Dissertation Defence: Classification of Fall Risk in Older Adults Using Foot-Tapping

Shamim Noroozi, supervised by Dr. Jennifer Jakobi, will defend their dissertation titled “Classification of Fall Risk in Older Adults Using Foot-Tapping: The Application of Wearable Sensors” in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Mechanical Engineering.

An abstract for Shamim Noroozi’s dissertation is included below.

Examinations are open to all members of the campus community as well as the general public. Please email jennifer.jakobi@ubc.ca to receive the Zoom link for this exam.

Abstract

This dissertation investigated the application of an accelerometer-instrumented unilateral foot-tapping test as a brief, space-efficient, and practical tool for assessing mobility and predicting clinically meaningful outcomes in aging. A total of 67 older adults (65-95 years) and 20 young-to-midlife adults (19-64 years) participated. Participants completed 30-second unilateral foot-tapping with the dominant and nondominant foot, a 4-m gait speed test, the Timed Up and Go test, the frailty phenotype, and fall-risk assessments based on the World Guidelines for Falls Prevention and Management. The test order was randomized. Foot-tapping was recorded using tri-axial accelerometers, and features were extracted from the time, frequency, and time-frequency domains. The specific objectives were: 1) to validate automated tap detection from accelerometer signals during unilateral foot-tapping; 2) to examine associations between foot-tapping features with gait speed and Timed Up and Go times to predict mobility outcomes across the adult lifespan; 3) to classify frailty in older adults; and 4) to classify fall-risk in older adults. Across the four studies, the findings showed that the proposed wavelet-based tap-detection algorithms identified individual taps with high sensitivity and precision, especially when the discrete wavelet transform was combined with an energy detector. Foot-tapping features were significantly associated with gait speed and Timed Up and Go performance, with prediction driven mainly by temporal and time-frequency features; the dominant foot provided the strongest prediction of Timed Up and Go time. Unilateral foot-tapping also distinguished non-frail older adults from a merged pre-frail and frail group, with the dominant foot better identifying pre-frail and frail individuals and the nondominant foot better identifying non-frail individuals. In addition, unilateral foot-tapping differentiated low fall-risk from elevated-risk in older adults, with the nondominant foot showing stronger overall classification performance than the dominant foot. Across the dissertation, the most informative predictors were concentrated mainly in the time and time-frequency domains, while frequency-domain features contributed less. Overall, this dissertation supports unilateral accelerometer-based foot-tapping as an objective and portable approach for mobility, frailty, and fall-risk assessment in older adults.