EEG-correlated functional MRI in epilepsy

This thesis concerns the application of simultaneous EEG-correlated functional MRI (EEG/fMRI) to imaging epileptiform activity in vivo. The limitations of fMRI and scalp EEG are discussed together with the key advantages and limitations of combining the two. Findings are presented from work on over 100 patient volunteers with epilepsy, including the first continuous EEG/fMRI patient study, and results from imaging focal electrographic seizures and absence seizures (serving as human evidence for the reciprocal participation of focal thalamic and widespread cortical networks during generalized spike-wave activity). The core comprises a study of interictal epileptiform EEG activity (IEDs) in 63 patients with focal epilepsy. Semi-automated spike detection and advanced modeling strategies are introduced to account for different EEG event types, and to minimize false activations from uncontrolled motion. We show that: (1) significant hemodynamic correlates were detectable in over 68% of patients in whom discharges were captured and were highly, but not entirely, concordant with site(s) of presumed seizure generation where known (2) deactivations were less concordant and may non-specifically reflect the consequential or downstream effects of IEDs on brain activity (3) a striking pattern of retrosplenial deactivation was observed in 7 cases mainly with focal discharges (4) the general hemodynamic response to IEDs is physiological (5) incorporating information about different types of IEDs, their durations and saturation effects resulted in more powerful models for the detection of fMRI correlates (6) focal activations were more likely when there was good electroclinical localization, frequent stereotyped spikes, less head motion and less background EEG abnormality, but were also seen in patients in whom the electroclinical focus localization was uncertain. Data is also presented from patients with generalized epilepsy and from a study of primary reading epilepsy using simultaneous EMG and voice recordings, exploring the relationship between cognitive and epileptic activity