Decomposition of Sensory Nerve Action Potential and Motor Unit Action Potential in Patients with Diabetic Neuropathy using Wavelet Transform and Fast Fourier Transform

J Korean Assoc EMG Electrodiagn Med 1999;1(2):156-164.
Published online May 9, 1999.

Abstract

Objectives : To evaluate the characteristics of nerve conduction abnormalities and activities of different types of muscle fibers in diabetic neuropathy with comparison to clinical and electrophysiological features, using analysis of power spectral density and relative energy ratio. Methods : For analysis of sensory nerve action potentials (SNAPs) we used the distribution of conduction velocity (DCV) based upon power spectral analysis, power spectral density (PSD) by fast fourier transformation, and relative energy ratio of the low frequency versus high frequency by wavelet transform method. For investigation of electromyography (EMG), PSD was determined during minimal and maximal volitions. Results : DCV demonstrated reduction of slow fibers (<40 m/sec) in small or large fiber dysfunction and attenuation of fast fibers (40~60 m/sec) in severe mixed fiber lesions. PSD of SNAPs showed reduction of low frequencies with increase of high and intermediate frequencies. PSD of EMG signals on maximal volition was significantly decreased in diabetics. PSD of SNAPs and EMG revealed a correlation with electrodiagnostic severity. Conclusion : Attenuation of slower nerve fibers and reduction of lower frequencies of SNAPs could be early findings of diabetic neuropathy. In addition, dysfunction of type II myofibers would be highly suggestive. Power spectral analysis could be utilized for early diagnosis.

Key Words: Diabetic neuropathy, Action potentials, Electromyography, Spectrum analysis, Transformation