Slip-induced falls are a growing health concern for older adults, and near-fall events are associated with an increased risk of falling. To detect older adults at a high risk of slip-related falls, this study aimed to develop models for near-fall event detection based on accelerometry data collected by body-fixed sensors. Thirty-four healthy older adults who experienced 24 laboratory-induced slips were included. The slip outcomes were first identified as loss of balance (LOB) and no LOB (NLOB), and then the kinematic measures were compared between these two outcomes. Next, all the slip trials were split into a training set (90%) and a test set (10%) at sample level. The training set was used to train both machine learning models (n = 2) and deep learning models (n = 2), and the test set was used to evaluate the performance of each model. Our results indicated that the deep learning models showed higher accuracy for both LOB (>64%) and NLOB (>90%) classifications than the machine learning models. Among all the models, the Inception model showed the highest classification accuracy (87.5%) and the largest area under the receiver operating characteristic curve (AUC), indicating that the model is an effective method for near-fall (LOB) detection. Our approach can be helpful in identifying individuals at the risk of slip-related falls before they experience an actual fall.

Stroke survivors are at a relatively higher risk of falling than their healthy counterparts. To identify the key gait characteristics affecting fall risk in this population, this study analyzed the gait kinematics and gait asymmetries for 36 community-dwelling people with chronic stroke (PwCS). According to their fall history in the last 12 months, they were divided into a fall group (n = 21) and non-fall group (n = 15), and then the gait kinematics (step length, stride length, stance time, swing time, trunk angle, and segment angles for lower limbs) and their asymmetries (symmetry ratio and symmetry index) were compared between these two groups. To investigate the relationship between fall types and gait characteristics, these variables were also compared between 11 slip-fallers and non-fallers, as well as between 7 trip-fallers and non-fallers. Our results indicated that the fallers showed smaller trunk and thigh angle, larger shank angle, and higher gait asymmetries (trunk and foot). Such changes in gait pattern could also be found in the trip-fallers, except the trunk angle. Additionally, the trip-fallers also showed a shorter step length, shorter stride length, shorter swing time, larger foot angle on the paretic side, and higher asymmetries in shank angle and step length, while the slip-fallers only showed changes in trunk angle and thigh angle and higher asymmetries in step length and foot angle compared to the non-fall group. Our results indicated that improper or pathological gait patterns (i.e., smaller thigh angle or higher foot asymmetry) increases the risk of falling in PwCS, and different fall types are associated with different gait characteristics. Our findings would be helpful for the development of fall risk assessment methods that are based on kinematic gait measurements. Implementation of objective fall risk assessments in PwCS has the potential to reduce fall-related injuries, leading to a reduction in associated hospital costs.

Trip perturbations are proposed to be a leading cause of falls in older adults. To prevent trip-falls, trip-related fall risk should be assessed and subsequent task-specific interventions improving recovery skills from forward balance loss should be provided to the individuals at risk of trip-fall. Therefore, this study aimed to develop trip-related fall risk prediction models from one’s regular gait pattern using machine-learning approaches. A total of 298 older adults (≥60 years) who experienced a novel obstacle-induced trip perturbation in the laboratory were included in this study. Their trip outcomes were classified into three classes: no-falls (n = 192), falls with lowering strategy (L-fall, n = 84), and falls with elevating strategy (E-fall, n = 22). A total of 40 gait characteristics, which could potentially affect trip outcomes, were calculated in the regular walking trial before the trip trial. The top 50% of features (n = 20) were selected to train the prediction models using a relief-based feature selection algorithm, and an ensemble classification model was selected and trained with different numbers of features (1–20). A ten-times five-fold stratified method was utilized for cross-validation. Our results suggested that the trained models with different feature numbers showed an overall accuracy between 67% and 89% at the default cutoff and between 70% and 94% at the optimal cutoff. The prediction accuracy roughly increased along with the number of features. Among all the models, the one with 17 features could be considered the best model with the highest AUC of 0.96, and the model with 8 features could be considered the optimal model, which had a comparable AUC of 0.93 and fewer features. This study revealed that gait characteristics in regular walking could accurately predict the trip-related fall risk for healthy older adults, and the developed models could be a helpful assessment tool to identify the individuals at risk of trip-falls.

This study assessed the effect of filter parameters on gait characteristics and the performance of machine-learning models. Overground walking trials (n = 99) with and without perturbations (slips, trips) were collected for 33 healthy older adults. Kinematics were collected by a motion capture system. Different Butterworth low-pass parameters were applied to the raw data, including three orders (2–6) and nine cutoffs (4–20 Hz). Spatiotemporal gait outcomes were then calculated to develop classification models to automatically identify the trial type (gait, gait–slip, or gait–trip) using Logistic Regression, Support Vector Classification, and Random Forest Classification. A 3 × 9 ANOVA showed main effects of order and cutoff (p < 0.01 for all) on gait characteristics during both perturbed and regular walking trials. However, the gait characteristics were different between them. The filter parameters significantly affected the performance of classification models using different classifiers, with significant main effects of the filter order (p < 0.05) and cutoff (p < 0.01) on AUC and overall accuracy for all of the models. Our results suggest that the standard Butterworth filter (fourth-order, cutoff: 6 Hz) is suitable for the development of classification models with low–medium complexity, while for models with high complexity (i.e., ensemble models), a filter with a higher order and cutoff (sixth-order, cutoff 10–12 Hz) might yield better performance.

Inertial sensors (IS) enable the kinematic analysis of human motion with fewer logistical limitations than the silver standard optoelectronic motion capture (MOCAP) system. However, there are no data on the validity of IS for perturbation training and during the performance of dance. The aim of this present study was to determine the concurrent validity of IS in the analysis of kinematic data during slip and trip-like perturbations and during the performance of dance. Seven IS and the MOCAP system were simultaneously used to capture the reactive response and dance movements of fifteen healthy young participants (Age: 18–35 years). Bland Altman (BA) plots, root mean square errors (RMSE), Pearson’s correlation coefficients (R), and intraclass correlation coefficients (ICC) were used to compare kinematic variables of interest between the two systems for absolute equivalency and accuracy. Limits of agreements (LOA) of the BA plots ranged from −0.23 to 0.56 and −0.21 to 0.43 for slip and trip stability variables, respectively. The RMSE for slip and trip stabilities were from 0.11 to 0.20 and 0.11 to 0.16, respectively. For the joint mobility in dance, LOA varied from −6.98–18.54, while RMSE ranged from 1.90 to 13.06. Comparison of IS and optoelectronic MOCAP system for reactive balance and body segmental kinematics revealed that R varied from 0.59 to 0.81 and from 0.47 to 0.85 while ICC was from 0.50 to 0.72 and 0.45 to 0.84 respectively for slip–trip perturbations and dance. Results of moderate to high concurrent validity of IS and MOCAP systems. These results were consistent with results from similar studies. This suggests that IS are valid tools to quantitatively analyze reactive balance and mobility kinematics during slip–trip perturbation and the performance of dance at any location outside, including the laboratory, clinical and home settings.

During balance recovery from slip perturbations, forward flexion (elevation) of the arms serves to counterbalance the posteriorly displaced center of mass (CoM). We aimed to investigate whether aging affects modulation of arm responses to various intensities of unpredictable slip perturbations and whether arm responses are related to compensatory stepping stability. Ten healthy young adults and ten healthy older adults participated. Participants were asked to react naturally to three randomly administered levels of slip-like surface perturbations (intensity 1 (7.75 m/s²), intensity 2 (12.00 m/s²) and intensity 3 (16.75 m/s²), which occurred by means of forward acceleration of the treadmill belt while standing. Kinematic data were collected using a motion capture system. Outcomes included arm elevation displacement, velocity, and margin of stability (MoS) of compensatory stepping. The results reveal no modulation of arm elevation velocity in older adults from perturbation intensity 1 to 2, whereas younger adults demonstrated progressive increases from intensity 1 to 2 to 3. At intensity 3, older adults demonstrated reduced maximal arm elevation velocity compared to younger adults (p = 0.02). The results in both groups combined reveal a positive correlation between maximal arm elevation velocity and first compensatory step MoS at intensity 3 (p = 0.01). Together, these findings indicate age-related decreases in arm response modulation and the association of arm elevation response with protective stepping stability, suggesting that fall prevention interventions may benefit from an emphasis on arm elevation velocity control in response to greater perturbation intensities.