Gait parameter and event estimation using smartphones.

BACKGROUND AND OBJECTIVES: The use of smartphones can greatly help for gait parameters estimation during daily living, but its accuracy needs a deeper evaluation against a gold standard. The objective of the paper is a step-by-step assessment of smartphone performance in heel strike, step count, step period, and step length estimation. The influence of smartphone placement and orientation on estimation performance is evaluated as well. METHODS: This work relies on a smartphone app developed to acquire, process, and store inertial sensor data and rotation matrices about device position. Smartphone alignment was evaluated by expressing the acceleration vector in three reference frames. Two smartphone placements were tested. Three methods for heel strike detection were considered. On the basis of estimated heel strikes, step count is performed, step period is obtained, and the inverted pendulum model is applied for step length estimation. Pearson correlation coefficient, absolute and relative errors, ANOVA, and Bland-Altman limits of agreement were used to compare smartphone estimation with stereophotogrammetry on eleven healthy subjects. RESULTS: High correlations were found between smartphone and stereophotogrammetric measures: up to 0.93 for step count, to 0.99 for heel strike, 0.96 for step period, and 0.92 for step length. Error ranges are comparable to those in the literature. Smartphone placement did not affect the performance. The major influence of acceleration reference frames and heel strike detection method was found in step count. CONCLUSION: This study provides detailed information about expected accuracy when smartphone is used as a gait monitoring tool. The obtained results encourage real life applications.

A smartphone-based architecture to detect and quantify freezing of gait in Parkinson's disease.

INTRODUCTION: The freezing of gait (FOG) is a common and highly distressing motor symptom in patients with Parkinson's Disease (PD). Effective management of FOG is difficult given its episodic nature, heterogeneous manifestation and limited responsiveness to drug treatment. METHODS: In order to verify the acceptance of a smartphone-based architecture and its reliability at detecting FOG in real-time, we studied 20 patients suffering from PD-related FOG. They were asked to perform video-recorded Timed Up and Go (TUG) test with and without dual-tasks while wearing the smartphone. Video and accelerometer recordings were synchronized in order to assess the reliability of the FOG detection system as compared to the judgement of the clinicians assessing the videos. The architecture uses two different algorithms, one applying the Freezing and Energy Index (Moore-Bächlin Algorithm), and the other adding information about step cadence, to algorithm 1. RESULTS: A total 98 FOG events were recognized by clinicians based on video recordings, while only 7 FOG events were missed by the application. Sensitivity and specificity were 70.1% and 84.1%, respectively, for the Moore-Bächlin Algorithm, rising to 87.57% and 94.97%, respectively, for algorithm 2 (McNemar value=28.42; p=0.0073). CONCLUSION: Results confirm previous data on the reliability of Moore-Bächlin Algorithm, while indicating that the evolution of this architecture can identify FOG episodes with higher sensitivity and specificity. An acceptable, reliable and easy-to-implement FOG detection system can support a better quantification of the phenomenon and hence provide data useful to ascertain the efficacy of therapeutic approaches.

Experimental evaluation of a smartphone based Step Length estimation.

Step Length (SL) is an essential parameter in the healthcare field to monitor the gait of patients affected by motor disorders such as Freezing of Gait (FoG), a motor block that provokes an interruption of the normal gait cycle. As a consequence spatio-temporal parameters of gait, in particular SL, are strongly altered before and during a FoG event. In this work we present a non-intrusive and non-invasive architecture applicable in this clinical scenario and we evaluate its reliability of SL estimation on 8 healthy subjects. We obtained mean errors of 7.77%, 6.99% and 6.44% for low, normal and high velocity respectively, which is a sufficient accuracy for FoG detection.