Energy entropy feature for the discrimination between the patients with amyotrophic lateral sclerosis and healthy subjects

Amyotrophic Lateral Sclerosis (ALS) is a type of neurological disease due to the degeneration of motor neurons. Hypothesising that ALS patient will experience an altered gait rhythm, with changes in the magnitude of the stride-to-stride fluctuations and perturbations in the fluctuation dynamics, the gait rhythm of ALS patients is compared with those of healthy subjects using energy entropy (EE) feature and approximate entropy (ApEn). It is found that EE (ApEn) was increased (decreased) in ALS patients compared to EE feature in healthy controls. It is found that in the diagnostic ability test ApEn performed well with AROC = 0.8222, accuracy = 83.3%, specificity = 66.7%, sensitivity = 93.3%, and precision = 85.7%, while EE feature outperformed with AROC = 0.9803, accuracy = 93.8%, specificity = 94.1%, sensitivity = 96.3%, and precision = 96.1%. Results also showed that the diagnostic performance of EE feature is much superior compared to that of ApEn in discriminating ALS from patients with Parkinson and Huntington diseases.