Biosignal Processing Group

© TIMELY-Consortium

Coronary artery disease (CAD) remains the leading cause of disease burden globally. CAD develops slowly, usually over decades, and depends on multiple (often modifiable) risk factors and their interactions. Self-management and patient activation are of rising importance as current restrictions in healthcare budgets impose great difficulties to enable the provision of qualitative secondary prevention to all cardiac patients in an era facing a huge cardiovascular disease epidemic.

The main hypothesis in the patient-centered TIMELY pathway, is that a modular, collaborative eHealth platform, supported by Artificial Intelligence (AI) for the continuous and in-time prediction of cardiac risks and complications and the induction of targeted behavioural change interventions, can be effective and cost-efficient for the secondary prevention of CAD by limiting the physiological and psychological effects of the disease and improving risk factor and symptom management. Improvements in patients’ self-care and empowerment and clinicians’ efficiency are also expected.

Along the continuum of the disease, prediction of the individual risk for disease progression, including physical impairment and severe events, is mandatory for timely intervention. TIMELY is a platform that provides AI-powered apps and dashboards and decision support tools assisting patients and clinicians to personalize healthcare based on risk evaluation, outcome prediction and tailored interventions. The platform will be developed based on a functional platform for Interoperability with electronic health records and security mechanisms, to ensure information completeness and continuity and to simplify data sharing. AI in TIMELY, built with big retrospective datasets of >23.000 CAD patients, will constantly monitor and evaluate risks and will indicate any deviation from defined therapy goals or unfavorable changes as well as propose proper interventions.

This project has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under Grant agreement ID: 101017424. Further informations can be found at the website of the project.

The autonomic nervous system (ANS) controls numerous body functions and shows specific reactions depending on personal conditions. The assessment of the ANS allows extensive statements, for example on the risk of suffering life-threatening illnesses, an assessment of sleep and the recording of stress levels.

Our group is developing novel methods of signal processing in order to be able to make such statements. In addition to the currently established analysis of heart rate variability, we focus at novel measures e.g. the QT interval variability. For their evaluation, innovative approaches have been and are being developed. The aim is a substantial improvement of the diagnostic possibilities, in particular the prediction and prevention, without creating an additional burden for the patients.

As part of the project, a close cooperation is maintained with the group of Prof. Mathias Baumert (The University of Adelaide). More information about the project can be found on the website of Two-Dimensional Signal Warping (2DSW).

© Martin Schmidt (Own work) [CC BY-SA 4.0]

Publications

• M. Schmidt, M. Baumert, T. Penzel, H. Malberg, und S. Zaunseder, „Nocturnal ventricular repolarization lability predicts cardiovascular mortality in the Sleep Heart Health Study“, American Journal of Physiology-Heart and Circulatory Physiology, Bd. 316, Nr. 3, S. H495–H505, Dez. 2018, doi: 10.1152/ajpheart.00649.2018.

• M. Schmidt, M. Baumert, H. Malberg, und S. Zaunseder, „T Wave Amplitude Correction of QT Interval Variability for Improved Repolarization Lability Measurement“, Front. Physiol., Bd. 7, 2016, doi: 10.3389/fphys.2016.00216.

• M. Schmidt, M. Baumert, H. Malberg, und S. Zaunseder, „Iterative two-dimensional signal warping—Towards a generalized approach for adaption of one-dimensional signals“, Biomedical Signal Processing and Control, Bd. 43, S. 311–319, Mai 2018, doi: 10.1016/j.bspc.2018.03.016.

• M. Schmidt, M. Baumert, A. Porta, H. Malberg, und S. Zaunseder, „Two-Dimensional Warping for One-Dimensional Signals—Conceptual Framework and Application to ECG Processing“, IEEE Transactions on Signal Processing, Bd. 62, Nr. 21, S. 5577–5588, Nov. 2014, doi: 10.1109/TSP.2014.2354313.

© Felix Gräßer

Especially in sleep medicine, but also in non-clinical applications, the evaluation of sleep quality plays a crucial role. In this context, the Machine Learning group is engaged with the automated classification of sleep stages, the detection of sleep-related breathing and sleep disorders and the automated derivation of recommendations for action. Modern methods of pattern recognition, machine learning (ML) and artificial intelligence (AI) are applied.

© Felix Gräßer

In the context of the research project Teleschlafmedizin, the research group is developing a telemedical platform for sleep medicine together with partners from the clinic and industry. The project comprises three main development focuses:

• Advancement, adaptation and validation of contactless medical measurement technologies for sleep medicine applications.
• Integration of various diagnostic and therapeutic biosignals and data from sleep medicine in a medical database (data integration). This concerns polysomnographic data from the clinic for neurology as well as monitoring data from the home environment.
• Development of a telemedical system that uses AI methods in order to automate data analysis and thus render it simpler and more practical.

© efre

The project is funded by the European Regional Development Fund (EFRE).

© ESF

The goal of the ESF junior research group MEDICOS (Medical Implant Communication System) is the development of extremely energy-efficient, compact and adaptive communication and biosignal processing systems, and thus the creation of a technical basis for a new generation of intelligent implants:

• New system concepts for risk prediction and prevention
• Implant networking and better data connection
• Powerful data processing strategies
• Long service life due to high energy efficiency

The project is granted by the European Social Fund (ESF) and from tax revenue on the basis of the budget approved by the members of the Landtag of the Free State of Saxony.

© fast

The aim of the fast athletics project is to optimize the biomechanical and sport physiological real-time performance diagnostics in recreational and professional sports for the movement-controlled self-education process and for the broadcast through the integration of

• novel body-near sensor infrastructures and control principles and
• novel training methods for sensory perception and motion control.

© Daniel Schwensow

The project addresses the technical and sports physiological fundamentals (methods, systems, knowledge) of a real-time physiological biofeedback (focus on real-time electromyogram, real-time EMG) in order to optimize the training with sport-related cyclically executed movements, such as rowing, cross country skiing or cycling. For this purpose, a prototypical realization for the rowing ergometer is implemented and evaluated.

Cluster members:

© fast

© BMBF

Further informations can be found at the website of the project

Fetal monitoring is currently limited in various ways. Consequently, there is a special need for non-invasive and simple methods that can be applied over longer periods, ideally also outside the clinic. Solutions eventually provide the needed all-day applications.

© Jonas Weiß

The fetal electrocardiogram (fECG), obtained through abdominal electrodes, is a method which tackles those problems. Therefore, electrical signals are recorded from electrodes placed on the belly of the mother. The signals contain the maternal ECG, fECG and noise. The separation of those signals is still mainly the subject of research.

This problem of source separation was addressed in the research group on an internationally high level. This success is particularly evident in two first places in the Computing in Cardiology Challenge 2013. Further research results on simulation, extraction and quality assessment of fetal ECG data were combined in a toolbox (FECGSYN). 

© Jonas Weiß

Since 2018, work has been taking place within the DFG project "Enhanced Fetal Monitoring". A large database of longitudinal recordings during pregnancy is being established in cooperation with the University Hospital Leipzig and the Humbold University Berlin.

The focus of future work includes the development of innovative algorithms extract the fECG and do further research on characteristics of the fECG itself. In addition, we test novel fECG based markers to identify fetal development and predict pregnancy complications. Overall, we seek to develop new fetal monitoring techniques and systems.

Publications

© DFG

• J. Weiß, H. Malberg, und M. Schmidt, „Detection Quality Indices for Improved Heart Beat Assessment in Non-Invasive Fetal ECG“, in 2020 Computing in Cardiology, Sep. 2020, S. 1–4, doi: 10.22489/CinC.2020.231.

• F. Andreotti, F. Gräßer, H. Malberg, und S. Zaunseder, „Non-invasive Fetal ECG Signal Quality Assessment for Multichannel Heart Rate Estimation“, IEEE Transactions on Biomedical Engineering, Bd. 64, Nr. 12, S. 2793–2802, Dez. 2017, doi: 10.1109/TBME.2017.2675543.

• J. Behar, F. Andreotti, S. Zaunseder, J. Oster, und G. D. Clifford, „A practical guide to non-invasive foetal electrocardiogram extraction and analysis“, Physiol. Meas., Bd. 37, Nr. 5, S. R1–R35, Apr. 2016, doi: 10.1088/0967-3334/37/5/R1.

• F. Andreotti, J. Behar, S. Zaunseder, J. Oster, und G. D. Clifford, „An open-source framework for stress-testing non-invasive foetal ECG extraction algorithms“, Physiol. Meas., Bd. 37, Nr. 5, S. 627–648, Apr. 2016, doi: 10.1088/0967-3334/37/5/627.