Automated Detection of Older Adults’ Naturally-Occurring Compensatory Balance Reactions: Translation From Laboratory to Free-Living Conditions

Abstract

Objective: Older adults' falls are a critical public health problem. The majority of free-living fall risk assessment methods have investigated fall predictive power of step-related digital biomarkers extracted from wearable inertial measurement unit (IMU) data. Alternatively, the examination of characteristics and frequency of naturally-occurring compensatory balance reactions (CBRs) may provide valuable information on older adults' propensity for falls. To address this, models to automatically detect naturally-occurring CBRs are needed. However, compared to steps, CBRs are rare events. Therefore, prolonged collection of criterion standard data (along with IMU data) is required to validate model's performance in free-living conditions. Methods: By investigating 11 fallers' and older non-fallers' freeliving criterion standard data, 8 naturally-occurring CBRs, i.e., 7 trips (self-reported using a wrist-mounted voice-recorder) and 1 hit/bump (verified using egocentric vision data) were localized in the corresponding trunk-mounted IMU data. Random forest models were trained on independent/unseen datasets curated from multiple sources, including in-lab data captured using a perturbation treadmill. Subsequently, the models' translation/generalization to older adults' out-of-lab data were assessed. Results: A subset of models differentiated between naturally-occurring CBRs and free-living activities with high sensitivity (100%) and specificity (≥ 99%). Conclusions: The findings suggest that accurate detection of naturally-occurring CBRs is feasible. Clinical/Translational Impact-As a multi-institutional validation study to detect older adults' naturally-occurring CBRs, suitability for larger-scale free-living studies to investigate falls etiology, and/or assess the effectiveness of perturbation training programs is discussed.INDEX TERMS Compensatory balance reactions, free-living digital biomarker, falls, machine learning, fall risk assessment.

Egocentric vision-based detection of surfaces: towards context-aware free-living digital biomarkers for gait and fall risk assessment

“…In our future work, other details of the walking surfaces will be considered. For instance in OA4’s multimodal data (IMU and FPV), 1 naturally-occurring (hit and bump) misstep was automatically detected by applying an IMU-based model, where a light pole was visually verified as the environmental fall risk [ 15 ]. Therefore, algorithms to detect such a static obstacle, as well as other tripping hazards (including dynamic obstacles such as pedestrians and pets, Fig.…”

“…3 ) and cracks in pavement [ 55 ] will be considered to provide complementary information on the properties of environment, towards a comprehensive context-aware free-living gait and fall risk assessment method. The automated identification of contexts associated with falls (and missteps [ 15 ]) using egocentric vision data would increase the interpretability of IMU-based FLDBs and address more specific intervention strategies, including the environmental modification (e.g., removing obstacles, securing fall areas, using non-slip flooring materials) as well as rehabilitation interventions (e.g., training to negotiate obstacles), which can potentially reduce the frequency of future falls in older adults. For instance, if a high frequency of slips (a form of misstep) is observed while walking on indoor tiles, non-slip flooring materials can be integrated to avoid future imbalance events.…”

“…In the case of observing unsatisfactory results from the former approach, the LOSO cross-validation approach (similar to leave-one-dataset-out in [ 21 ]) was considered as the next best in the present study to evaluate the framework’s generalizability and robustness against inter-participant variance in the MAGFRA-W dataset (although higher accuracies compared to the first approach is expected to be achieved [ 15 , 20 , 21 ]).…”

“…For instance, higher variability in acceleration signal (measured by the amplitude of the dominant frequency in the mediolateral direction, as a FLDB) during gait could indicate appropriate adaptation to the environment [ 13 ] (and potentially a lower risk of falls) and/or exhibit gait impairment (and potentially a higher risk of falls) [ 14 ]. Similarly, frequent missteps (as a FLDB) detected in free-living IMU data can be an indicator of impaired dynamic balance control (and a higher risk for falls [ 8 ]) and/or false alarms generated by anticipatory locomotion adjustment while walking on an irregular terrain (e.g., construction site) [ 15 ]. This ambiguity in interpretation leads to less precise intervention strategies to prevent falls.…”

“…The incorporated IMU-based approach exhibited satisfactory performance in our previous research [ 15 ] addressing a different research question, and was hypothesized to be a suitable baseline candidate for the purpose of terrain type identification. This approach demonstrated satisfactory performance for the identification of stair walking episodes, as discussed in Additional file 1 : Materials (I).…”