Van Graan, Louis André; (2023) Functional network correlates of language and semiology in epilepsy. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Van Graan L. Andre. Functional network correlates of language and semiology in epilepsy.pdf - Other Download (6MB) | Preview

Abstract

Epilepsy surgery is appropriate for 2-3% of all epilepsy diagnoses. The goal of the presurgical workup is to delineate the seizure network and to identify the risks associated with surgery. While interpretation of functional MRI and results in EEG-fMRI studies have largely focused on anatomical parameters, the focus of this thesis was to investigate canonical intrinsic connectivity networks in language function and seizure semiology. Epilepsy surgery aims to remove brain areas that generate seizures. Language dysfunction is frequently observed after anterior temporal lobe resection (ATLR), and the presurgical workup seeks to identify the risks associated with surgical outcome. The principal aim of experimental studies was to elaborate understanding of language function as expressed in the recruitment of relevant connectivity networks and to evaluate whether it has value in the prediction of language decline after anterior temporal lobe resection. Using cognitive fMRI, we assessed brain areas defined by parameters of anatomy and canonical intrinsic connectivity networks (ICN) that are involved in language function, specifically word retrieval as expressed in naming and fluency. fMRI data was quantified by lateralisation indices and by ICN_atlas metrics in a priori defined ICN and anatomical regions of interest. Reliability of language ICN recruitment was studied in 59 patients and 30 healthy controls who were included in our language experiments. New and established language fMRI paradigms were employed on a three Tesla scanner, while intellectual ability, language performance and emotional status were established for all subjects with standard psychometric assessment. Patients who had surgery were reinvestigated at an early postoperative stage of four months after anterior temporal lobe resection. A major part of the work sought to elucidate the association between fMRI patterns and disease characteristics including features of anxiety and depression, and prediction of postoperative language outcome. We studied the efficiency of reorganisation of language function associated with disease features prior to and following surgery. A further aim of experimental work was to use EEG-fMRI data to investigate the relationship between canonical intrinsic connectivity networks and seizure semiology, potentially providing an avenue for characterising the seizure network in the presurgical workup. The association of clinical signs with the EEG-fMRI informed activation patterns were studied using the data from eighteen patients’ whose seizures and simultaneous EEG-fMRI activations were reported in a previous study. The accuracy of ICN_atlas was validated and the ICN construct upheld in the language maps of TLE patients. The ICN construct was not evident in ictal fMRI maps and simulated ICN_atlas data. Intrinsic connectivity network recruitment was stable between sessions in controls. Amodal linguistic processing and the relevance of temporal intrinsic connectivity networks for naming and that of frontal intrinsic connectivity networks for word retrieval in the context of fluency was evident in intrinsic connectivity networks regions. The relevance of intrinsic connectivity networks in the study of language was further reiterated by significant association between some disease features and language performance, and disease features and activation in intrinsic connectivity networks. However, the anterior temporal lobe (ATL) showed significantly greater activation compared to intrinsic connectivity networks – a result which indicated that ATL functional language networks are better studied in the context of the anatomically demarked ATL, rather than its functionally connected intrinsic connectivity networks. Activation in temporal lobe networks served as a predictor for naming and fluency impairment after ATLR and an increasing likelihood of significant decline with greater magnitude of left lateralisation. Impairment of awareness served as a significant classifying feature of clinical expression and was significantly associated with the inhibition of normal brain functions. Canonical intrinsic connectivity networks including the default mode network were recruited along an anterior-posterior anatomical axis and were not significantly associated with clinical signs.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item