48][49][50].…”

“…A single prospective cohort study investigated the association of inertial sensor-based PC assessment scores with sports related injuries in a cohort of collegiate American football players, reporting that players with poorer PC were at a 5.19 times greater-odds of sustaining a sports related injury [61]. Table here 3.6 MEASUREMENT RELIABILITY Thirteen of the included studies were identified as assessing measurement reliability [16,27,28,33,40,41,45,47,58,[62][63][64][65]. Eleven of these studies investigated intra-session test retest reliability, demonstrating reliability ranging from poor to excellent (ICC = 0.03 to 0.97) [16,27,33,40,41,45,47,58,[63][64][65], while five studies reported poor to excellent (ICC = 0.02 to 0.95) inter-session reliability [16,28,[62][63][64].…”