Functional neuro-imaging as a pre-surgical tool in epilepsy

Abstract

Functional neuro-imaging techniques are helpful in the pre-surgical evaluation of epilepsy for localization of the epileptogenic zone as ancillary tools to electroencephalography (EEG) and magnetic resonance imaging (MRI) or when other localization techniques are normal, non-concordant or discordant. Positron emission tomography (PET) and ictal single photon emission computed tomography (ictal SPECT) imaging are traditional tests that have been reported to have good sensitivity and specificity although the results are better with more expertise as is true for any technique. More recently magnetoencephalogram/magnetic source imaging (MEG/MSI), diffusion tensor imaging and functional magnetic resonance imaging (fMRI) have been used in localization and functional mapping during the pre-surgical work-up of epilepsy. Newer techniques such as fMRI-EEG, functional connectivity magnetic resonance imaging and near infra-red spectroscopy, magnetic resonance spectroscopy and magneto nanoparticles hold promise for further development that could then be applied in the work-up of epilepsy surgery. In this manuscript, we review these techniques and their current position in the pre-surgical evaluation of epilepsy.

Keywords: Diffusion tensor imaging; functional connectivity magnetic resonance imaging; functional magnetic resonance imaging; functional magnetic resonance imaging-electroencephalography; magnetoencephalogram; positron emission tomography; single photon emission computed tomography; temporal lobe epilepsy.

Figure 1

Multi-modality imaging in a patient with left occipital lobe seizures. Magnetic resonance imaging (MRI) shows left sub-occipital cortical dysplasia in T2 and fluid attenuated inversion recovery sequences (a, b). Positron emission tomography (PET) imaging shows hypometabolism in corresponding areas visible as a lighter shade of gray in grayscale images (c, d), but is much easier appreciated as a break in the pink cortical ribbon in color coded PET imaging co-registered to MRI (e, f). Prior to surgery the patient also had diffusion tensor imaging tractography of the visual pathway fibers which are seen to be displaced superiorly by the lesion (yellow arrows in g, h)

Figure 2

Single photon emission computed tomography (SPECT) imaging and subtraction ictal SPECT co-registered to MRI (SISCOM): Patient with non-lesional left temporal lobe seizures. SPECT images show left temporal hypermetabolism in ictal images (a), compared to inter-ictal (b). Co-registration of these images to MRI makes the asymmetry more obvious in ictal (c) and inter-ictal (d) reconstructions (note that the slice angle and thickness in the reconstructions c and d are different from the raw SPECT images in a and b. Subtraction of the inter-ictal from the ictal SPECT shows the regions of maximal difference in intensity (e). Image orientation is radiological (patient left is on image right)

Figure 3

MEG acquired with a 306-channel whole-head MEG system showing epileptiform spike activity mainly from the posterior-lateral aspect of the left inferior-frontal/superior-temporal lobes (red circle and green arrow in a). MSI shows that the MEG spikes (red circles in b) are localized to the inferior margin of a small cavity in the posterior part of the left inferior frontal gyrus (red arrow in c). The primary auditory cortex in the left hemisphere was located in the vicinity of the spike cluster (blue box in b)

Figure 4

Schematic showing analysis methodology in functional magnetic resonance imaging-electroencephalography (fMRI-EEG). EEG is acquired using a specialized system within the magnetic resonance imaging (MRI) machine while acquiring blood oxygenation level dependent (BOLD) sequences (a). Subsequently the EEG is analyzed for spikes and the corresponding BOLD fMRI change is detected (b). Multiple spike related BOLD signal changes are summated to improve the signal to noise ratio (c). The resultant summated signal is co-registered to the structural brain MRI to show the location of the summated BOLD signal change (d). This is the same patient in Figure 1. See how the summated signal in d corresponds to the lesion visualized with other structural and functional imaging modalities