Association of structural brain networks with cognition in suspected epilepsy and epilepsy surgery
The human brain can be modelled as a network of interconnected anatomical regions using structural magnetic resonance imaging data. Understanding variations in network structure could help understand variations in higher brain functions such as cognition. This thesis investigates whether brain network structure is associated with cognition in patients investigated for suspected epilepsy and in those undergoing epilepsy surgery. Epilepsy is frequently associated with cognitive impairments and seizures are known to disrupt both structural and functional brain networks. First, a systematic literature review was undertaken to compare structural brain networks in epilepsy with healthy controls. Patients with epilepsy were found to have less efﬁcient networks with increased average path lengths compared to healthy controls. Networks constructed from cortical thickness covariance also showed increased clustering coefﬁcients compared to controls. Second, the association between network structure and cognitive dysfunction in a cohort of children undergoing investigation for suspected epilepsy was analysed. Patients with cognitive dysfunction had networks with longer average path lengths, longer normalised average path lengths, and lower global efﬁciency, even after controlling for the number and weight of network edges. These ﬁndings were consistent across network construction methods. Third, in a cohort of children undergoing resective epilepsy surgery, a post-operative increase in intelligence quotient was associated with increased global efﬁciency in the structural network within the healthy, contralateral non-operated hemisphere. Although cognition was associated with clinical features such as seizure frequency and age at onset of seizures, differences in network characteristics could not be completely explained by differences in clinical features. Different modelling techniques created different representative models but ﬁndings were broadly consistent across model types. This thesis suggests that the widespread alterations in brain structure described in epilepsy may lead to less efﬁcient brain networks which could contribute to cognitive dysfunction. Adequately treating seizures in those with an efﬁcient underlying brain network structure may facilitate cognitive development and allow patients to achieve their cognitive potential.