This manuscript presents a multimodal deep learning pipeline for real-time decoding of Parkinsonian gait states, enabling predictive biomarkers that can drive adaptive DBS controllers.

Contribution (First Author): Conceptualization, methodology design, implementation, experimental evaluation, and manuscript drafting.

The paper documents the CaliBrain software stack, outlining its benchmarking datasets, uncertainty metrics, simulation engines, and visualization utilities for reproducible evaluation of source imaging algorithms.

Contribution (First Author): Software architecture, benchmarking framework development, data simulation, uncertainty calibration analysis, documentation, and tutorials.

The abstract introduces a generative modeling approach that upscales sparse sensor configurations to virtual high-density fNIRS measurements, improving spatial specificity for mobile neuroimaging experiments.

Contribution (First Author): Model development, experimental protocol design, implementation, quantitative evaluation, and performance analysis.

This work benchmarks Bayesian and frequentist inverse solvers on synthetic and empirical datasets to quantify calibration fidelity across multiple uncertainty metrics.

Contribution (Second Author): Benchmarking framework implementation, methodological analysis, statistical evaluation, and manuscript drafting.

The poster summarized the pipeline for simulating brain sources, propagating them through head models, and evaluating calibration metrics that inform practical sensor design choices.

Contribution (Second Author): Poster preparation, calibration experiments, and evaluation pipeline development.

This software poster showcased recently merged multivariate connectivity estimators, real-data tutorials, and validation utilities that shipped with the mne-connectivity package.

Contribution (Second Author): Implementation and extension of the MNE-Python connectivity toolbox.

This project involved constructing a lead-field matrix for the AAL2 atlas to be used within the neurolib framework for whole-brain modeling.