CardiacState: Learning Cyclic Cardiac State through Predictive Video Pretraining

Predicting how an observed physical system will evolve requires inferring the underlying state that governs its motion. While controllable video generation relies on external action or text signals, action-free predictive models leave this conditioning implicit within high-dimensional representations. Autonomous biological systems, such as the beating heart, lack external controllers—their driving dynamics unfold continuously within the visual stream itself. Yet, core clinical tasks like keyframe localization and Left Ventricular Ejection Fraction (LVEF) estimation depend heavily on tracking this physiological state. We introduce CardiacState, a joint-embedding predictive architecture (JEPA) that addresses this by formulating physiological progression as an explicit, self-derived conditioning state on a continuous cyclic manifold $S^1$. By imposing a causal temporal bottleneck that denies the future-feature predictor direct access to historical frames, CardiacState forces all temporal context to route exclusively through this $S^1$ state alongside physical elapsed time $\Delta {\tau }_t$. Pretrained on 1.88 million clinical echocardiogram videos without external text, action labels, or hardware ECG gating, CardiacState advances both feature representation and structural interpretability. Its frozen visual representations achieve $4.87\pm 0.12\%$ LVEF MAE on EchoNet-Dynamic and $4.73\pm 0.18\%$ on EchoNet-Pediatric—outperforming prior self-supervised baselines ($5.76\%$ and $5.12\%$). Furthermore, unlike implicit representations, this explicit cyclic state directly enables non-parametric keyframe localization across diverse acoustic views via phase-anchor matching, reducing mean localization error from $2.25$ frames to $1.99\pm 0.02$ frames. Controlled ablations show that cardiac progression regularization is important for learning a useful cyclic coordinate, while the temporal bottleneck provides further gains in localization and representation quality. These results support cyclic state conditioning for predictive learning in quasi-periodic systems.