Neuro-Degenerative Disease Diagnosis Using Human Gait

Abstract

Falls are one of the most serious implications of the gait disturbance in neuro-degenerative disease. Neurodegenerative diseases affect the ability to control muscle movements. Muscle tone, involuntary movements and smoothness of movement are significantly impacted while range of motion and muscle mass remain unaffected. Some of the major neuro-degenerative diseases include Parkinson’s Disease (PD), Amyotrophic Lateral Sclerosis (ALS), and Huntington Disease (HD). These diseases are generally unpredictable in their rate of progression and exhibition of degeneration. Neuro-degenerative diseases are caused by loss or death of neurons. These diseases are characterized by progressive nervous system dysfunction. This dissertation reviews how the human gait is related to neuro-degenerative diseases. Gait is basically the pattern of movement, means how we move or walk. For comparison of the gait pattern of patients suffering from neuro-degenerative disease, a healthy control group is taken. Heel and Toe strike detection and computation for gait analysis using computer aided programming for patients and healthy control was done After computing heel and toe strike intervals of each left and right foot, Mean and Standard Deviation (SD) were calculated to obtain the Coefficient of variation (CV) and compared with healthy control.The gait pattern of both left and right foot is recorded, analyzed and compared for detection and computation of neuro-degenerative disease diagnosis, using several techniques. Three phases of classification were used to classify the data more appropriately. First phase classifies normal and neuro-degenerative subjects. Second phase classify Parkinson and non-Parkinson subjects. Third phase classify Huntington and ALS subjects. Coefficient of Variation of left and right foot heel and toe strike intervals is taken as input vector to Artificial Neural Network (ANN) for these classifications. In phase 1, we obtained accuracy of 94%. In phase 2, accuracy was 100%, and in phase 3, we obtained accuracy of 88%.