Graph theoretical analysis of EEG effective connectivity in vascular dementia patients during a visual oddball task

•EEG-based brain networks were analyzed for vascular dementia (VaD) patients during a visual oddball task.•The parietal region showed weakened information flow and altered topological properties in VaD patients.•VaD patients have more random brain networks for information processing.

ObjectiveThe study was dedicated to investigating the change in information processing in brain networks of vascular dementia (VaD) patients during the process of decision making.MethodsEEG was recorded from 18 VaD patients and 19 healthy controls when subjects were performing a visual oddball task. The whole task was divided into several stages by using global field power analysis. In the stage related to the decision-making process, graph theoretical analysis was applied to the binary directed network derived from EEG signals at nine electrodes in the frontal, central, and parietal regions in δ (0.5-3.5 Hz), θ (4-7 Hz), α1 (8-10 Hz), α2 (11-13 Hz), and β (14-30 Hz) frequency bands based on directed transfer function.ResultsA weakened outgoing information flow, a decrease in out-degree, and an increase in in-degree were found in the parietal region in VaD patients, compared to healthy controls. In VaD patients, the parietal region may also lose its hub status in brain networks. In addition, the clustering coefficient was significantly lower in VaD patients.ConclusionsImpairment might be present in the parietal region or its connections with other regions, and it may serve as one of the causes for cognitive decline in VaD patients. The brain networks of VaD patients were significantly altered toward random networks.SignificanceThe present study extended our understanding of VaD from the perspective of brain functional networks, and it provided possible interpretations for cognitive deficits in VaD patients.