A Reproducible Approach to the Analysis of Data Consistency in Neuroimaging Using Open-Source Tools

A Reproducible Approach to the Analysis of Data Consistency in Neuroimaging Using Open-Source Tools

Zvereva A.K. (MIPT, Dolgoprudny, Moscow Region, Russia)
Grabovoy A.V. (MIPT, Dolgoprudny, Moscow Region, Russia; ICS RAS, Moscow, Russia)
Kaprielova M.S. (MIPT, Dolgoprudny, Moscow Region, Russia)

Abstract

Studies involving multi-source data analysis often require representation harmonization to address discrepancies caused by technical differences in data acquisition. We describe a reproducible software pipeline for multi-echo functional MRI (fMRI) data, aimed at mapping signals from different echo channels into a common, aligned latent space. The pipeline is based on open data and tools (BIDS, DataLad) and includes preprocessing, time-windowing, and training of lightweight representation alignment models. Efficacy is evaluated using a protocol based on the gain in inter-echo first principal component (ΔPC1) correlation. Using a sample of 100 sessions with a fixed configuration of procedures and hyperparameters, the median ΔPC1 gain was ≈ +0.11, and the proportion of sessions with a positive effect was ≈ 61%, confirming a moderate but consistent improvement. The 95% bootstrap CI for the median does not include zero. This pipeline serves as an open, reproducible baseline for comparison against more complex harmonization methods.

Keywords

software pipeline; reproducible analysis; data agreement; representation alignment; multi-echo fMRI; neuroimaging; latent representations; DataLad; BIDS.

Edition

Proceedings of the Institute for System Programming, vol. 38, issue 3, part 1, 2026, pp. 269-282

ISSN 2220-6426 (Online), ISSN 2079-8156 (Print).

DOI: 10.15514/ISPRAS-2026-38(3)-17

For citation

Zvereva A.K., Grabovoy A.V., Kaprielova M.S. A Reproducible Approach to the Analysis of Data Consistency in Neuroimaging Using Open-Source Tools. Proceedings of the Institute for System Programming, vol. 38, issue 3, part 1, 2026, pp. 269-282 DOI: 10.15514/ISPRAS-2026-38(3)-17.

Full text of the paper in pdf (in Russian) Back to the contents of the volume