Explaining the unique nature of individual gait patterns with deep learning

Fabian Horst,Wojciech Samek,Klaus-Robert Müller,Sebastian Lapuschkin,Wolfgang I Schöllhorn

doi:10.1038/s41598-019-38748-8

Abstract

Machine learning (ML) techniques such as (deep) artificial neural networks (DNN) are solving very successfully a plethora of tasks and provide new predictive models for complex physical, chemical, biological and social systems. However, in most cases this comes with the disadvantage of acting as a black box, rarely providing information about what made them arrive at a particular prediction. This black box aspect of ML techniques can be problematic especially in medical diagnoses, so far hampering a clinical acceptance. The present paper studies the uniqueness of individual gait patterns in clinical biomechanics using DNNs. By attributing portions of the model predictions back to the input variables (ground reaction forces and full-body joint angles), the Layer-Wise Relevance Propagation (LRP) technique reliably demonstrates which variables at what time windows of the gait cycle are most relevant for the characterisation of gait patterns from a certain individual. By measuring the time-resolved contribution of each input variable to the prediction of ML techniques such as DNNs, our method describes the first general framework that enables to understand and interpret non-linear ML methods in (biomechanical) gait analysis and thereby supplies a powerful tool for analysis, diagnosis and treatment of human gait.

Highlights

The ability to walk is crucial for human mobility and enables to predict quality of life, morbidity and mortality[1,2,3,4,5,6,7,8,9,10]
Layer-wise Relevance Propagation (LRP) decomposes the prediction f(x) of a learned function f given an input sample x into into time-resolved input relevance values Ri for each time-discrete input xi, which enables to explain the prediction of deep artificial neural networks (DNN) as partial contributions from individual input components (Fig. 1 III: Explain prediction using LRP)
By decomposing the prediction of machine learning methods such as artificial neural networks back to the input variables, the LRP technique demonstrated which gait variables were most relevant for the characterisation of gait patterns from a certain individual

Summary

Introduction

The ability to walk is crucial for human mobility and enables to predict quality of life, morbidity and mortality[1,2,3,4,5,6,7,8,9,10]. Machine learning techniques are solving very successfully a variety of classification tasks and provide new insights from complex physical, chemical, biological, or social systems, in most cases they go along with the disadvantage of acting as a black box, rarely providing information about what made them arrive at a particular decision[47,48] This non-transparent operating and decision-making of most non-linear machine learning methods leads to the problem that their predictions are not straightforward understandable and interpretable. Interpreting linear and non-linear models have helped to gain interesting insights in neuroscience[62,63], bioinformatics[64,65,66] and physics[67]

Objectives

Methods

Results

Conclusion