Automated Electromechanical Wave Imaging at Reduced Frame Rates During Sinus Rhythm Using Machine Learning

Spatial resolution is typically prioritized over temporal in conventional echocardiography, which restricts the use of time-shifted based techniques at higher frame rates (FR). Electromechanical Wave Imaging (EWI) is an echocardiography-based modality that non-invasively maps the cardiac activation sequence at a high frame rate. At lower FRs, EWI strain curves have a different profile which constrains manual input and interpretation. In this study, we investigate how machine learning (ML) can assist in the accurate activation time estimation at low FR data. We show that with the use of a Random Forest classifier for automated EWI estimation, we successfully identify the normal, early activated basal septum in normal sinus rhythm lower FR cases ( $\mathrm{N}=4$ , FRs of 500 Hz, 250 Hz, 125 Hz) while maintaining the electromechanical activation propagation pattern over the rest of the myocardium. These findings indicate that ML can be used to generate accurate EWI activation maps with a more flexible FR range that includes clinical systems at low FRs, where manual processing would otherwise be impractical and/or inaccurate.