Portal de Periódicos da CAPES

Long‐term Reproducibility of f MRI Activation in Epilepsy Patients with Fixation Off Sensitivity

2005; Wiley; Volume: 46; Issue: 7 Linguagem: Inglês

10.1111/j.1528-1167.2005.11005.x

1528-1167

Carlo Di Bonaventura, Anna Elisabetta Vaudano, Marco Carnì, Patrizià Pantano, Valter Nucciarelli, Girolamo Garreffa, B. Maraviglia, Massimiliano Prencipe, L. Bozzao, M. Manfredi, Anna Teresa Giallonardo,

Epilepsy research and treatment

Recently EEG/functional magnetic resonance imaging (fMRI) has been proposed in epilepsy research to study and map electrophysiological activity. It can detect regions in the brain activated during interictal or ictal EEG abnormalities and thus contribute to a clearer understanding of neurophysiologic mechanisms underlying epileptic phenomena (1-3). However, fMRI studies are limited by a number of factors: some are related to the technique itself (nonstandardized technical approach, heterogeneous procedures used in the experimental paradigms, image processing and statistical analysis, and varying sensitivity of the instrument), whereas others are related to the seizures (unpredictable occurrence, technical problems, and artifact on images in case of motor activity). It may, consequently, be difficult to verify the reproducibility of results over time, as is suggested by the lack of studies focusing on this aspect. In March 2002, our group published a work on the hemodynamic correlates of epileptiform discharges elicited by the elimination of central vision and fixation (so-called Fixation Off Sensitivity, FOS) (4). In this article, paroxysmal activity in the temporooccipital regions elicited by closing the eyes correlated significantly with a highly localized increased blood oxygen level–dependent (BOLD) signal in extrastriate areas (Fig. 1A). This result was obtained in three epilepsy patients with a high within-subject and between-subject concordance during an EEG/fMRI session performed according to the classic block design, with rest consisting of an "eyes open" state, and activation of an "eyes-closed" state. A: Previously published statistical parametric maps illustrating cortical areas activated during the "eyes closed" state in the three patients with Fixation Off Sensitivity (FOS). B: Activated areas during the same state in EEG/fMRI study performed in the same patients, 3 years later, by using the same experimental setting and paradigm. Statistical parametric maps thresholded at p < 0.05 (corrected for multiple comparisons) are superimposed on a structural image. In cases 1 and 2, the "eyes closed" state elicited main activation clusters in the extrastriate areas in the temporooccipital cortex (bilateral in case 1, unilateral in case 2) with a high long-term within-subject reproducibility; in case 3, as the FOS phenomenon had previously disappeared after modification of the therapy, no fMRI activation was found. C: EEG tracings recorded in the three patients outside the scanner; note the disappearance of FOS phenomenon in case 3. D: EEG paroxysmal activity (2.5-Hz spike-and-wave complexes in bilateral temporoparietooccipital regions) elicited by elimination of central vision and fixation in case 1, recorded during fMRI session after off-line artifact subtraction; black arrows, eye closure. In the same patients, we repeated, 3 years later, an EEG/fMRI study using the same experimental setting and paradigm. From an epileptologic point of view, the patients' clinical condition was unchanged, with partial seizures persisting at a monthly frequency. The EEG showed an unchanged FOS phenomenon in two cases (cases 1 and 2 in the previous article), whereas in one patient (case 3 in the previous article) this pattern had disappeared 6 months previously after a modification in therapy. Before the fMRI session, both the patients still displaying the FOS phenomenon underwent video-EEG monitoring (Telefactor System, 21 channels, International 10–20 System, West Warwick, U.S.A.) to assess the appearance of the FOS phenomenon as well as its reactivity to specific stimuli (opened and closed eyes, darkness, etc.). fMRI data (20 axial slices, 5-mm thickness, TR/TE = 3,000/50 ms, image matrix 64 × 64) were acquired by using a clinical 1.5-T magnet (Philips Gyroscan, Eindhoven, Netherlands) continuously (two series of 200 temporal dynamics, 10-min scan time for each session). The functional study consisted of multiple 15-s epochs at baseline ("off," eyes open) and during activation ("on," eyes closed), in a boxcar configuration. The EEG was recorded throughout the fMRI scan by using an MR-compatible cap connected to a nonferrous shielded head-box. EEG signals (sampling rate, 1,024 Hz) were transmitted through a fiberoptic cable to a digital recording system (Micromed, Treviso, Italy). The gradient-induced artifact on the EEG data was digitally removed off-line by using custom-developed software (Micromed). The EEG tracings were reviewed to verify the constant presence of FOS phenomenon and the temporal coincidence of this phenomenon with the block-design paradigm we used. The fMRI data were preprocessed and analyzed by using SPM99 software (http://www.fil.ion.ucl.ac.uk/spm, Wellcome Department of Cognitive Neurology, London, U.K.). All the images were realigned to correct for the subject's head movements inside the scanner and smoothed (gaussian kernel, 8 mm) to increase the signal-to-noise ratio. The signal intensity of each voxel of the resulting time series of images was correlated with the expected hemodynamic response function for each discharge that occurred during acquisition. The resulting statistical parametric maps of the brain showing significantly correlated voxels were thresholded for amplitude (p < 0.05, corrected for multiple comparisons) of the activated brain areas. The results were overlaid onto the structural image for presentation. The images acquired were normalized, and anatomic labeling was performed by using the Talairach atlas. The analysis of the fMRI data showed a main activation cluster in the temporooccipital regions (bilateral in case 1, monolateral on the right side in case 2; Fig. 1B) related to epileptiform abnormalities in the closed-eyes condition (Fig. 1C and D); in case 3, in whom the FOS phenomenon had disappeared, no fMRI changes were observed. The fact that these features are highly concordant with previously published data (4) indicates good long-term reproducibility of this technique in the study of epileptiform activities. This result is noteworthy on account of the lack of EEG/fMRI studies testing the replications of within- and between-subject results over time. Pathologic activity findings obtained by means of EEG/fMRI are, indeed, very difficult to reproduce because of the marked intrinsic limitations of this technique when applied to epilepsy patients (e.g., the unpredictable occurrence of interictal or ictal abnormalities and the technical problems involved in recording seizures with motor activity); moreover, as EEG/fMRI is, by nature, an instrument of varying sensitivity, nonstandardized techniques or heterogeneous procedures are often used (5, 6). It is for all these reasons that data on the reproducibility of fMRI results in epilepsy studies are scarce, and that the results of the few studies that have focused on this topic are too limited for any conclusive considerations (few patients, no long-term repeated examination, studies limited to interictal spikes) (7-9). Given its monomorphic, constantly elicitable pattern, the FOS phenomenon has proved to be an ideal model with which to study the hemodynamic correlates of epileptiform activity and to replicate or verify results over the time.