Estimating fall risk with inertial sensors using gait stability measures that do not require step detection

Riva, Federico; Toebes, Marcel J.P.; Pijnappels, Mirjam; Stagni, Rita; Dieën, Jaap H. van

doi:10.1016/j.gaitpost.2013.05.002

Cited by 135 publications

(136 citation statements)

References 42 publications

Supporting

Mentioning

122

Contrasting

Unclassified

Order By: Relevance

“…In particular, MSE, sLE and RQA were not affected by the presence of turns during the walk; having also recently proved to be related to fall history in treadmill walking tests [23,24], such measures could contribute to the definition of a fall risk index in free-walking conditions. Further research is needed to assess the capability of these measures to identify fall-prone subjects in an over-ground walking task.…”

Section: Resultsmentioning

confidence: 99%

Are gait variability and stability measures influenced by directional changes?

Riva

Grimpampi

Mazzà

et al. 2014

BioMedical Engineering OnLine

View full text Add to dashboard Cite

BackgroundMany gait variability and stability measures have been proposed in the literature, with the aim to quantify gait impairment, degree of neuro-motor control and balance disorders in healthy and pathological subjects. These measures are often obtained from lower trunk acceleration data, typically acquired during rectilinear gait, but relevant experimental protocols and data processing techniques lack in standardization. Since directional changes represent an essential aspect of gait, the assessment of their influence on such measures is essential for standardization. In addition, their investigation is needed to evaluate the applicability of these measures in laboratory trials and in daily life activity analysis. A further methodological aspect to be standardized concerns the assessment of the sampling frequency, which could affect stability measures. The aim of the present study was hence to assess if gait variability and stability measures are affected by directional changes, and to evaluate the influence of sampling frequency of trunk acceleration data on the results.MethodsFifty-one healthy young adults performed a 6-minute walk test along a 30 m straight pathway, turning by 180 deg at each end of the pathway. Nine variability and stability measures (Standard deviation, Coefficient of variation, Poincaré plots, maximum Floquet multipliers, short-term Lyapunov exponents, Recurrence quantification analysis, Multiscale entropy, Harmonic ratio and Index of harmonicity) were calculated on stride duration and trunk acceleration data (acquired at 100 Hz and 200 Hz) coming from straight walking windows and from windows including both straight walking and the directional change.ResultsHarmonic ratio was the only measure that resulted to be affected by directional changes and sampling frequency, decreasing with the presence of a directional change task. HR was affected in the AP and V directions for the 200 Hz, but only in AP direction for the 100 Hz group.ConclusionMultiscale entropy, short term Lyapunov exponents and Recurrence quantification analysis were generally not affected by directional changes nor by sampling frequency, and could contribute to the definition of a fall risk index in free-walking conditions.

show abstract

Section: Resultsmentioning

confidence: 99%

Are gait variability and stability measures influenced by directional changes?

Riva

Grimpampi

Mazzà

et al. 2014

BioMedical Engineering OnLine

View full text Add to dashboard Cite

show abstract

“…Factor analysis uses a statistical technique to examine variability between correlated features, and represents that variability as fewer factor variables; for example, Riva et al [5] used this method to represent 24 features as seven factors. Principal component analysis (PCA) is similar to factor analysis but uses orthogonal transformation to represent features as linearly uncorrelated variables called principal components.…”

Section: Introductionmentioning

confidence: 99%

“…This method reduced feature-space size from 67 features to as few as one feature [8]. Only these few studies, which used wearable sensor-derived features for faller classification, reduced feature-space size before faller classification [5–8]. …”

Section: Introductionmentioning

confidence: 99%

Feature selection for elderly faller classification based on wearable sensors

Howcroft

Kofman

Lemaire

2017

J NeuroEngineering Rehabil

View full text Add to dashboard Cite

BackgroundWearable sensors can be used to derive numerous gait pattern features for elderly fall risk and faller classification; however, an appropriate feature set is required to avoid high computational costs and the inclusion of irrelevant features. The objectives of this study were to identify and evaluate smaller feature sets for faller classification from large feature sets derived from wearable accelerometer and pressure-sensing insole gait data.MethodsA convenience sample of 100 older adults (75.5 ± 6.7 years; 76 non-fallers, 24 fallers based on 6 month retrospective fall occurrence) walked 7.62 m while wearing pressure-sensing insoles and tri-axial accelerometers at the head, pelvis, left and right shanks. Feature selection was performed using correlation-based feature selection (CFS), fast correlation based filter (FCBF), and Relief-F algorithms. Faller classification was performed using multi-layer perceptron neural network, naïve Bayesian, and support vector machine classifiers, with 75:25 single stratified holdout and repeated random sampling.ResultsThe best performing model was a support vector machine with 78% accuracy, 26% sensitivity, 95% specificity, 0.36 F1 score, and 0.31 MCC and one posterior pelvis accelerometer input feature (left acceleration standard deviation). The second best model achieved better sensitivity (44%) and used a support vector machine with 74% accuracy, 83% specificity, 0.44 F1 score, and 0.29 MCC. This model had ten input features: maximum, mean and standard deviation posterior acceleration; maximum, mean and standard deviation anterior acceleration; mean superior acceleration; and three impulse features. The best multi-sensor model sensitivity (56%) was achieved using posterior pelvis and both shank accelerometers and a naïve Bayesian classifier. The best single-sensor model sensitivity (41%) was achieved using the posterior pelvis accelerometer and a naïve Bayesian classifier.ConclusionsFeature selection provided models with smaller feature sets and improved faller classification compared to faller classification without feature selection. CFS and FCBF provided the best feature subset (one posterior pelvis accelerometer feature) for faller classification. However, better sensitivity was achieved by the second best model based on a Relief-F feature subset with three pressure-sensing insole features and seven head accelerometer features. Feature selection should be considered as an important step in faller classification using wearable sensors.

show abstract

“…These signals represent the overall motion pattern given the proximity of the sensor to the center of mass. Their processing enables the assessment of trunk stability, balance control and fall risk 18. In addition, the analysis of the recorded signals permits the identification of gait events 11 and the extraction of spatio-temporal gait parameters that are sensitive to patients with motor symptoms of PD 5…”

Section: Introductionmentioning

confidence: 99%

Is the Assessment of 5 Meters of Gait with a Single Body-Fixed-Sensor Enough to Recognize Idiopathic Parkinson’s Disease-Associated Gait?

et al. 2017

View full text Add to dashboard Cite

Quantitative assessment of gait in patients with Parkinson’s disease (PD) is an important step in addressing motor symptoms and improving clinical management. Based on the assessment of only 5 meters of gait with a single body-fixed-sensor placed on the lower back, this study presents a method for the identification of step-by-step kinematic parameters in 14 healthy controls and in 28 patients at early-to-moderate stages of idiopathic PD. Differences between groups in step-by-step kinematic parameters were evaluated to understand gait impairments in the PD group. Moreover, a discriminant model between groups was built from a subset of significant and independent parameters and based on a 10-fold cross-validated model. The discriminant model correctly classified a total of 89.5% participants with four kinematic parameters. The sensitivity of the model was 95.8% and the specificity 78.6%. The results indicate that the proposed method permitted to reasonably recognize idiopathic PD-associated gait from 5-m walking assessments. This motivates further investigation on the clinical utility of short episodes of gait assessment with body-fixed-sensors.Electronic supplementary materialThe online version of this article (doi:10.1007/s10439-017-1794-8) contains supplementary material, which is available to authorized users.

show abstract

Estimating fall risk with inertial sensors using gait stability measures that do not require step detection

Cited by 135 publications

References 42 publications

Are gait variability and stability measures influenced by directional changes?

Are gait variability and stability measures influenced by directional changes?

Feature selection for elderly faller classification based on wearable sensors

Is the Assessment of 5 Meters of Gait with a Single Body-Fixed-Sensor Enough to Recognize Idiopathic Parkinson’s Disease-Associated Gait?

Contact Info

Product

Resources

About