An efficient Gait Dynamics classification method for Neurodegenerative Diseases using Brain signals

Neurons of the human brain are primarily affected by the Huntington's disease (HD), Amyotrophic Lateral Sclerosis (ALS), Parkinson's disease and so on. Classification of these neurodegenerative diseases (NDD) is clinically important to analyze the destruction of nerve cells. Early diagnosis of NDD'S helps in saving the human life. Based on the report of previous studies, motor impairment or human gait cycle is largely affected by the clinical symptoms of NDD. Accurate diagnosis of various neurodegenerative diseases in correct time is very important for early diagnosis of the disease. Diseases can be diagnosed earlier by means of characterizing the gait cycle. In this work, a gait dynamics classification method is proposed for determining the neurodegenerative diseases from the brain signals using multilevel feature extraction method. From force sensitive resistors, the left and right feet signals recorded in 60 one minute are included in the input database. It is obtained through fixing 16 healthy subjects, 13 ALS, 20 HD, and 15 PD. Using six levels of Discrete Wavelet Transform (DWT), the features are determined by means of decomposing the raw signal. Ultimately, the pathological gait signals are classified through exploiting three multilevel feature extraction techniques named as, (Detrended Fluctuation Analysis (DFA), Positive, Negative Peak Histogram Analysis (PNPHA) (proposed Method) and Statistical Temporal parameter Analysis (STA)). Experimental outcomes proved that the gait dynamics are successively distinguished between NDD and group of healthy controls using the proposed method.