EEG-fMRI fusion-based source localization for the identification and mechanistic elucidation of paroxysmal kinesigenic dyskinesia.
The GTBL-AF model provides a promising approach for improving PKD recognition through EEG-fMRI fusion, achieving high classification accuracy.
Where it sits
this study against the rest of the adamax corpusSummary and findings
The study developed a multimodal EEG-fMRI framework to improve source localization in paroxysmal kinesigenic dyskinesia (PKD). The GTBL-AF model achieved 94.2% classification accuracy in PKD recognition. This approach may enhance understanding of PKD pathophysiology.
Abstract
Paroxysmal kinesigenic dyskinesia is a rare neurological disorder characterized by brief, recurrent motor attacks that significantly impair quality of life. Prior studies have largely relied on unimodal data, which offer partial insights into neural regulation but are constrained by trade-offs between temporal and spatial resolution. To address this limitation, we developed a multimodal recognition and tracing framework integrating electroencephalography and functional magnetic resonance imaging. We propose GTBL-AF, a deep multimodal neural architecture that captures spatial connectivity and temporal dynamics of brain function through graph attention, Transformers, and bidirectional long short-term memory networks, with cross-attention enabling modality-level fusion. GTBL-AF achieved 94.2% classification accuracy in paroxysmal kinesigenic dyskinesia recognition, significantly outperforming unimodal methods. Incorporating dipole-based electroencephalography source localization and phase-locking value connectivity, we observed increased temporal complexity and reorganized functional connections in key cortical regions, including the prefrontal cortex, temporal pole, and parietal association areas. Whole-brain analyses using sample entropy and small-world metrics revealed greater dynamic uncertainty and enhanced small-world properties in paroxysmal kinesigenic dyskinesia patients, indicative of compensatory neural regulation. Furthermore, network-based statistics identified aberrant synchronous connectivity within circuits mediating cognitive control and motor initiation. This study presents a deep EEG-fMRI multimodal fusion framework for PKD and provides evidence of widespread network reorganization. These findings may contribute to a better understanding of PKD pathophysiology and provide a methodological reference for future multimodal-assisted diagnosis and individualized clinical assessment.<h4>Supplementary information</h4>The online version contains supplementary material available at 10.1007/s11571-026-10504-5.
Background
Paroxysmal kinesigenic dyskinesia (PKD) is a rare disorder characterized by sudden motor attacks, impacting quality of life. Previous studies have used unimodal data, which are limited by their inability to simultaneously capture high temporal and spatial resolution. This study aims to address these limitations by integrating EEG and fMRI data to provide a more comprehensive understanding of PKD pathophysiology.
Methods
The study developed a deep multimodal neural architecture called GTBL-AF, which integrates EEG and fMRI data using graph attention, Transformers, and bidirectional long short-term memory networks. The framework was designed to capture spatial connectivity and temporal dynamics, with cross-attention enabling modality-level fusion. The primary outcome was the classification accuracy of PKD recognition.
Results
The GTBL-AF model achieved a classification accuracy of 94.2% in recognizing PKD, significantly outperforming unimodal methods. The study observed increased temporal complexity and reorganized functional connections in key cortical regions, including the prefrontal cortex, temporal pole, and parietal association areas. Whole-brain analyses indicated greater dynamic uncertainty and enhanced small-world properties in PKD patients.
Interpretation
The study's findings suggest that the GTBL-AF model offers a significant improvement over unimodal methods in classifying PKD, with a high classification accuracy of 94.2%. However, the clinical significance of these findings remains unclear, as the study focuses on mechanistic insights rather than direct clinical outcomes. The use of advanced EEG-fMRI fusion techniques may not be feasible in all clinical settings, limiting the immediate applicability of the results.
Key findings
- 94.2% classification accuracy in PKD recognition.
- GTBL-AF significantly outperformed unimodal methods.
- Increased temporal complexity observed in PKD patients.
- Reorganized functional connections in prefrontal cortex, temporal pole, and parietal association areas.
- Enhanced small-world properties in PKD patients.
Limitations
- EEG-fMRI fusion may not be widely available.
- Clinical significance not directly addressed.
- Focuses on mechanistic insights rather than clinical outcomes.