A Generalized Correlation Index for Quantifying Signal Morphological Similarity

TL;DR

This paper introduces a generalized correlation index that enhances the measurement of signal similarity, particularly for assessing injury severity in biomedical signals, by extending existing methods to include multiple levels of comparison.

Contribution

It generalizes the adaptive signed correlation index to incorporate specific signal features and multiple levels, improving injury assessment resolution.

Findings

Enhanced injury severity quantification

More refined signal similarity measurement

Demonstrated effectiveness in spinal cord injury assessment

Abstract

In biomedical applications, the similarity between a signal measured from an injured subject and a reference signal measured from a normal subject can be used to quantify the injury severity. This paper proposes a generalization of the adaptive signed correlation index (ASCI) to account for specific signal features of interest and extend the trichotomization of conventional ASCI to an arbitrary number of levels. In the context of spinal cord injury assessment, a computational example is presented to illustrate the enhanced resolution of the proposed measure and its ability to offer a more refined measure of the level of injury.