In within-dataset classification on Dataset A, combining real and harmonized EEG improved accuracy from 56.57% to 66.20% for EEGNet and from 61.96% to 72.54% for FBCNet, and the harmonized representation supported both source-only cross-dataset transfer and target-domain fine-tuning, providing a practical basis for signal-level harmonization and cross-dataset MI decoding. PMID 42537670, DOI 10.1088/1741-2552/ae9344
Graph Convolutional Network-Based Harmonization of EEG for Cross-Dataset Transfer in MI-BCI
Summary
The study presents a spatial harmonization framework built on a two-layer graph convolutional network (GCN) that maps heterogeneous EEG recordings onto a unified physical electrode layout while preserving motor imagery information, addressing electrode-configuration mismatches across MI-BCI datasets. Each trial is modeled as a graph so the GCN captures spatio-temporal relations, and harmonized EEG showed lower error than spherical spline interpolation while retaining the key temporal-spectral-spatial features. Combining real and harmonized EEG lifted EEGNet accuracy from 56.57% to 66.20% and FBCNet from 61.96% to 72.54% in within-session classification on Dataset A, and supported source-only cross-dataset transfer and target-domain fine-tuning.
Why it matters
Harmonizing heterogeneous montages at the signal level — instead of retraining per device — is what lets motor imagery decoders transfer across datasets and hardware, a prerequisite for moving beyond a single lab setup.
Compiled by BCIwiki from public sources
Sources · 1
pubmed.ncbi.nlm.nih.gov 2026-08-01