Research in the Glen Group - Dr. Yuriy Chesnokov

Statistical complexity measure as a diagnosis tool for cardiac rhythms

The project is about screening and prediction of adverse heart conditions using statistical complexity of the heart data. We are trying to develop new insights into symbolic and time-frequency space of the cardiac data to reveal hidden patterns not obvious from conventional time domain analysis.

Research and application software code development:

• Automated computer annotation of the ECG data (P,QRS,T waves, PQ, QT, RR, RRn, ectopics, artefacts, noise), wavelet-transform based denoising, compression of the ECG data
• Statistical complexity analysis of the cardiac data, symbolization methods for cardiac data
• Development of multiscale complexity methods based on continuous and fast wavelet transforms for cardiac data
• Wavelet and statistical complexity analysis of the heart rate variability and EEG data, breathing induced changes in the signals
• AI classification methods (artificial neural networks, genetic algorithms, N-dimensional self-organizing maps)

Current applications:

• Automatic ECG annotation, 1st place in QT interval Physionet Computers in Cardiology 2006
• Paroxysmal atrial fibrillation screening and distant prediction with HRV data spectral analysis, automatic AI classification
• Autonomous failure patients screening with multiscale statistical complexity, automatic AI classification