- Email: [email protected]
Functional MRI interactions between dysplastic nodules and overlying cortex in periventricular nodular heterotopia
Epilepsy & Behavior 19 (2010) 631–634
Contents lists available at ScienceDirect
Epilepsy & Behavior j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / ye b e h
Brief Communication
Functional MRI interactions between dysplastic nodules and overlying cortex in periventricular nodular heterotopia John S. Archer a,b,c,⁎, David F. Abbott a,b, Richard A.J. Masterton b, Susan M. Palmer b, Graeme D. Jackson a,b,c a b c
University of Melbourne, Melbourne, Australia Brain Research Institute, Florey Neuroscience Institutes (Austin), Melbourne, Australia Austin Health, Melbourne, Australia
a r t i c l e
i n f o
Article history: Received 13 August 2010 Revised 10 September 2010 Accepted 10 September 2010 Available online 27 October 2010 Keywords: Epilepsy Electroencephalography–functional magnetic resonance imaging Functional connectivity Malformations of cortical development Cortical dysplasia
a b s t r a c t Periventricular nodular heterotopia (PVNH) is a malformation of cortical development associated with epilepsy. It is unclear whether the epileptogenic focus is the nodule, overlying cortex, or both. We performed electroencephalography (EEG)–functional magnetic resonance imaging (fMRI) in a patient with bilateral PVNH, capturing 45 “left temporal” epileptiform discharges. The relative time at which fMRI-involved regions became active was assessed. Additionally, nodule–cortex interactions were explored using fMRI functional connectivity. There was EEG–fMRI activity in specific periventricular nodules and overlying cortex in the left temporoparietal region. In both nodules and cortex, the peak BOLD response to epileptiform events occurred earlier than expected from standard fMRI hemodynamic modeling. Functional connectivity showed nodule– cortex interactions to be strong in this region, even when the influence of fMRI activity fluctuations due to spiking was removed. Nonepileptogenic, contralateral nodules did not show connectivity with overlying cortex. EEG–fMRI and functional connectivity can help identify which of the multiple abnormal regions are epileptogenic in PVNH. © 2010 Elsevier Inc. All rights reserved.
1. Introduction Periventricular nodular heterotopia (PVNH) is a malformation of cortical development characterized by persistence of cells in the ventricular zone during cortical development. Most patients with familial bilateral PVNH have mutations in the filamin-A gene, which assists locomotion of neurons along radial glial fibers after their initial proliferation in the periventricular zone. Many patients with bilateral PVNH have epilepsy, and there is likely to be a complex interaction between epileptogenic nodular tissue and overlying cortex. Most have an electroclinical pattern of temporal lobe epilepsy, but routine temporal lobe resections have had limited success [1]. Depth EEG recordings have variably found seizures to arise in nodular tissue, overlying cortex, or simultaneously in both [2,3]. There are some reports of seizure reduction in patients with focal PVNH following tailored wedge resections of heterotopic nodules and overlying brain [3]. The simultaneous acquisition of continuous EEG and functional MRI (EEG–fMRI) recordings allows whole-brain visualization of
⁎ Corresponding author. University of Melbourne, Neurosciences Building, Repatriation Campus, Austin Health, Banksia Street, Heidelberg West 3081, Victoria, Australia. Fax: +61 3 9496 4071. E-mail address: [email protected] (J.S. Archer). 1525-5050/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.yebeh.2010.09.018
changes in neural activity during interictal epileptiform events [4]. Functional connectivity analysis detects brain regions showing temporally correlated fluctuations in fMRI activity, to display networks of brain regions presumed to be involved in a common function [5]. This study had two aims: (1) using EEG–fMRI, to reveal the cerebral structures active during epileptiform activity in PVNH, and to explore their temporal association; (2) using functional connectivity of nodular tissue to assess the nature of nodule–cortex interactions and to determine whether it is possible to distinguish between “epileptogenic” and “nonepileptogenic” nodules.
2. Case report A 38-year-old, filamin-A-positive woman had a 14-year history of recurrent complex partial seizures. Structural MRI confirmed multiple, bilateral periventricular nodules of heterotopic gray matter (Fig. 1). Neuropsychometric testing confirmed mild intellectual disability and normal memory. Video/EEG monitoring revealed seizures arising from the left temporal region, characterized by psychoparesis, leftward eye deviation, and variable oral automatisms. Interictal EEG (Supplementary Fig. 1 [see Appendix]) showed frequent sharp waves, sharp–slow complexes, and runs of delta slowing broadly over the left temporal region, often with admixed focal fast activity over the left posterior
632
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
Fig. 1. Timing of regions involved during epileptiform activity in periventricular nodular heterotopia. In the left column are two T1-weighted axial MR images showing the nodules of abnormal gray matter studding the ventricular walls (L N R). The middle column shows the EEG–fMRI activations within key regions thresholded at P b 0.05 corrected for familywise error (for an activation map providing more complete slice coverage, see Supplementary Fig. 2). The right column shows associated time courses of 3.4 × 3.4-mm regions of interest identified from the activation map, displayed as a peristimulus time histogram with 3.2-second time bins, covering 30 seconds before and after the epileptiform event time (vertical green line). Activation (~ 1.5% signal change) is seen in periventricular nodules and overlying cortex.
temporal region (T5). Seizures were characterized by fast activity building out of T5, evolving into ictal theta over the left temporal region, occasionally extending over the left hemisphere. Flurodeoxyglucose (FDG) PET revealed left temporal pole hypometabolism, and co-registration with MRI confirmed metabolic activity in periventricular heterotopic tissue. Ictal SPECT showed regional hyperperfusion over the left lateral temporal lobe (Supplementary Fig. 1), compared with the interictal study.
3. Methods 3.1. fMRI acquisition Using an in-house EEG–fMRI system and previously published methods [6], we recorded EEG during 30 minutes of continuous fMRI imaging. The BOLD fMRI was acquired on a 3-T GE Signa LX scanner (General Electric, Milwaukee, WI, USA) with a birdcage quadrature head coil. Whole-brain T2⁎-weighted functional magnetic resonance images were acquired using a gradient-recalled echo-planar imaging (EPI) sequence (TR = 3.2 seconds, TE = 40 ms, flip angle = 80°, FOV= 24 × 24 cm, 64 × 64 matrix) with 40 slices giving a resolution of 3.4 × 3.4 × 3.2 (+0.2 gap) mm.
3.2. Electroencephalography–functional magnetic resonance imaging The EEG record was reviewed offline [7], identifying 45 epileptiform events, for an event-related fMRI analysis in SPM8 (http://www. fil.ion.ucl.ac.uk/spm/). Standard pre-processing steps were: slicetiming correction, rigid-body realignment to correct for subject motion, and spatial smoothing with a gaussian kernel (FWHM = 6 mm). The realignment parameters were incorporated in the analysis to model subject motion. Twenty-five analyses were performed, each based on hemodynamic response functions (HRFs) of different onset relative to the EEG event [8], ranging from –19.2 to + 19.2 seconds, in 1.6-second steps. Each analysis produced a statistical parametric map (SPM), thresholded at P b 0.05 corrected for familywise error (FWE). The SPM showing the strongest t statistic was used to identify key cortical and nodular regions for more detailed time course analysis. Peristimulus time histograms were plotted, showing the average BOLD signal within 3.2-second time bins (one TR) from 30 seconds before to 30 seconds after the event. 3.3. Functional connectivity Functional connectivity analysis was performed in SPM8 using pre-processed images from the EEG–fMRI dataset. Active and control
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
periventricular nodules were chosen as seeds for the connectivity analysis, based on results of the EEG–fMRI analysis. Time courses of average signal intensity within each seed region were generated in iBrain [9]. Each seed time course was compared with all other voxels in the brain using separate regression analyses in SPM8. After bandpass filtering (0.008–0.08 Hz) of the fMRI signal, nine regressors were included in the design matrix to reduce the effects of physiological noise: global signal change in the ventricles, gray + white matter; and focal signal change in the white matter and ventricles [5]. Before analysis, each seed time course was orthogonalized with respect to these regressors of no interest. Finally, we repeated the connectivity analysis with the additional step of including a regressor of no interest that modeled the expected event-related BOLD fluctuations. Functional connectivity images were displayed to show voxels significant to P b 0.05, corrected for familywise error. 4. Results EEG–fMRI showed BOLD activity changes time-locked to epileptiform activity in several regions of periventricular tissue and overlying cortex (Fig. 1, Supplementary Fig. 2). The most statistically
633
significant BOLD signal increases were observed in the large area of periventricular tissue in the left trigone and in overlying left posterior temporal and parietal cortex. BOLD increases extended into adjacent, more anterior periventricular tissue. Peak activation (highest-significance activation map) was seen on the analysis assuming event onset 1.6 seconds prior to epileptiform events (HRF peak 3.4 seconds postevent). Peristimulus time histograms (Fig. 1) showed BOLD signal rising around the same time in nodular and cortical regions, suggesting a rapid interplay between the two regions. Although BOLD signal appeared to begin rising earlier in nodular tissue than cortex, this difference did not reach statistical significance. Functional connectivity confirmed strong co-fluctuations in fMRI activity between “active” left periventricular tissue, other “active” nodular areas, and overlying cortex (Fig. 2). This persisted even when analysis removed the influence of scalp detected spiking (Fig. 2B). Comparable co-fluctuations in BOLD activity between periventricular nodules and overlying cortex were not evident for a “not involved” contralateral nodule. An ipsilateral “not involved” nodule in the frontal horn provided a more noisy image, with edge effects suggesting contamination by motion, but not the clear pattern of activation with overlying cortex.
Fig. 2. Functional connectivity of nodular tissue in periventricular nodular heterotopia. For each analysis the “seed” region is indicated by the green circle; warm colors indicate regions showing a positive correlation with the seed region, and cool colors represent negative correlations. The color scale represents statistical significance (t statistics). (A) A seed placed in nodular tissue in the left trigone, a region that showed EEG–fMRI involvement, reveals other key elements of the “epilepsy network.” (B) Even after activation related to epileptiform events has been regressed out, the left trigone nodule continues to show strong co-fluctuations with overlying cortex, suggesting activity in this “network” even when epileptiform activity is not visible at the scalp. (C) A contralateral nodule shows minimal functional connectivity with overlying cortex. (D) A left frontal nodule, not active during epileptiform discharges, shows minimal co-fluctuation with overlying cortex. Speckled activation in the L N R hemisphere overlies boundaries of brain ventricle and brain skull, suggesting contamination by residual motion artifact. Images are thresholded at P b 0.05, corrected for familywise error.
634
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
5. Discussion Epileptiform activity in this subject with bilateral PVNH is associated with fMRI activity changes in specific periventicular nodules and overlying cortex. The interaction between cortical and deep structures is complex, and, within the temporal sensitivity of fMRI, nodular regions are involved around the same time as cortex. However, fMRI activity builds in this network earlier than expected, suggesting there may be rising neuronal activity prior to scalpdetected spikes. We and others have observed a similar phenomenon with EEG–fMRI studies in focal [10] and generalized [11] epilepsies. It is unclear whether this reflects building nonepileptic activity in the nodule–cortex “network’” or building epileptiform activity not detectable at the scalp. A recent EEG–fMRI study of four patients with bilateral PVNH showed nodule activation in half, during epileptiform events and seizures [12]. Because the amplitude and significance of activation were greater in cortex than in nodular tissue, the authors concluded that the seizures originated from overlying cortex. The study modeled seizures as a block of activation, an approach that might fail to detect nodular regions involved in initiating, but not sustaining a seizure. Plots of event-related fMRI signal change reveal the time course of involved regions. Many elements of the “epileptogenic network” were revealed by a functional connectivity analysis seeding one nodular component of the network. Invasive neurophysiological studies have shown similar functional integration across this “epilepsy network” in PVNH [2]. Functional connectivity with overlying cortex was not present in a similar contralateral nodule not involved in epileptic discharges. Intuitively, nodules require cortical interactions to generate seizures, but it is unclear whether all nodules with cortical interactions are epileptogenic. Synchronized co-fluctuations in fMRI activity were present in the “epileptogenic network,” even when the influence of epileptiform-related activity was removed. Again, it is unclear whether this reflects nonepileptic activity in the nodule–cortex “network” or epileptiform activity not detectable through the filtering effects of the scalp. This patient proceeded to craniotomy, resection of left trigone nodular tissue including a wedge of overlying posterior temporal cortex, and clearance of adjacent ventricular nodules. There had been one seizure at 9-month review, compared with weekly clusters previously. 5.1. Ethical approval We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this article is consistent with those guidelines.
5.2. Conflict of interest statement None of the authors has any conflict of interest to disclose.
Acknowledgments This study was supported by the National Health and Medical Research Council of Australia (NHMRC Project Grant 628725) and the Operational Infrastructure Support Program of the State Government of Victoria. We gratefully acknowledge the assistance of Mr. Danny Flanagan for performing and reviewing the EEG record.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.yebeh.2010.09.018.
References [1] Sisodiya SM. Surgery for malformations of cortical development causing epilepsy. Brain 2000;123(Pt 6):1075–91. [2] Valton L, Guye M, McGonigal A, et al. Functional interactions in brain networks underlying epileptic seizures in bilateral diffuse periventricular heterotopia. Clin Neurophysiol 2008;119:212–23. [3] Tassi L, Colombo N, Cossu M, et al. Electroclinical, MRI and neuropathological study of 10 patients with nodular heterotopia, with surgical outcomes. Brain 2005;128: 321–37. [4] Waites AB, Shaw ME, Briellmann RS, Labate A, Abbott DF, Jackson GD. How reliable are fMRI–EEG studies of epilepsy? A nonparametric approach to analysis validation and optimization. NeuroImage 2005;24:192–9. [5] Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 2007;8:700–11. [6] Masterton RA, Abbott DF, Fleming SW, Jackson GD. Measurement and reduction of motion and ballistocardiogram artefacts from simultaneous EEG and fMRI recordings. NeuroImage 2007;37:202–11. [7] Flanagan D, Abbott DF, Jackson GD. How wrong can we be? The effect of inaccurate mark-up of EEG/fMRI studies in epilepsy. Clin Neurophysiol 2009;120:1637–47. [8] Hawco CS, Bagshaw AP, Lu Y, Dubeau F, Gotman J. BOLD changes occur prior to epileptic spikes seen on scalp EEG. NeuroImage 2007;35:1450–8. [9] Abbott D, Jackson G. iBrain: software for analysis and visualisation of functional MR images. NeuroImage 2001;13:s59. [10] Masterton RA, Harvey AS, Archer JS, et al. Focal epileptiform spikes do not show a canonical BOLD response in patients with benign rolandic epilepsy (BECTS). NeuroImage 2010;51:252–60. [11] Carney PW, Masterton RA, Harvey AS, Scheffer IE, Berkovic SF, Jackson GD. The core network in absence epilepsy: differences in cortical and thalamic BOLD response. Neurology 2010;75:904–11. [12] Tyvaert L, Hawco C, Kobayashi E, LeVan P, Dubeau F, Gotman J. Different structures involved during ictal and interictal epileptic activity in malformations of cortical development: an EEG–fMRI study. Brain 2008;131:2042–60.
Contents lists available at ScienceDirect
Epilepsy & Behavior j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / ye b e h
Brief Communication
Functional MRI interactions between dysplastic nodules and overlying cortex in periventricular nodular heterotopia John S. Archer a,b,c,⁎, David F. Abbott a,b, Richard A.J. Masterton b, Susan M. Palmer b, Graeme D. Jackson a,b,c a b c
University of Melbourne, Melbourne, Australia Brain Research Institute, Florey Neuroscience Institutes (Austin), Melbourne, Australia Austin Health, Melbourne, Australia
a r t i c l e
i n f o
Article history: Received 13 August 2010 Revised 10 September 2010 Accepted 10 September 2010 Available online 27 October 2010 Keywords: Epilepsy Electroencephalography–functional magnetic resonance imaging Functional connectivity Malformations of cortical development Cortical dysplasia
a b s t r a c t Periventricular nodular heterotopia (PVNH) is a malformation of cortical development associated with epilepsy. It is unclear whether the epileptogenic focus is the nodule, overlying cortex, or both. We performed electroencephalography (EEG)–functional magnetic resonance imaging (fMRI) in a patient with bilateral PVNH, capturing 45 “left temporal” epileptiform discharges. The relative time at which fMRI-involved regions became active was assessed. Additionally, nodule–cortex interactions were explored using fMRI functional connectivity. There was EEG–fMRI activity in specific periventricular nodules and overlying cortex in the left temporoparietal region. In both nodules and cortex, the peak BOLD response to epileptiform events occurred earlier than expected from standard fMRI hemodynamic modeling. Functional connectivity showed nodule– cortex interactions to be strong in this region, even when the influence of fMRI activity fluctuations due to spiking was removed. Nonepileptogenic, contralateral nodules did not show connectivity with overlying cortex. EEG–fMRI and functional connectivity can help identify which of the multiple abnormal regions are epileptogenic in PVNH. © 2010 Elsevier Inc. All rights reserved.
1. Introduction Periventricular nodular heterotopia (PVNH) is a malformation of cortical development characterized by persistence of cells in the ventricular zone during cortical development. Most patients with familial bilateral PVNH have mutations in the filamin-A gene, which assists locomotion of neurons along radial glial fibers after their initial proliferation in the periventricular zone. Many patients with bilateral PVNH have epilepsy, and there is likely to be a complex interaction between epileptogenic nodular tissue and overlying cortex. Most have an electroclinical pattern of temporal lobe epilepsy, but routine temporal lobe resections have had limited success [1]. Depth EEG recordings have variably found seizures to arise in nodular tissue, overlying cortex, or simultaneously in both [2,3]. There are some reports of seizure reduction in patients with focal PVNH following tailored wedge resections of heterotopic nodules and overlying brain [3]. The simultaneous acquisition of continuous EEG and functional MRI (EEG–fMRI) recordings allows whole-brain visualization of
⁎ Corresponding author. University of Melbourne, Neurosciences Building, Repatriation Campus, Austin Health, Banksia Street, Heidelberg West 3081, Victoria, Australia. Fax: +61 3 9496 4071. E-mail address: [email protected] (J.S. Archer). 1525-5050/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.yebeh.2010.09.018
changes in neural activity during interictal epileptiform events [4]. Functional connectivity analysis detects brain regions showing temporally correlated fluctuations in fMRI activity, to display networks of brain regions presumed to be involved in a common function [5]. This study had two aims: (1) using EEG–fMRI, to reveal the cerebral structures active during epileptiform activity in PVNH, and to explore their temporal association; (2) using functional connectivity of nodular tissue to assess the nature of nodule–cortex interactions and to determine whether it is possible to distinguish between “epileptogenic” and “nonepileptogenic” nodules.
2. Case report A 38-year-old, filamin-A-positive woman had a 14-year history of recurrent complex partial seizures. Structural MRI confirmed multiple, bilateral periventricular nodules of heterotopic gray matter (Fig. 1). Neuropsychometric testing confirmed mild intellectual disability and normal memory. Video/EEG monitoring revealed seizures arising from the left temporal region, characterized by psychoparesis, leftward eye deviation, and variable oral automatisms. Interictal EEG (Supplementary Fig. 1 [see Appendix]) showed frequent sharp waves, sharp–slow complexes, and runs of delta slowing broadly over the left temporal region, often with admixed focal fast activity over the left posterior
632
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
Fig. 1. Timing of regions involved during epileptiform activity in periventricular nodular heterotopia. In the left column are two T1-weighted axial MR images showing the nodules of abnormal gray matter studding the ventricular walls (L N R). The middle column shows the EEG–fMRI activations within key regions thresholded at P b 0.05 corrected for familywise error (for an activation map providing more complete slice coverage, see Supplementary Fig. 2). The right column shows associated time courses of 3.4 × 3.4-mm regions of interest identified from the activation map, displayed as a peristimulus time histogram with 3.2-second time bins, covering 30 seconds before and after the epileptiform event time (vertical green line). Activation (~ 1.5% signal change) is seen in periventricular nodules and overlying cortex.
temporal region (T5). Seizures were characterized by fast activity building out of T5, evolving into ictal theta over the left temporal region, occasionally extending over the left hemisphere. Flurodeoxyglucose (FDG) PET revealed left temporal pole hypometabolism, and co-registration with MRI confirmed metabolic activity in periventricular heterotopic tissue. Ictal SPECT showed regional hyperperfusion over the left lateral temporal lobe (Supplementary Fig. 1), compared with the interictal study.
3. Methods 3.1. fMRI acquisition Using an in-house EEG–fMRI system and previously published methods [6], we recorded EEG during 30 minutes of continuous fMRI imaging. The BOLD fMRI was acquired on a 3-T GE Signa LX scanner (General Electric, Milwaukee, WI, USA) with a birdcage quadrature head coil. Whole-brain T2⁎-weighted functional magnetic resonance images were acquired using a gradient-recalled echo-planar imaging (EPI) sequence (TR = 3.2 seconds, TE = 40 ms, flip angle = 80°, FOV= 24 × 24 cm, 64 × 64 matrix) with 40 slices giving a resolution of 3.4 × 3.4 × 3.2 (+0.2 gap) mm.
3.2. Electroencephalography–functional magnetic resonance imaging The EEG record was reviewed offline [7], identifying 45 epileptiform events, for an event-related fMRI analysis in SPM8 (http://www. fil.ion.ucl.ac.uk/spm/). Standard pre-processing steps were: slicetiming correction, rigid-body realignment to correct for subject motion, and spatial smoothing with a gaussian kernel (FWHM = 6 mm). The realignment parameters were incorporated in the analysis to model subject motion. Twenty-five analyses were performed, each based on hemodynamic response functions (HRFs) of different onset relative to the EEG event [8], ranging from –19.2 to + 19.2 seconds, in 1.6-second steps. Each analysis produced a statistical parametric map (SPM), thresholded at P b 0.05 corrected for familywise error (FWE). The SPM showing the strongest t statistic was used to identify key cortical and nodular regions for more detailed time course analysis. Peristimulus time histograms were plotted, showing the average BOLD signal within 3.2-second time bins (one TR) from 30 seconds before to 30 seconds after the event. 3.3. Functional connectivity Functional connectivity analysis was performed in SPM8 using pre-processed images from the EEG–fMRI dataset. Active and control
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
periventricular nodules were chosen as seeds for the connectivity analysis, based on results of the EEG–fMRI analysis. Time courses of average signal intensity within each seed region were generated in iBrain [9]. Each seed time course was compared with all other voxels in the brain using separate regression analyses in SPM8. After bandpass filtering (0.008–0.08 Hz) of the fMRI signal, nine regressors were included in the design matrix to reduce the effects of physiological noise: global signal change in the ventricles, gray + white matter; and focal signal change in the white matter and ventricles [5]. Before analysis, each seed time course was orthogonalized with respect to these regressors of no interest. Finally, we repeated the connectivity analysis with the additional step of including a regressor of no interest that modeled the expected event-related BOLD fluctuations. Functional connectivity images were displayed to show voxels significant to P b 0.05, corrected for familywise error. 4. Results EEG–fMRI showed BOLD activity changes time-locked to epileptiform activity in several regions of periventricular tissue and overlying cortex (Fig. 1, Supplementary Fig. 2). The most statistically
633
significant BOLD signal increases were observed in the large area of periventricular tissue in the left trigone and in overlying left posterior temporal and parietal cortex. BOLD increases extended into adjacent, more anterior periventricular tissue. Peak activation (highest-significance activation map) was seen on the analysis assuming event onset 1.6 seconds prior to epileptiform events (HRF peak 3.4 seconds postevent). Peristimulus time histograms (Fig. 1) showed BOLD signal rising around the same time in nodular and cortical regions, suggesting a rapid interplay between the two regions. Although BOLD signal appeared to begin rising earlier in nodular tissue than cortex, this difference did not reach statistical significance. Functional connectivity confirmed strong co-fluctuations in fMRI activity between “active” left periventricular tissue, other “active” nodular areas, and overlying cortex (Fig. 2). This persisted even when analysis removed the influence of scalp detected spiking (Fig. 2B). Comparable co-fluctuations in BOLD activity between periventricular nodules and overlying cortex were not evident for a “not involved” contralateral nodule. An ipsilateral “not involved” nodule in the frontal horn provided a more noisy image, with edge effects suggesting contamination by motion, but not the clear pattern of activation with overlying cortex.
Fig. 2. Functional connectivity of nodular tissue in periventricular nodular heterotopia. For each analysis the “seed” region is indicated by the green circle; warm colors indicate regions showing a positive correlation with the seed region, and cool colors represent negative correlations. The color scale represents statistical significance (t statistics). (A) A seed placed in nodular tissue in the left trigone, a region that showed EEG–fMRI involvement, reveals other key elements of the “epilepsy network.” (B) Even after activation related to epileptiform events has been regressed out, the left trigone nodule continues to show strong co-fluctuations with overlying cortex, suggesting activity in this “network” even when epileptiform activity is not visible at the scalp. (C) A contralateral nodule shows minimal functional connectivity with overlying cortex. (D) A left frontal nodule, not active during epileptiform discharges, shows minimal co-fluctuation with overlying cortex. Speckled activation in the L N R hemisphere overlies boundaries of brain ventricle and brain skull, suggesting contamination by residual motion artifact. Images are thresholded at P b 0.05, corrected for familywise error.
634
J.S. Archer et al. / Epilepsy & Behavior 19 (2010) 631–634
5. Discussion Epileptiform activity in this subject with bilateral PVNH is associated with fMRI activity changes in specific periventicular nodules and overlying cortex. The interaction between cortical and deep structures is complex, and, within the temporal sensitivity of fMRI, nodular regions are involved around the same time as cortex. However, fMRI activity builds in this network earlier than expected, suggesting there may be rising neuronal activity prior to scalpdetected spikes. We and others have observed a similar phenomenon with EEG–fMRI studies in focal [10] and generalized [11] epilepsies. It is unclear whether this reflects building nonepileptic activity in the nodule–cortex “network’” or building epileptiform activity not detectable at the scalp. A recent EEG–fMRI study of four patients with bilateral PVNH showed nodule activation in half, during epileptiform events and seizures [12]. Because the amplitude and significance of activation were greater in cortex than in nodular tissue, the authors concluded that the seizures originated from overlying cortex. The study modeled seizures as a block of activation, an approach that might fail to detect nodular regions involved in initiating, but not sustaining a seizure. Plots of event-related fMRI signal change reveal the time course of involved regions. Many elements of the “epileptogenic network” were revealed by a functional connectivity analysis seeding one nodular component of the network. Invasive neurophysiological studies have shown similar functional integration across this “epilepsy network” in PVNH [2]. Functional connectivity with overlying cortex was not present in a similar contralateral nodule not involved in epileptic discharges. Intuitively, nodules require cortical interactions to generate seizures, but it is unclear whether all nodules with cortical interactions are epileptogenic. Synchronized co-fluctuations in fMRI activity were present in the “epileptogenic network,” even when the influence of epileptiform-related activity was removed. Again, it is unclear whether this reflects nonepileptic activity in the nodule–cortex “network” or epileptiform activity not detectable through the filtering effects of the scalp. This patient proceeded to craniotomy, resection of left trigone nodular tissue including a wedge of overlying posterior temporal cortex, and clearance of adjacent ventricular nodules. There had been one seizure at 9-month review, compared with weekly clusters previously. 5.1. Ethical approval We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this article is consistent with those guidelines.
5.2. Conflict of interest statement None of the authors has any conflict of interest to disclose.
Acknowledgments This study was supported by the National Health and Medical Research Council of Australia (NHMRC Project Grant 628725) and the Operational Infrastructure Support Program of the State Government of Victoria. We gratefully acknowledge the assistance of Mr. Danny Flanagan for performing and reviewing the EEG record.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.yebeh.2010.09.018.
References [1] Sisodiya SM. Surgery for malformations of cortical development causing epilepsy. Brain 2000;123(Pt 6):1075–91. [2] Valton L, Guye M, McGonigal A, et al. Functional interactions in brain networks underlying epileptic seizures in bilateral diffuse periventricular heterotopia. Clin Neurophysiol 2008;119:212–23. [3] Tassi L, Colombo N, Cossu M, et al. Electroclinical, MRI and neuropathological study of 10 patients with nodular heterotopia, with surgical outcomes. Brain 2005;128: 321–37. [4] Waites AB, Shaw ME, Briellmann RS, Labate A, Abbott DF, Jackson GD. How reliable are fMRI–EEG studies of epilepsy? A nonparametric approach to analysis validation and optimization. NeuroImage 2005;24:192–9. [5] Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 2007;8:700–11. [6] Masterton RA, Abbott DF, Fleming SW, Jackson GD. Measurement and reduction of motion and ballistocardiogram artefacts from simultaneous EEG and fMRI recordings. NeuroImage 2007;37:202–11. [7] Flanagan D, Abbott DF, Jackson GD. How wrong can we be? The effect of inaccurate mark-up of EEG/fMRI studies in epilepsy. Clin Neurophysiol 2009;120:1637–47. [8] Hawco CS, Bagshaw AP, Lu Y, Dubeau F, Gotman J. BOLD changes occur prior to epileptic spikes seen on scalp EEG. NeuroImage 2007;35:1450–8. [9] Abbott D, Jackson G. iBrain: software for analysis and visualisation of functional MR images. NeuroImage 2001;13:s59. [10] Masterton RA, Harvey AS, Archer JS, et al. Focal epileptiform spikes do not show a canonical BOLD response in patients with benign rolandic epilepsy (BECTS). NeuroImage 2010;51:252–60. [11] Carney PW, Masterton RA, Harvey AS, Scheffer IE, Berkovic SF, Jackson GD. The core network in absence epilepsy: differences in cortical and thalamic BOLD response. Neurology 2010;75:904–11. [12] Tyvaert L, Hawco C, Kobayashi E, LeVan P, Dubeau F, Gotman J. Different structures involved during ictal and interictal epileptic activity in malformations of cortical development: an EEG–fMRI study. Brain 2008;131:2042–60.