Fusion of EEG and fMRI for the investigation of functional connectivity during a visual oddball task

The combined use of EEG and fMRI allows the fusion of electrophysiological and hemodynamic information for the study of the human brain function. In order to investigate functional connectivity during a visual oddball task, we performed simultaneous EEG/fMRI recording from 14 healthy subjects. A completely data-driven approach, based on trial-by-trial coupling of concurrent EEG and fMRI, was used for the retrieval of cerebral networks linked to P300. An ICA-based analysis of fMRI BOLD data allowed to characterize neuro-anatomically defined patterns, each of them with a definite spatial map and a specific time-course of activation. In addition, we performed single-trial analysis on EEG data, reconstructing the temporal evolution of the P300 amplitude. This time-course, convolved with a canonical HRF, was considered as an estimate of the P300-related activity in the BOLD signals, and allowed to identify two cerebral networks linked to target detection. They showed spatial maps already observed in previous studies, which could be respectively associated with the dorsal attention and the change-detection networks. Time-domain analysis of BOLD data suggested that the attention network might account for sustained activity reflecting amplitude modulations of P300, whereas the change-detection network for transient activity associated with the event-related P300 response.