ISDN2012_0290: A study of advanced multi‐channel ENMG in Writer's cramp and Parkinsonism

V. Rama Raju 1,2, R. Borgohain 1, S.N.N. Pandit 3, Rukmini Mridula 1 1 Department of Neurology, Nizam’s Institute of Medical Sciences (NIMS), Hyderabad (AP) 500082, India 2 CMR College of Engineering and Technology (CMRCET), CMR Group of Institutions (CMRGI), Hyderabad, India 3 Center for Quantitative Methods, OU College of Engineering, Osmania University, Hyderabad, India Background: Mirror movements (MMs) are seen in the right hand in Writers Cramp (WC) patients when writing with the left hand. These MMs can be similar to the original dystonic movement while writing with the right hand (concordant) or in the opposite direction (discordant). Objectives: The aim of present experimental investigation is to attempt to differentiate between those with concordant (C) and discordant (D) MMs in WC, in order to establish that there is a quantifiable difference between these two groups and to design and fabricate a multi-channel ENMG system. Materials and methods: This study was conducted at Nizam’s Institute of Medical Sciences a tertiary care centre. Study period was from August 1997 till 2003. A suitable multi-channel EMG system was designed and fabricated to record digitized EMG signals simultaneously with a set of five, intramuscular, nylon coated fine wire electrodes (50 m diameter). The basic signal data consisted of EMG-data gathered from 5 muscles of the right hand, when the patient wrote first with right hand and then, wrote with his/her left hand, i.e., right hand writing signal (RHWS) and left hand writing signal (LHWS). Duration of signal recording was 10 seconds, with 3 kHz sampling frequency, giving 30,000 readings for each muscle. Means analysis, differences in means, standard deviations, variances, t, F and p-values between RHWS and LHWS were compared using student t, 2 (Chi-square) and Fisher’s tests. Advanced statistical methods/techniques employed in the present study include singular value decomposition (SVD)/Eigen analysis, principal component analysis (PCA), distance function, simple and (hierarchical) clustering, canonical correlation/multidimensional scaling and EMG coherence. Results: 12 patients with Writer’s cramp (8 with concordant and 4 with discordant MMs) were assessed. On comparison of the measures of dispersion; D group had statistically significant difference between LHWS and RHWS (variance, standard deviation and F ratio) with a larger variance in RHWS, as compared to C group where variances and SD were equal or smaller in the RHWS compared to LHWS. Mean amplitudes for RHWS and LHWS for the same muscles, though differ significantly in statistical terms, showed a consistent pattern only in the fifth muscle with a larger mean amplitude on left side in all patients and were not of value in differentiating between concordant (C) and discordant (D) groups of patients. The principal component (PC) scores of the 12 patients showed 80% variance in our computation in the scatter plot diagram. The cluster analysis based on dissimilarity among the patients’ signals show a possibility that, in addition to the grouping of patients as C or D, some other groupings may also be meaningful. EMG–EMG coherence was assessed in the Writer’s cramp hand muscles, namely ECR, ECU, FCR, FCU, followed by 5th muscle. The coherence computed, evaluated and compared between flexor aspect of forearm and extensor aspect of intrinsic hand muscles was showed significant coherence in both groups. These observations suggest that the nature of EMG–EMG coherence in dystonia Writer’s cramp may be constrained by the descending motor systems, both in terms of their anatomical distribution and their frequency characteristics. Conclusions: This study showed significant quantifiable EMG differences in the signals seen while writing with the right and left hands between those Writer’s cramp patients with concordant mirror movements (C group) versus those with discordant mirror movements (D group). This was mainly seen in the measures of dispersion of the signal, i.e., standard dispersion, variances and their ratio (F-ratio). These were statistically significantly different between the two groups, C and D, and the pattern of differences were consistent with the hypothesis that the discordant group had a compensatory force which overcame the dystonic force resulting in the final abnormal posture. This was seen in the form of larger variances and standard differences in the RHWS in the D group as compared to the C group, as the dystonic and compensatory forces both contribute to the instability. These differences were robust and seen in every measure of dispersion, such as in the patterns of significance of f-values for ratios of variances. Cluster analysis and more sophisticated analyses using advanced multivariate techniques leading to effective data summarization and measures of dissimiliarity between patients as reflected in the signals recorded and consequent possible clustering among them, however, did not lead to any meaningful clinical conclusions. These analyses could possibly be applied to longitudinal follow-ups and correlations with a normal control population in future to better comprehend the phenomenon of Writer’s cram.