Gait Rhythm Fluctuation Analysis for Neurodegenerative Diseases by Empirical Mode Decomposition.

Previous studies have indicated that gait rhythm fluctuations are useful for characterizing certain pathologies of neurodegenerative diseases such as Huntington's disease (HD), amyotrophic lateral sclerosis (ALS), and Parkinson's disease (PD). However, no previous study has investigated the properties of frequency range distributions of gait rhythms. Therefore, in our study, empirical mode decomposition was implemented for decomposing the time series of gait rhythms into intrinsic mode functions from the high-frequency component to the low-frequency component sequentially. Then, Kendall's coefficient of concordance and the ratio for energy change for different IMFs were calculated, which were denoted as W and R , respectively. Results revealed that the frequency distributions of gait rhythms in patients with neurodegenerative diseases are less homogeneous than healthy subjects, and the gait rhythms of the patients contain much more high-frequency components. In addition, parameters of W and R can significantly differentiate among the four groups of subjects (HD, ALS, PD, and healthy subjects) (with the minimum p-value of 0.0000493). Finally, five representative classifiers were utilized in order to evaluate the possible capabilities of W and R to distinguish the patients with neurodegenerative diseases from the healthy subjects. This achieved maximum area under the curve values of 0.949, 0.900, and 0.934 for PD, HD, and ALS detection, respectively. In sum, our study suggests that gait rhythm features extracted in the frequency domain should be given consideration seriously in the future neurodegenerative disease characterization and intervention.