Higher order dynamic mode decomposition: From fluid dynamics to heart disease analysis.

Area of Science:

• Cardiology
• Biomedical Engineering
• Data Science

Background:

• Higher Order Dynamic Mode Decomposition (HODMD) is a data-driven technique typically used in fluid dynamics and complex non-linear dynamical systems.
• Echocardiography is a crucial imaging modality for assessing cardiac function and diagnosing heart conditions.

Purpose of the Study:

• To evaluate the performance of HODMD for pattern recognition in echocardiography images.
• To apply HODMD for identifying characteristic patterns associated with various cardiac diseases in mice.

Main Methods:

• HODMD was applied to echocardiography video loops (LAX and SAX views) from healthy mice and those with conditions like Diabetic Cardiomyopathy, Obesity, Hypertrophy, and Myocardial Infarction.
• The algorithm analyzed datasets using a maximum of 200 snapshots to identify dominant frequencies (heart and respiratory rates) and patterns (modes).

Main Results:

• HODMD successfully captured distinct frequency branches corresponding to heart and respiratory rates.
• The technique identified specific modes representing dominant features related to cardiac and pulmonary function in different disease states.
• Six datasets, including healthy controls and five disease models, were analyzed, demonstrating HODMD's capability.

Conclusions:

• HODMD is a robust and suitable tool for pattern recognition in echocardiography.
• The method can effectively identify characteristic patterns for classifying different cardiac pathologies.
• This study represents the first known application of HODMD for pattern recognition in echocardiography.