Page 218 - ILAE_Lectures_2015

Voxel-based morphometry may demonstrate areas of hippocampal atrophy in individual
patients with clear-cut hippocampal sclerosis49 but for the detection of occult abnormalities
in individual patients it appears to be relatively insensitive at thresholds that do not give false
positive results50. These methods are complemented by anatomical atlases, which allow
quantification of brain imaging data on a lobar and sub-lobar basis51,52. Analysis of the texture
of the neocortex on a T1-weighted volume scan may give increased sensitivity to identify
focal cortical dyplasia27. Curvilinear reconstructions may increase the visibility of subtle
neocortical lesions53. Three-dimensional reconstruction of the neocortex may assist
visualisation of abnormalities and surgical planning54, and this can now be automated55.

FUNCTIONAL MRI

Ictal and inter-ictal epileptiform activity

Activation of eloquent cortex has been shown in patients with frequent partial seizures and
in patients undergoing epilepsy surgery who had invasive EEG evaluation, the ictal-onset
phase-related maps were concordant with the presumed seizure onset zone for all patients.
The most statistically significant haemodynamic cluster within the presumed seizure onset
zone was between 1.1 and 3.5 cm from the invasively defined seizure onset zone. Resection
of this region was associated with a good surgical outcome185.

Focal increases in cerebral blood delivery have been identified in patients with frequent inter-
ictal spikes56-58. Continuous recording of EEG and functional MRI (fMRI) is possible,
following introduction of methods to remove the artifact on the EEG trace caused by the
fMRI acquisition, and results in much more detail and analysis of the time course of
haemodynamic changes59,60. However, even in well-selected patients, approximately 50% do
not exhibit inter-ictal epileptiform discharges during the 1060 minute EEG/fMRI study. Of
the remaining patients, approximately 50–90% lead to significant signal changes which are
concordant with electroclinical data61,62,186. These results may be used to re-evaluate patients
who have, for example, been previously rejected for epilepsy surgery63.

The most recent and largest of these studies reported 76 patients with refractory focal
epilepsy, 33 of whom had extratemporal lobe epilepsy. No discharges were seen during the
3560 minute acquisition in 64% of the patients. The mean number of discharges during the
recording in the remaining patients was 89.3. Ten patients underwent surgical treatment,
seven of whom achieved seizure freedom. In six of these seven patients, the area of resection
included the area of maximum BOLD (blood-oxygen-level dependent) signal increase.
Interestingly, in the remaining three patients who continued to have seizures after surgery,
the areas of maximal BOLD activation did not overlap with the resected area suggesting that
EEG/fMRI may possess a negative predictive value for a good surgical outcome, that is, a
lack of concordance with other localising investigations may predict a poor postoperative
outcome64.

The interrogation of the EEG/fMRI data for typical spike-related haemodynamic responses
in patients without spikes during the recording period may also provide useful localising
information and improve the yield of EEG/fMRI studies187.

The feasibility of combining fMRI with intracranial EEGs is being studied65. As EEG-fMRI
covers the whole brain, and as intracranial EEG has very limited spatial sampling, the
BOLD changes may assist in the interpretation of intracranial EEG data. This is particularly
relevant if the irritative and seizure onset zones are not being sampled directly by the EEG
contacts, but these are detecting the propagation of epileptic activity, as the fMRI data can
be examined for BOLD changes prior to the first detected EEG changes.