Adaptive gait segmentation algorithm for walking bout detection using tri-axial accelerometers.

Gait analysis has many potential applications in understanding the activity profiles of individuals in their daily lives, particularly when studying the progression of recovery following injury, or motor deterioration in pathological conditions. One of the many challenges of conducting such analyses in the home environment is the correct and automatic identification of bouts of gait activity. To address this, a novel method for determining bouts of gait from accelerometer data recorded from the shank is presented. This method is fully automated and includes an adaptive thresholding approach which avoids the necessity for identifying subject-specific thresholds. The algorithm was tested on data recorded from 15 healthy subjects during self-selected slow, normal and fast walking speeds ranging from 0.48 ± 0.19 to 1.38 ± 0.33m/s and a single subject with PD walking at their normal walking speed (1.41 ± 0.08m/s) using accelerometers on the shanks. Intra-Class Correlation (ICC) confirmed high levels of agreement between bout onset/offset times and durations estimated using the algorithm, experimentally recorded stopwatch times and manual annotation for the healthy subjects (r=0.975, p <; 0.001; r=0.984, p<; 0.001) and moderate agreement for the PD subject (r=0.663, p<; 0.001). Mean absolute errors between accelerometer-derived and manually-annotated times were calculated, and ranged from 0.91 ± 0.05 s to 1.17 ± 2.26 s for bout onset detection, 0.80 ± 0.23 s to 2.41 ± 3.77 s for offset detection and 1.27 ± 0.13 s to 3.67 ± 4.59 s for bout durations.

Quantitative clinical assessment of motor function during and following LSVT-BIG® therapy.

BACKGROUND: LSVT-BIG® is an intensively delivered, amplitude-oriented exercise therapy reported to improve mobility in individuals with Parkinson's disease (PD). However, questions remain surrounding the efficacy of LSVT-BIG® when compared with similar exercise therapies. Instrumented clinical tests using body-worn sensors can provide a means to objectively monitor patient progression with therapy by quantifying features of motor function, yet research exploring the feasibility of this approach has been limited to date. The aim of this study was to use accelerometer-instrumented clinical tests to quantify features of gait, balance and fine motor control in individuals with PD, in order to examine motor function during and following LSVT-BIG® therapy. METHODS: Twelve individuals with PD undergoing LSVT-BIG® therapy, eight non-exercising PD controls and 14 healthy controls were recruited to participate in the study. Functional mobility was examined using features derived from accelerometry recorded during five instrumented clinical tests: 10 m walk, Timed-Up-and-Go, Sit-to-Stand, quiet stance, and finger tapping. PD subjects undergoing therapy were assessed before, each week during, and up to 13 weeks following LSVT-BIG®. RESULTS: Accelerometry data captured significant improvements in 10 m walk and Timed-Up-and-Go times with LSVT-BIG® (p <  0.001), accompanied by increased stride length. Temporal features of the gait cycle were significantly lower following therapy, though no change was observed with measures of asymmetry or stride variance. The total number of Sit-to-Stand transitions significantly increased with LSVT-BIG® (p <  0.001), corresponding to a significant reduction of time spent in each phase of the Sit-to-Stand cycle. No change in measures related to postural or fine motor control was observed with LSVT-BIG®. PD subjects undergoing LSVT-BIG® showed significant improvements in 10 m walk (p <  0.001) and Timed-Up-and-Go times (p = 0.004) over a four-week period when compared to non-exercising PD controls, who showed no week-to-week improvement in any task examined. CONCLUSIONS: This study demonstrates the potential for wearable sensors to objectively quantify changes in motor function in response to therapeutic exercise interventions in PD. The observed improvements in accelerometer-derived features provide support for instrumenting gait and sit-to-stand tasks, and demonstrate a rescaling of the speed-amplitude relationship during gait in PD following LSVT-BIG®.

Gait Event Detection From Accelerometry Using the Teager-Kaiser Energy Operator.

OBJECTIVE: A novel method based on the application of the Teager-Kaiser Energy Operator is presented to estimate instances of initial contact (IC) and final contact (FC) from accelerometry during gait. The performance of the proposed method was evaluated against four existing gait event detection (GED) methods under three walking conditions designed to capture the variance of gait in real-world environments. METHODS: A symmetric discrete approximation of the Teager-Kaiser energy operator was used to capture simultaneous amplitude and frequency modulations of the shank acceleration signal at IC and FC during flat treadmill walking, inclined treadmill walking, and flat indoor walking. Accuracy of estimated gait events were determined relative to gait events detected using force-sensitive resistors. The performance of the proposed algorithm was assessed against four established methods by comparing mean-absolute error, sensitivity, precision, and F1-score values. RESULTS: The proposed method demonstrated high accuracy for GED in all walking conditions, yielding higher F1-scores (IC: >0.98, FC: >0.9) and lower mean-absolute errors (IC: <0.018s, FC: <0.039s) than other methods examined. Estimated ICs from shank-based methods tended to exhibit unimodal distributions preceding the force-sensitive resistor estimated ICs, whereas estimated gait events for waist-based methods had quasiuniform random distributions and lower accuracy. CONCLUSION: Compared with the established gait event detection methods, the proposed method yielded comparably high accuracy for IC detection, and was more accurate than all other methods examined for FC detection. SIGNIFICANCE: The results support the use of the Teager-Kaiser Energy Operator for accurate automated GED across a range of walking conditions.

Application of the Teagar-Kaiser energy operator and wavelet transform for detection of finger tapping contact and release times using accelerometery.

The Teager-Kaiser energy operator (TKEO), when applied to a signal gives an estimation of the instantaneous energy of that signal. It, therefore, accentuates both frequency and amplitude changes in a signal. To date, it has been primarily used in communications systems and most popularly in electromyographic signal analysis to detect bursts of muscle activity, however, it has the potential to be used in a number of applications including accelerometer and movement data.A new algorithm was developed which used the TKEO to detect contact times during a finger tapping task from accelerometer data recorded from the index finger. The accuracy of the algorithm was assessed in 7 healthy control subjects during continuous finger tapping across a range of frequencies from 0.5Hz to 2.5Hz. The algorithm proved to be sensitive, correctly identifying at least 99% of all contacts in each of the finger tapping conditions that were tested. The mean absolute error of the contact detection is 14.7 ± 6 ms, while the mean absolute error of the release detection is 36.5 ± 36.3 ms. The proposed algorithm provides a method for the automatic detection of the temporal occurrences of the events of the finger tapping task using only a tri-axial accelerometer. The approach presented provides a means for objective assessment of finger tapping tasks for evaluation of the fine dexterity of the upper limb.