Secondary Computational Reanalysis of Published Brownian Trajectories: Predictive Information, Temporal Conditioning, and Robustness Assessment

This repository presents an independent methodological reanalysis of experimental Brownian trajectory data originally published by Lokesh Muruga, Felix Ginot, Sarah A. M. Loos, and Clemens Bechinger in Nature Physics (2026), under the title The Extraction of Work from Hidden Degrees of Freedom. The original study experimentally demonstrated how environmental memory stored in hidden degrees of freedom of a viscoelastic medium can influence the relaxation dynamics of a trapped Brownian particle and contribute to enhanced work extraction. The experimental findings, original measurements, and physical discovery are fully attributed to the original authors. Nature and Scope of the Contribution This work constitutes a secondary computational and methodological reanalysis of previously published experimental data. It does not represent an independent experimental replication or validation of the original physical discovery. No new experimental measurements were performed. The experimental design, original measurements, and demonstration of environmental memory and enhanced work extraction are entirely attributed to Muruga, Ginot, Loos, and Bechinger (2026). The present contribution is limited to complementary predictive modeling, temporal-history analysis, cross-recording validation, and robustness assessments conducted using the publicly available datasets. The reported predictive improvements and history-conditioned contrasts should be interpreted within the stated methodological limitations. Exploratory findings are not presented as independently confirmed physical discoveries. Objective The present work provides a complementary computational investigation of the temporal information contained in the published experimental trajectories, focusing on predictive sufficiency, historical-state dependence, conditional relaxation, and methodological robustness. Data and Methodology The reanalysis uses the publicly released experimental datasets, including Lokesh_IE_Data_V2.zip and Lokesh_Information_Data.zip, together with the authors' processed data and MATLAB analysis materials. Nine consecutive experimental recordings, each containing 360,000 position measurements, were examined across five computational analysis stages (M1–M5). The methodology includes: Reproduction and examination of published memory and normalized work curves. Predictive comparisons between present-state models and models incorporating previous particle positions. Leave-one-recording-out validation using extended autoregressive models and synthetic Markovian controls. Analysis of history-conditioned relaxation with matched initial-position distributions. Robustness assessment using repeated sampling, alternative matching tolerances, and comparisons between short and long measurement intervals. Principal Results The initial predictive analysis showed an average 9.61% reduction in mean squared prediction error when historical position information was incorporated into the reference model. Under a more demanding autoregressive baseline and leave-one-recording-out evaluation, an average improvement of approximately 3.13% remained, with positive improvements across all nine analyzed recordings. The history-conditioned relaxation analysis also identified an exploratory temporal contrast between the 11 and 01 measurement histories. This contrast showed a mean difference of approximately 0.375 kBT between measurement separations of 1 and 200 seconds, with consistent direction across the nine recordings under the evaluated matching conditions. These findings quantify predictive and statistical properties of the published experimental data. They do not establish a previously unknown physical memory mechanism or independently replicate the experimental discovery. Scientific Scope and Limitations The results constitute a secondary methodological contribution rather than a new experimental discovery. The analyzed recordings were acquired consecutively and should not automatically be treated as statistically independent experiments. Some robustness comparisons were defined after exploratory analysis and must therefore be interpreted as exploratory rather than confirmatory. Predictive improvements depend on the model classes, measured observables, and validation procedures. They do not, by themselves, uniquely establish the physical origin of environmental memory. Repository Contents The deposited materials include an English-language scientific manuscript, a technical methodology and reproducibility report, Python analysis scripts covering stages M1–M5, a consolidated execution workflow, derived numerical results, figures, and supporting documentation. The computational workflow is designed for reproducible execution in Google Colab using the original publicly available experimental datasets. Original Research and Data Attribution Muruga, L., Ginot, F., Loos, S. A. M., and Bechinger, C. (2026). The Extraction of Work from Hidden Degrees of Freedom. Nature Physics. Original article: https://doi.org/10.1038/s41567-026-03442-w Original experimental data repository: https://doi.org/10.5281/zenodo.19881972 All experimental data provenance, original experimental methods, and primary physical findings are credited to the original research team. Contribution This independent reanalysis contributes a reproducible framework for evaluating historical information, out-of-sample predictability, and robustness of conditional relaxation measurements in experimentally observed Brownian dynamics.