Latent Variate Factorial and Clustering Analysis of EMG Writer`s Cramp Neuromuscular Signals Dr. Raju Venkateshwarla Rama*, Sreeniva B. CMR College of Engineering & Technology (UGC Autonomous), Kandlakoya, Medchal Road, Hyderabad-501 401, Telangan State, India *Corresponding Author Email: drvrr@cmrcet.org
Online published on 20 December, 2018. Abstract In this study, multivariate techniques applied on EMG Writer`s Cramp signals data of 12 subjects. Data acquired with technically advanced intelligent EMG system from five muscles of right-hand, when subject inscribed first with intrinsic dominant right-hand and then, with innate non-dominant left-hand i.e., right hand writing signal (RHWS) and left hand writing signal (LHWS). Latent variate-factorial (or factor) principal components (PCs of differences of 12 subject's scores) and cluster analysis based on means and variances of twelve subjects were analyzed. Eigen values and corresponding dominant Eigen vectors were computed. In our computation, 12 subjects (subjects) points are well scattered out, without clear-cut-pattern. While cluster analysis based on dissimilarity among the subject's waveforms—signals showed a possibility that in addition to the groupings of subjects as C or D, some other groupings may also be meaningful. The data analysis made in this direction showed significant findings which led to attempts at more sophisticated analysis using advanced multivariate techniques leading to effective data summarization and measures of dissimilarity between subjects as reflected in the waveforms recorded and consequent possible clustering among them. However, these did not lead to any meaningful clinical conclusions. Hence, these analyses could be possibly applied to longitudinal follow ups, microanalysis and correlations with a normal control population in the future to better comprehend phenomenon of Writer's cramp at micro-structure-level. Top Keywords Electromyography, Writer's Cramp, Principal Component Analysis, Cluster analysis. Top |