Spectral Analysis of Gait Variability of Stride Interval Time Series: Comparison of Young, Elderly and Parkinson's Disease Patients

[Purpose] The purpose of this study was to verify a method of spectral analysis of stride-to-stride variability, and to apply this method to the analysis of the gait pattern of patients with Parkinson's disease. [Subjects] Ten healthy young individuals, ten community-dwelling elderly individuals and nine individuals with Parkinson's disease were recruited. [Methods] To quantitatively evaluate stride-to-stride variability, we used fast Fourier transform, calculated the power spectrum, and separated it into three frequency ranges. We plotted a double-log graph for the power spectrum and calculated the slope of the line using least-squares regression. For each of the participants we measured 10-meter walking speed, stride-to-stride variability, and disease severity (only Parkinson's disease). [Results] Each of three ranges of the power spectrum was four times larger for Parkinson's disease patients than for the healthy groups (p < 0.05). The severity of Parkinson's disease correlated significantly with the very low frequency range (0.01-0.05 Hz) of the power spectrum (r = 0.767, p = 0.0159) and the scaling factor β of the power spectrum (r = 0.850, p = 0.0037). [Conclusion] This method of spectral analysis of stride-to-stride variability may be useful for gait analysis of patients with Parkinson's.