Our brain is a complex organ with different levels of interaction and therefore can be thought as a complex network. In high level interactions, the brain network is formed by interconnected areas called nodes and their connections are the links. Hubs play a key role in information processing in the brain. Brain networks can be extracted using several imaging modalities, here we focus on networks based on resting state fMRI (rsfMRI) where spontaneous brain activity is indirectly measured resulting in a blood-oxygen-level dependent (BOLD) signal for each voxel. Brain regions or voxels are said to be functionally connected and therefore, a link exists, if they present temporal correlation. The advantages of rsfMRI are the easiness of the acquisition suitable for children and clinical population and to be able to uncover networks related to spontaneous or "default mode" of the brain. Moreover, it has been shown that the resting state networks are impaired in psychiatric and neurological disorders. Rolandic epilepsy (RE) is one of the most common epilepsy in childhood manifesting abnormal EEG activity in central-temporal areas. Despite seizure remission during adolescence, recent studies have shown a serious of comorbidities. Moreover, the risk of cognitive impairments has been linked to inter-ictal epileptic discharges (IED). Nevertheless, the underlying mechanisms are not fully understood. Here, we applied two novel methods to resting state fMRI. the blind deconvolution method to recover the neural activity and to extract the hemodynamic response (HRF) and functional connectivity density (FCD). FCD is a data-driven voxel-wise new tool combining graph theory and functional connectivity that unveils densely connected regions that can work as functional hubs of information in the brain. The goal was to identify hubs of information flow and possible network disruption in RE in patients with and without IEDs. FCD maps revealed main hubs in the posterior cingulate, precuneus, cuneus and calcarine. Patients with IEDs during the scanner showed higher FCD as compared to healthy controls and larger hub in the postcentral precentral gyri, key focal areas in RE. Patients with no IEDs during the scanner showed overall lower FCD as compared to controls and IED groups. Group comparison revealed hyper local connectivity in bilateral thalamus in the patients with IEDs compared to patients without IEDs. Additional exploratory HRF analysis showed that patients with IEDs presented higher response height in the HRF in the thalamus evidencing the inhomogeneity of the HRF among groups. We speculate that locally abnormal information flow in bilateral thalamus might suggest the involvement of this region in the generation of spikes in RE. It also provides additional evidence for an epileptic as a network disease rather than a focus dysfunction. This hypothesis could be further confirmed in meta analysis, small group size is the main limitation of this study. To the best of our knowledge, this is the first study to combine blind deconvolution and FCD to the whole brain analysis in RE.
展开▼
