# Enhancing Brain Source Reconstruction by Initializing 3D Neural Networks with Physical Inverse Solutions

Source: [Enhancing Brain Source Reconstruction by Initializing 3D Neural Networks with Physical Inverse Solutions](<https://www.sota2.com/research/paper/68fa75751e209131cea68155>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-13T05:00:04.336Z

Authors: Marco Morik, Ali Hashemi, Klaus-Robert M\\"uller, Stefan Haufe, Shinichi Nakajima

Published: 2024-10-31T18:43:38.000Z

## Abstract

Reconstructing brain sources is a fundamental challenge in neuroscience, crucial for understanding brain function and dysfunction. Electroencephalography (EEG) signals have a high temporal resolution. However, identifying the correct spatial location of brain sources from these signals remains difficult due to the ill-posed structure of the problem. Traditional methods predominantly rely on manually crafted priors, missing the flexibility of data-driven learning, while recent deep learning approaches focus on end-to-end learning, typically using the physical information of the forward model only for generating training data. We propose the novel hybrid method 3D-PIUNet for EEG source localization that effectively integrates the strengths of traditional and deep learning techniques. 3D-PIUNet starts from an initial physics-informed estimate by using the pseudo inverse to map from measurements to source space. Secondly, by viewing the brain as a 3D volume, we use a 3D convolutional U-Net to capture spatial dependencies and refine the solution according to the learned data prior. Training the model relies on simulated pseudo-realistic brain source data, covering different source distributions. Trained on this data, our model significantly improves spatial accuracy, demonstrating superior performance over both traditional and end-to-end data-driven methods. Additionally, we validate our findings with real EEG data from a visual task, where 3D-PIUNet successfully identifies the visual cortex and reconstructs the expected temporal behavior, thereby showcasing its practical applicability.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Source Localization on 2000 single source activations 20 dB SNR (val)](<https://www.sota2.com/research/sota/source-localization-on-2000-single-source-activations-20-db-snr-val.md>): Source Localization; 2000 single source activations 20 dB SNR (val); EMD: 0.19.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2411.00143>)

[PDF](<https://arxiv.org/pdf/2411.00143.pdf>)
