Objective: In this study, we aim to build an automated-FoG-detector for subject-independent. After extracting enormously relevant features, we apply variance detection-techniques to identify-FoG-events.

Background: Gait is one of the most affected motor-characteristics in Parkinson’s disease (PD). Freezing of gait (FoG), defined as a motor block of movement (particularly prior to gait-initiation) during turns or when meeting obstacles [1], is one of the most common symptoms ([2] reported that 47% of > 6000 subjects had 28% FoG events every day). Furthermore, there is a sturdy relationship between FoG and falls [1], [3], [4]. PD in advanced stage presents FoG symptom which is common and robustly relates to falls [5]-[7]. In search of new methodologies and equipment to aid improve patients lifestyle, a new non-invasive wireless-system is proposed in this study to identify FoG. Freezing of gait (FoG) is frequently encountered in Parkinsonian-gait and firmly-relates to falls. Current clinical FoG assessments are patients’ self-report diaries and manually experts’ video-analysis. Both are subjective and give rise to reliably-moderate, i.e., moderate reliability. Current detection algorithms have been principally designed in subject-dependent-settings.

Methods: We first build our algorithm with a dataset from the Daphnet project [1]. Specifically, feature selection is performed using statistical-cross-correlation and Pearson`s correlation and clustering-techniques (clusterability-metrics). From a list of 244 feature candidates, 36 candidates were selected using saliency and robustness-criterion. We develop an anomaly score-detector with adaptive thresholding to identify FoG events. Then, using the clusterability-accuracy-metrics, we minimize the feature list to 7 candidates.

Results: Our novel multichannel-freezing-gait index is most selective across all window-sizes, attaining sensitivity(specificity) of 95.6%(circa ~79%). On the other hand, freezing-index from the perpendicular i.e., vertical-axis was the best-choice for a single-input (I/P), achieving sensitivity (specificity) of 93.6% (circa ~84%) for ankle and 88.6% (circa ~94%) for back-sensors.

Conclusions: The freezing-index-feature from a single-channel (X or Y-axis) at ankle or hip-sensor-location can be used for an incongruity-detection based scheme to detect-FoG-events. Our proposed method is objective and significantly outperforms (e.g., mean (±SD) of sensitivity, specificity are 93.6% (±23%) and 83.6% (±36%) for ASD ankle y-axis) other automated methods in the literature.

References: 1.Bloem et al., “Falls and freezing of gait in Parkinson’s disease: A review of two interconnected, episodic phenomena,” Mov Disord, vol. 19, no. 8, pp. 871–884, Aug. 2004. 2. M. Macht et al., “Predictors of freezing in Parkinson’s disease: A survey of 6,620 patients,” Mov Disord, vol. 22, no. 7, pp. 953–956, May 2007. 3. M. Latt et al., “Clinical and physiological assessments for elucidating falls risk in Parkinson’s disease,” Mov Disord, vol. 24, no. 9, 1280–1289, Jul. 2009. 4. S. Paul et al., “Three simple clinical tests to accurately predict falls in people with Parkinson’s disease,” Mov Disord, vol. 28, no. 5, Pp: 655-662, May 2013. 5. M. Bachlin et al., “Wearable assistant for Parkinson’s disease patients with the freezing of gait symptom,” IEEE Trans. Inf. Technol. Biomed., vol. 14, no. 2, pp. 436–446, Mar. 2010. 6. Giladi et al., “Validation of the freezing of gait questionnaire in patients with Parkinson’s disease,” Mov Disord, vol. 24, no. 5, Pp: 655–661, 2009. 7. M. Bachlin et al., “Wearable assistant for Parkinson’s disease patients with the freezing of gait symptom,” IEEE Trans. Inf. Technol. Biomed., vol. 14, no. 2, pp. 436–446, Mar. 2010. 8. M. M. Hoehn et al., “Parkinsonism: Onset, progression, and mortality,” Neurology, vol. 50, no. 2, pp. 318–318, 1998. 9. W. Gibb and A. Lees, “A comparison of clinical and pathological features of young-and old-onset Parkinson’s disease,” Neurology, vol. 38, no. 9, pp. 1402–1402, 1988.

« Back to 2018 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/freezing-of-gait-recognition-in-parkinsons-disease-a-subject-independent-objective-method/