Pune IISER scientists find a new method to detect heart abnormalities
Scientists at Indian Institutes of Science Education and Research (IISER) Pune have developed a new method to analyse heart dynamics which can potentially help in the diagnosis of abnormalities of the heart.
The study titled ‘Detecting abnormality in heart dynamics from multifractal analysis of ECG signals’ was released on the IISER website in December 2017. Published in Scientific Reports, an online independent publication under the Nature magazine, the paper has been authored by Snehal M Shekatkar, Yamini Kotriwar, KP Harikrishnan and G Ambika.
The analysis was based on the mathematical concept of fractals, which exhibit a complex never-ending pattern that are similar at any scale. These are created by repeated processes in ongoing feedback loop and can be observed in nature, like in the spiral patterns of sea shells, or branching pattern of trees.
Hence, these fractals have continued to be used by the researchers as frameworks to understand and define complex systems.
“As per our research, the pattern of heart dynamics that is displayed by electrocardiograph (ECG) signals and are read out for heartbeat, can be analysed and studied through the fractal analysis based on the non-uniform distribution of signal points in the underlying dynamics.
“In simple words, we are trying to understand the complexity of the heart, which when affected with any abnormality, has reduced complexity. Our method is to detect and diagnose the abnormality,” said Ambika from IISER Pune, who had been working on the project collaborating with KP Harikrishnan of Cochin College.
Both led the research on the characterisation of heart dynamics in order to distinguish between healthy and abnormal ECG signals. The project under Science and Engineering Research Board (SERB) was funded by department of science and technology (DST).
The team of researchers, had recreated the dynamics of the heart using derived measures from multi-fractal spectrum of ECG signals which led them to successfully categorise healthy and unhealthy data sets of signals. In order to build a model to bring more accuracy to the predictions, the study used a supervised machine learning approach, the IISER researchers said.
One of the prominent findings of the research, includes the indication that the dynamics of a healthy heart , despite high complexity, shows less variability. According to the researchers, this method could lead to a qualitative process of analysing ECG signals to facilitate more accurate diagnosis, therapy and continued check on patients.