Use of SPART method for automatic rhythm analysis in presence of body position changes and respiration artifacts

Accurate identification of ectopic beats is important in frequency HRV analysis in presence of ECG signal artifacts related both to body position changes and respiration. The objective of this research was to show results of QRS morphology analysis obtained in different spatial planes: frontal, horizontal and optimal. The optimal plane of QRS loop was found with SPART method. Fourier descriptors of plane loop were used for classification of QRS loops morphology. The authors distinguished three morphological beats categories: normal, ventricular and supraventricular. The XYZ ECG signals of 18 patients after complete repair of tetralogy of Fallot recorded during supine controlled respiration (6/min) and after +60/spl deg/ passive head-up tilt (spontaneous respiration) were examined. All ECG signals were drawn from the authors' own multilead IGS-WH database. The prevalence of abnormal beats was 0.81%. The SPART identification of optimal QRS spatial loop orientation was accurate in all analyzed beats. The study showed changes in spatial loops morphology. Statistical indices showed low values of QRS classification in the frontal plane. On contrary, the results obtained for XZ and UV planes were comparable. The SPART identification of QRS loop optimal spatial orientation was excellent. The study confirmed usefulness of the method for automatic rhythm analysis in the presence of body position changes and respiration artifacts. However, further investigations for optimization of classification algorithm are needed.