Introduction:​ Diabetic peripheral neuropathy is one of the common complications of type-2 diabetes mellitus (DM). Changes in the intrinsic plantar tissue coupled with repetitive mechanical loads and loss of sensation may lead to foot related complications (skin break down, ulcerations, and amputations) in persons with neuropathy if left untreated. The purpose of this dissertation was to stratify individuals with pre-diabetes, diabetes with and without neuropathy using dynamic plantar pressure parameters during walking, using machine learning algorithms.

Methods:​ Plantar pressure data was collected from one hundred participants during walking with pressure measuring insoles fixed between the feet and thin socks. Simultaneously high-definition videos were collected using a camera placed behind the participants. Walking speed was computed in the narrow walkways established in laboratory and clinical settings using a single calibrated camera. Support vector machine algorithms were implemented using dynamic plantar pressure parameters and participant-specific parameters to predict group classification.

Results:​ The camera calibration approach estimated the walking speed at three different physical locations with good to excellent intra-rater reliability. No significant differences in plantar pressure measures were found among the three participant groups using traditional statistical analysis. Support vector machine classifiers were able to successfully classify the participant groups with very high sensitivity and specificity.

Conclusions:​ The study’s findings may allow early detection of neuropathy in persons with DM using quantitative measures of the dynamic plantar pressure distribution, which may prevent foot ulcerations. Future implementation of machine learning algorithms will offer a data-driven approach for clinicians to provide better prevention and treatment strategies to persons with DM to avoid foot related complications and, improve health and quality of life.