A Machine Learning Approach to Unveil Balance Behavior Through Aging with an Auditory Cue

Moving through space involves different sensory cues coming from the environment. Previous works show that auditory cues are used in association with other sensory inputs to maintain balance. Moreover, testing spatial hearing under adverse conditions (e.g., eyes closed) provides additional insights into how the absence of one sensory modality could affect self-motion perception. This work aims to design an innovative approach combining kinematic methods and machine learning algorithms to increase the accuracy of a widely used clinical test (Fukuda-Unterberger stepping test). Thanks to the present protocol, we have quantified motor parameters (e.g., linear displacement, angular displacement, and angle of rotation), detecting variations between different audio conditions and age groups. The first hypothesis is to find an advantage in postural control with the use of an auditory cue, providing a better understanding of the interconnection between hearing and the vestibular organs. The second hypothesis is to discriminate motor test performance between different age groups, in line with the progressive vestibular deterioration due to age.