High-Specificity Digital Architecture for Real-Time Recognition of Loss of Balance Inducing Fall

Falls are a significant cause of loss of independence, disability and reduced quality of life 9 in people with Parkinson’s Disease (PD). Intervening quickly and accurately on the postural 10 instability could strongly reduce the consequences of falls. In this context, the paper proposes and 11 validates a novel architecture for the reliable recognition of losses of balance situations. The 12 proposed system addresses some challenges related to the daily life applicability of near fall 13 recognition systems: the high specificity and the system robustness against the Activities of Daily 14 Life (ADL). In this respect, the proposed algorithm has been tested on five different tasks: walking 15 steps, sudden curves, chair transfers via Timed Up&Go (TUG) test, balance-challenging obstacle 16 avoidance and slip-induced loss of balance. The system analyzes data from wireless acquisition 17 devices that capture electroencephalography (EEG) and electromyography (EMG) signals. The 18 collected data are sent to two main units: the Muscular Unit and the Cortical one. The first one 19 realizes a binary ON/OFF pattern from muscular activity (10 EMGs) and triggers the Cortical Unit. 20 This latter unit evaluates the rate of variation in the EEG power spectrum density (PSD), considering 21 five bands of interest. The neuromuscular features are then sent to a logical network for the final 22 classification, which distinguishes among falls and ADL. 23 In this preliminary study, we tested the proposed model on 9 healthy subjects (aged 26.3 ± 2.4), even 24 if the study on PD patients is under investigation. Experimental validation on healthy subjects 25 showed that the system reacts in 370.62 ± 60.85 ms with a sensitivity of the 93.33 ± 5.16 %. During 26 the ADL tests the system showed a specificity of 98.91 ± 0.44 % in steady walking steps recognition, 27 99.61 ± 0.66 % in sudden curves detection, 98.95 ± 1.27 % in contractions related to TUG tests and 28 98.42 ± 0.90 % in the obstacle avoidance protocol.