(I-334) USING A SMARTPHONE TO MEASURE STATIC BALANCE: VERIFICATION AND VALIDATION

Introduction:: Postural stability is an involuntary mechanism for balancing a person’s center of gravity and maintaining an upright posture. Balance testing can be used as an indicator of an older adult’s fall risk [1], a tool to diagnose disease [2-4], and a measure to determine the effectiveness of interventions [5, 6]. Force plates are the current gold standard to measure static balance, where common balance measurements used are the root mean square (RMS) of the center of pressure (COP) velocity in the anteroposterior (AP) and mediolateral (ML) directions, the primary axes of postural sway, and the COP 95% confidence ellipse area. However, such quantitative balance testing is rarely performed in a clinical setting due to the cost of physical equipment and expertise required to perform the testing; instead, qualitative methods are used, allowing for physician bias. Recent research investigates an alternative method to measure balance based on body accelerations using an inertial measurement unit (IMU) or a smartphone. The body’s acceleration characterizes postural sway, while the COP velocity indicates the efficiency of postural control, and thus a relationship should be observed in the data [7]. The widespread use of smartphones enables the potential to offer a low-cost, accessible way to measure static balance [8]. The objectives of this study were to determine if a smartphone’s acceleration values were comparable to those of an IMU and correlated to force plate balance metrics, and if the smartphone could detect differences in balance metrics across three balance poses of increasing difficulty.

Materials and Methods:: Protocols were approved by the Cal Poly Institutional Review Board. Thirty-eight participants (21 Female, 17 Male; aged 12-25) were included in the study. Participants stood on a force plate (AccuGait, AMTI, USA) shoulder width apart with their arms crossed over their chest, wearing a chest strap (Free Train, UK) which held a smartphone (Apple iPhone 11, USA) and an IMU (MMS, MetaMotions, USA). Three balance poses, including both legs (BL), semi-tandem (ST), and single leg (SL) (Figure 1), were each held for 30 seconds. COP positional data, converted into velocity through the time derivative, were collected from the force plate (150 Hz), while acceleration data were collected from the smartphone and IMU (100 Hz). Smartphone and IMU balance parameters were defined as the RMS of the acceleration values in the AP and ML directions. Force plate balance parameters were defined as the RMS of the COP velocity in the AP and ML directions and the COP 95% confidence ellipse area. Spearman correlations were run between the IMU and smartphone acceleration data, as well as between the smartphone and force plate balance data, to verify that the smartphone can similarly detect balance as a force plate would. Repeated measure analyses of variance (RMANOVA) with post-hoc Tukey tests were run for the smartphone parameters between the three poses to determine the smartphone’s ability to decipher between pose difficulties. For all statistical tests, p < 0.05 was defined as significant.

Results, Conclusions, and Discussions:: All IMU and smartphone balance parameters were significantly correlated (p < 0.001), verifying the smartphone acceleration balance parameters against the IMU data. The Spearman correlations between smartphone and force plate balance parameters showed significant correlations between all force plate balance parameters and the smartphone RMS values in the ML direction (p < 0.030), but not the AP direction (p >0.300). This suggests that the smartphone can reasonably detect static human balance in the ML direction, indicating that postural sway occurs mostly in the ML direction and is less prominent in the AP direction. Statistical significance (p=0.001) was observed between the smartphone RMS value in the ML direction and the COP 95% confidence ellipse, which is the area of the COP trajectory in both the ML and AP directions. This aligns with previous research considering the 95% confidence ellipse to be the most widely accepted measure of balance [7, 9, 10]. Statistically significant differences were detected between the BL and SL poses and between the ST and SL poses for ML smartphone acceleration RMS values (p < 0.002, Figure 2). This suggests that the smartphone can reasonably differentiate between pose difficulty, thereby demonstrating the smartphone’s ability to assess static human balance in the ML direction [11]. While further research is needed to fully validate the smartphone’s ability to effectively measure human balance, data from this study show promising results.

Acknowledgements (Optional): : Funding was provided by the Donald E. Bently Center for Engineering Innovation (SK)


References (Optional): : [1] Hsieh, K. et al., Gait Posture. 67:160-165, 2019. [2] Park, J. et al., Int J Ind Ergon. 48:139-148, 2015. [3] Horak, F. et al., Movement Disorders. 28(11):1544-1551, 2013. [4] Dibble, L. et al., Physical Therapy. 88(3): 323-332, 2008. [5] Otero, M. et al., Clin Interv Aging. 12:505, 2017. [6] Darekar, A. et al., J Neuroeng Rehabil. 12(1):1-14, 2015. [7] Chen, B. et al., Int J Environ Res Public Health. 18(5), 2696, 2021. [8] Poushter, J., Pew Research Center. 22(1), 1-44, 2016. [9] Swanenburg, J. et al., BMC Musculoskelet Disord. 9, 1-10, 2008. [10] Park, T., J Phys Ther Sci . 26(4), 625-627, 2014. [11] Heebner, N. et al., Gait Posture. 41(2):535-539, 2015.