LAUSR

Abstract Purpose Non-invasive estimation of the conduction velocity distribution (CVD) of a peripheral nerve has the potential to both improve clinical diagnoses of pathology and to observe the progress of the disease or the efficacy of treatments. Comparisons were made of the performance of three non-invasive CVD estimation methods proposed by independent research groups on peripheral nerve bundles under different conditions. Methods The first method (Cummins) uses a nerve compound action potential (CAP) with temporal dispersion and a mathematical single fiber action potential (SFAP). The second method (Barker) uses two CAPs and a non-mathematical SFAP waveform. The third method (Hirose) uses two CAPs recorded from distal and proximal sites. The Cummins and Barker methods have iterative solutions in the time domain while the Hirose method is a deconvolution estimator in the frequency domain. In order to compare these methods, we used cold exposure to affect primarily motor fibers and ischemia to affect primarily sensory fibers on rat caudal nerve bundles. Results The Cummins method is sensitive to changes in motor and sensory fiber percentages in CVD if it is used with the volume conductor model. The Barker and Hirose methods are sensitive to motor fiber percentages in CVD but they cannot detect changes in sensory fiber percentages accurately. Conclusions Estimation of the CVD using a priori SFAP created with a volume conductor model can non-invasively supply accurate and precise information about fiber groups in a peripheral nerve bundle.