Chaotic internal dynamics coexist with a stable temporal scaffold in a mesoscale sarcomere model informed by high-resolution recordings

Abstract

During hyperthermal sarcomeric oscillations (HSOs), serial sarcomeres differ in amplitude and phase. Their fast length changes can cancel in total length, while changing relative amplitudes can move the best phase arrangement. Each model node represented one sarcomere. I tested two requirements: correct amplitude-to-sarcomere pairing and timely phase adjustment. Lower normalized residuals meant better cancellation; subtracting the minimum set by the current amplitudes isolated the part that phase adjustment could remove. Correct pairing reduced the 95th percentile of this avoidable part relative to amplitude-blind or misassigned inputs in all 20 prespecified conditions. Dynamic adjustment outperformed the best of 13 fixed arrangements in all 12 conditions with target periods of at least four HSO cycles, but only 2/8 faster conditions. A ratio of target-change time to model response time almost perfectly ranked dynamic wins above fixed wins in held-out and speed-limit tests (AUC 0.976 and 0.996), although the best cutoff differed among condition families. Replaying measured amplitude histories from five sarcomeres in seven cardiomyocytes gave a median avoidable residual of 0.000802, about 1/78 of the lowest fixed-strategy median, and outperformed three fixed strategies in 7/7 cells. Returning each history to its source sarcomere gave the lowest residual among all 120 assignments in every cell. Observed phase motion favored the predicted direction relative to the circular-shift median in 5/7 cells; the prespecified cell-level test gave P=0.1094. Within this reduced model, dynamic phase balancing requires correct node-specific information and sufficient response time. The study establishes this conditional model capacity and identifies native phase dynamics as the next mechanochemical test.