Search papers, labs, and topics across Lattice.
This paper introduces a Layer-wise Interactive Dual-Stream Network (LI-DSN) for EEG decoding, addressing the "information silo" problem in existing dual-stream networks by enabling progressive, cross-stream communication at each layer. LI-DSN incorporates a Temporal-Spatial Integration Attention (TSIA) mechanism, using a Spatial Affinity Correlation Matrix (SACM) and a Temporal Channel Aggregation Matrix (TCAM) to capture spatial and temporal relationships, respectively. Experiments across eight EEG datasets demonstrate that LI-DSN outperforms 13 state-of-the-art baselines in motor imagery classification, emotion recognition, and steady-state visual evoked potentials.
By fusing temporal and spatial EEG features layer-by-layer, LI-DSN shatters the "information silo" problem of traditional dual-stream networks, unlocking significant gains in decoding accuracy across diverse BCI tasks.
Electroencephalography (EEG) provides a non-invasive window into brain activity, offering high temporal resolution crucial for understanding and interacting with neural processes through brain-computer interfaces (BCIs). Current dual-stream neural networks for EEG often process temporal and spatial features independently through parallel branches, delaying their integration until a final, late-stage fusion. This design inherently leads to an"information silo"problem, precluding intermediate cross-stream refinement and hindering spatial-temporal decompositions essential for full feature utilization. We propose LI-DSN, a layer-wise interactive dual-stream network that facilitates progressive, cross-stream communication at each layer, thereby overcoming the limitations of late-fusion paradigms. LI-DSN introduces a novel Temporal-Spatial Integration Attention (TSIA) mechanism, which constructs a Spatial Affinity Correlation Matrix (SACM) to capture inter-electrode spatial structural relationships and a Temporal Channel Aggregation Matrix (TCAM) to integrate cosine-gated temporal dynamics under spatial guidance. Furthermore, we employ an adaptive fusion strategy with learnable channel weights to optimize the integration of dual-stream features. Extensive experiments across eight diverse EEG datasets, encompassing motor imagery (MI) classification, emotion recognition, and steady-state visual evoked potentials (SSVEP), consistently demonstrate that LI-DSN significantly outperforms 13 state-of-the-art (SOTA) baseline models, showcasing its superior robustness and decoding performance. The code will be publicized after acceptance.
