FULL-LENGTH ORIGINAL RESEARCH
Altered thalamocortical functional connectivity in idiopathic generalized epilepsy *Jung Bin Kim, †Sang-il Suh, *Woo-Keun Seo, *Kyungmi Oh, *Seong-Beom Koh, and *Ji Hyun Kim Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
SUMMARY
Dr. Ji Hyun Kim is an associate professor in Neurology at Korea University Guro Hospital, Seoul, Republic of Korea.
Objective: Aberrant thalamocortical network has been hypothesized to play a crucial role in the fundamental pathogenesis underlying idiopathic generalized epilepsy (IGE). We aimed to investigate alterations of thalamocortical functional network in patients with IGE using thalamic seed-based functional connectivity (FC) analysis, and their relationships with frontal cognitive functions and clinical characteristics. Methods: Forty-nine IGE patients (31 with juvenile myoclonic epilepsy, 17 with IGE with generalized tonic–clonic seizures only, one with juvenile absence epilepsy) and 42 control subjects were prospectively recruited. Voxel-based morphometry (VBM) was first performed to detect thalamic region of gray matter (GM) reduction in patients compared to controls. Between-group comparison of thalamocortical FC was then carried out using resting-state functional magnetic resonance imaging (MRI) analysis seeding at thalamic region of volume difference. In addition, thalamocortical FC was correlated with frontal cognitive performance and clinical variables. Results: Neuropsychological assessment revealed that patients with IGE had poorer performance than controls on most of the frontal cognitive functions. VBM detected a reduction in GM in the anteromedial thalamus in patients relative to controls. FC analysis seeding at the anteromedial thalamus revealed a reduction of thalamocortical FC in the bilateral medial prefrontal cortex and precuneus/posterior cingulate cortex in patients with IGE compared to controls. Thalamocortical FC strength of bilateral medial prefrontal cortex correlated negatively with disease duration, but did not correlate with seizure frequency or frontal cognitive functions in patients with IGE. Significance: Our results indicate that IGE is associated with decreased thalamocortical FC between anteromedial thalamus and medial prefrontal cortex and precuneus/ posterior cingulate cortex. Our finding of greater reduction of medial prefrontal FC in relation to increasing disease duration suggests that thalamoprefrontal network abnormality, the proposed pathophysiologic mechanism underlying IGE, may be the consequence of the long-standing burden of the disease. KEY WORDS: Idiopathic generalized epilepsy, Resting-state functional connectivity, Thalamocortical network, Frontal cognitive dysfunction.
Idiopathic generalized epilepsy (IGE) constitutes a heterogeneous group of epilepsy syndromes with a nonfocal mechanism of seizure onset and no identifiable cause
other than a genetic predisposition. Childhood absence epilepsy (CAE), juvenile absence epilepsy, juvenile myoclonic epilepsy (JME), and IGE with generalized tonic–clonic seizures only (GTCS) are the well-recognized subsyndromes of IGE, according to predominant seizure semiology and age of seizure onset. Based on genetic traits, similar seizure semiology, and typical electroencephalography (EEG) features of 3- to 5-Hz generalized spike-wave discharges (GSWDs), these IGE subsyndromes are considered to share a common pathogenetic mechanism.
Accepted January 27, 2014. Departments of *Neurology and † Radiology, Korea University Guro Hospital, Korea University College of Medicine, Seoul, Korea Address correspondence to Ji Hyun Kim, Department of Neurology, Korea University Guro Hospital, Korea University College of Medicine, 152-703, Guro-dong Ro 148, Guro-gu, Seoul, Republic of Korea. E-mail: [email protected] Wiley Periodicals, Inc. © 2014 International League Against Epilepsy
1
2 J. B. Kim et al. The fundamental pathogenesis that underlies IGE is not fully elucidated; however, converging evidence has suggested a critical role of abnormal thalamocortical circuit in the generation of GSWDs. Recent advances in computational analysis of multimodal neuroimaging, such as voxel-based morphometry (VBM), magnetic resonance spectroscopy (MRS), diffusion tensor imaging (DTI), and functional magnetic resonance imaging (fMRI) have contributed greatly to the understanding of structural and functional changes of the thalamus and frontal cortex in IGE.1,2 In particular, contribution of the thalamocortical network to the generation of GSWDs has been highlighted in simultaneous EEG-fMRI studies that showed both thalamic activation and widespread cortical deactivation in IGE.3,4 Recently, resting-state functional connectivity MRI (fcMRI) has been used increasingly to investigate largescale functional networks at a whole-brain level, based on the temporal correlation of spontaneous fluctuations of blood oxygen level–dependent (BOLD) signals in a very low frequency range (0.01–0.1 Hz).5,6 Because fcMRI reliably identifies a variety of intrinsic corticocortical and corticosubcortical networks,5 it is widely used to noninvasively evaluate changes in functional brain networks in diverse neurologic and psychiatric disorders as well as in the normal aging process.6 Previous fcMRI studies have consistently shown decreased functional connectivity (FC) within the default mode network (DMN) in IGE, suggesting a disruption of functional integration of DMN in association with impaired consciousness and cognitive dysfunctions.7–11 Given the hypothesis of thalamocortical network abnormality in the fundamental pathomechanism underlying IGE, investigation of the thalamocortical FC seems more likely to provide valuable information regarding functional changes of this network. The purpose of the current study was to investigate FC alterations within the thalamocortical network in IGE, by using the thalamic seed derived from the VBM of group difference between patients with IGE and controls. We predicted that FC between thalamus and frontal cortex may be altered in patients relative to controls. Correlation analyses were additionally performed between thalamocortical FC and neuropsychological measures and clinical variables to explore the possible influence of the clinical and cognitive factors on FC changes.
Methods Subjects We prospectively recruited 54 right-handed adult patients with IGE (33 patients with JME, 20 patients with GTCS, one patient with juvenile absence epilepsy), who were followed at least 1 year in the epilepsy clinic of Korea University Guro Hospital. This study was an extension of our previous works.12,13 Diagnosis of IGE was based on Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
electroclinical criteria according to the International League Against Epilepsy (ILAE) classification. Inclusion criteria we used were as follows: (1) unequivocal seizure semiology of IGE—typical absence seizure, myoclonic seizure involving the bilateral upper extremities preferentially occurring early in the morning, or GTCS; (2) no evidence of developmental and neurologic abnormalities, and global cognitive impairment on Mini-Mental State Examination (MMSE score ≥ 28/30); (3) at least one EEG examination demonstrating typical GSWDs on a normal background; and (4) neither abnormal nor unusual findings on clinical MRI. Patients with comorbid neurologic, psychiatric, or chronic systemic disorders were excluded. None of the patients were taking medications other than antiepileptic drugs (AEDs) at the time of study inclusion. Demographic data and clinical information such as seizure semiology, age of seizure onset, duration of epilepsy, seizure frequency, and current AEDs were obtained through interviews with the patients and their parents and reviews of medical records. For group comparison, 45 right-handed healthy volunteers matched for age and gender were recruited to serve as controls. All control subjects underwent neurologic examination as well as a detailed interview to ensure that they had (1) no neurologic abnormality and global cognitive impairment (MMSE score ≥ 28/30); (2) no history of neurologic, psychiatric, or systemic disorders; (3) no family history of epilepsy; and (4) no history of alcohol or drug abuse. Control subjects with abnormal or unusual MRI findings were also excluded. The local ethics committee approved the study protocol, and all participants provided written informed consent prior to study inclusion. Neuropsychological assessment Neuropsychological assessment was carried out by an experienced neuropsychologist, who was blinded to the clinical diagnosis, on the same day of MRI scanning. Because IGE patients are known to have frontal cognitive dysfunctions,14,15 the neuropsychological battery was more weighted on the frontal cognitive functions. Assessed domains and the tests were as follows: (1) global cognitive function—Korean version of MMSE; (2) attention and working memory—Trail-Making Test part A, Digit Span Test (forward and backward); and (3) executive function – Trail-Making Test part B, Stroop Test (word, color, wordcolor), and Letter Fluency Test (words beginning with the three Korean letters). Group comparisons of demographic data and neuropsychological measures were undertaken with use of the student t-test and chi-square test, where appropriate (p < 0.05). z-Score was calculated for each subject for later correlation analysis, with the mean and standard deviation of data from control subjects used to define z-scores for both controls and patients. Domain scores for attention/working memory and executive function were computed by averaging the z-scores of each test.
3 Altered Thalamocortical FC in IGE Magnetic resonance imaging acquisition MR images were acquired on a Siemens Trio 3T scanner (Erlangen, Germany) with a 12-channel phased array head coil. For identification of structural abnormalities, the following conventional MR images were acquired simultaneously: axial T2-weighted and fluid-attenuated inversion recovery (FLAIR) images (4 mm thickness), oblique coronal T2-weighted and FLAIR images perpendicular to the long axis of hippocampus (3 mm thickness), and gadolinium-enhanced axial T1-weighted images (5 mm thickness). The MR images of all participants were reviewed by a board-certified neuroradiologist (S.I.S.) for any structural abnormalities and reported as normal. For VBM analysis, high-resolution three-dimensional (3D) magnetization prepared rapid gradient echo (MPRAGE) sequence was acquired with the following parameters: repetition time (TR) = 1,780 msec, echo time (TE) = 2.34 msec, voxel size = 1 mm3. For fcMRI analysis, 245 volumes of echo planar imaging (EPI) were acquired (TR = 2,000 msec, TE = 30 msec, voxel dimensions = 3.4 9 3.4 9 3.75 mm3). All participants were instructed to relax, think of nothing in particular, and keep their eyes closed but remain awake during scanning. No sedative drug was given and all participants reported no sleep and seizures during the scanning. Voxel-based morphometry Data were processed and analyzed using SPM8 (http:// www.fil.ion.ucl.ac.uk/spm), where we applied VBM implemented in the VBM8 toolbox (http://dbm.neuro.uni-jena. de/vbm.html) with default parameters. Images were bias corrected, tissue classified, and registered using linear (12parameter affine) and nonlinear transformations within a unified segmentation model. Subsequently, analyses were performed on the volumes of the gray matter (GM) segments, which were multiplied by nonlinear components derived from the normalization matrix to compensate for possible volume changes during the nonlinear spatial normalization procedure. These modulated GM images were then smoothed with an 8-mm full-width half-maximum (FWHM) isotropic Gaussian kernel. Regionally specific differences in GM volume between patients and controls were assessed using an analysis of covariance (ANCOVA) with total intracranial volume, age, and gender as nuisance variables. Statistical significance was set at a height threshold of p < 0.001 and an extent threshold of cluster-level p < 0.05, corrected for multiple comparisons using family-wise error (FWE). Seed-based functional connectivity analysis The fcMRI data were processed and analyzed using SPM8 and REST software (http://rfmri.org). The first five volumes were discarded to ensure steady-state longitudinal magnetization. Data preprocessing included slice-timing correction, head motion correction, coregistration of EPI
to high-resolution T1-weighted image, and spatial normalization by applying transformation matrix obtained during the segmentation of T1-weighted images. Five patients and three control subjects, who had excessive head motion (>1 mm of maximum displacement in the x, y, or z directions and 1 degree of angular rotation about each axis), were excluded from the statistical analysis. Spatially normalized images were then resampled to an isotropic voxel size of 3 9 3 9 3 mm3 and smoothed with a 6-mm FWHM Gaussian kernel. Linear detrending and temporal band pass (0.01–0.08 Hz) filtering were made to remove low-frequency drifts and physiologic high-frequency noise. In addition, global mean signal, six motion parameters, cerebrospinal fluid signal, and white matter signal were regressed out to reduce the effects of head motion and nonneuronal BOLD fluctuations.16 For seed-based FC analysis, thalamic region of GM volume reduction based on the VBM result was used as the seed. The mean time series of the thalamic seed was calculated by averaging the time series of all voxels within the seed. The time series of the thalamic seed was then correlated with the time series of the remaining voxels in the whole brain, resulting in a correlation map for each participant that contained correlation coefficient for each voxel with that of the thalamic seed. The resulting correlation coefficients for each voxel (r values) were converted to z scores by using Fisher’s r-to-z transformation to improve the gaussianity of their distribution. The z-transformed FC maps were used as the FC strength for between-group comparison and correlation analysis. Because there is debate concerning the physiologic significance of negative connectivity,16,17 only positive connectivity maps were analyzed in our study. Group comparison of FC maps between whole IGE patients and controls was assessed using two-sample t-test, and statistical significance was set at a height threshold of p < 0.001 and an extent threshold of cluster-level p < 0.05, corrected for multiple comparisons using FWE. Subgroup analyses were further performed to determine similarities and differences in thalamocortical FC between two subsyndromes of IGE, JME, and GTCS. To this end, one-way analysis of variance (ANOVA) was first performed to identify overall differences in thalamocortical FC among the three groups (42 controls, 31 JME patients, and 17 GTCS patients) using an F contrast (cluster-level FWE-corrected p < 0.05). Post hoc pairwise comparisons were then made between the three groups (control vs. JME, control vs. GTCS, and JME vs. GTCS) using T contrasts with Bonferroni correction (two-sample t-test, cluster-level FWE-corrected p < 0.017 [0.05/3]). Within-group correlation analysis was performed to delineate possible relationships between thalamocortical FC strength and neuropsychological measures and clinical variables in patients. The z-scaled domain scores for both attention/working memory and executive function were Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
4 J. B. Kim et al. correlated with z-scaled FC strength of thalamocortical alterations (partial correlation controlling for the effect of age followed by Bonferroni correction, p < 0.05). In addition, z-scaled FC strength was correlated with clinical variables such as disease duration (partial correlation controlling for the effect of age followed by Bonferroni correction, p < 0.05) and number of generalized tonic–clonic seizures per year (Pearson correlation followed by Bonferroni correction, p < 0.05) in the patient group. Multiple stepwise linear regression analysis was then performed to assess the influence of variables of clinical importance (duration of epilepsy, frequency of GTCS, attention/working memory domain score, and executive domain score) on the FC strength of altered thalamocortical network. Bonferroni correction was further applied to correct for multiple comparisons. Statistical analyses were performed by using the Statistical Package for Social Sciences (SPSS, Version 19.0; IBM, Armonk, NY, U.S.A.).
Results Clinical characteristics Forty-nine patients (23 women; mean age 26 [standard deviation] 7 years) and 42 control subjects (20 women; mean age 27 6 years) were finally included for statistical analysis. Two groups did not differ in age, gender, and education years (all p > 0.05, Table 1). Patient group consisted of 31 JME, 17 GTCS, and one juvenile absence epilepsy. Mean age of seizure onset was 15.5 3.8 years (range 9– 26 years), and mean duration of epilepsy was 10.4
6.1 years (range 1–27 years). Mean number of GTCS per year was 1.2 0.8 (range 0–3.5). AEDs at the time of study consisted of valproate monotherapy in 30 (61%), lamotrigine monotherapy in 6 (12%), levetiracetam monotherapy in 6 (12%), topiramate monotherapy in 2 (4%), valproate/lamotrigine polytherapy in 2 (4%), and valproate/ levetiracetam polytherapy in 3 (6%). Eighteen patients (37%) had positive family history of epilepsy in their firstdegree relatives. Neuropsychological assessment Forty-four patients and 38 controls completed neuropsychological assessment (Table 1). There was no difference in MMSE score between the groups. Results of attention and working memory tests showed that patients had poorer performance than controls in Trail-Making Test part A (p < 0.001), Digit Span forward score (p < 0.001), and Digit Span backward score (p < 0.001). In addition, patients performed worse than controls in executive function tests, including Trail-Making Test part B (p = 0.001), Stroop word-color test (p = 0.002), and Letter fluency test (p < 0.001). Voxel-based morphometry Compared to controls, patients with IGE had a significant regional GM volume reduction in the anteromedial thalamus (FWE-corrected p = 0.036; peak z score = 4.5; Montreal Neurological Institute [MNI] coordinates of local maxima = 5/ 7/6) (Fig. 1). No area of increased GM volume was found in patients compared to controls at the same threshold.
Table 1. Clinical characteristics and neuropsychological measures in patients with idiopathic generalized epilepsy and control subjects IGE patients (n = 49) Demographic and clinical data Age (years) Gender (F:M) Education years IGE subsyndrome (no. of patients) Age of seizure onset (years) Duration of epilepsy (years) Seizure frequency (no. of generalized tonic–clonic seizures per year) Neuropsychological data Mini-Mental State Examination score Trail-Making Part A (s) Digit Span Forward scaled score Digit Span Backward scaled score Trail-Making Part B (s) Stroop I (word, s) Stroop II (color, s) Stroop III (word-color, s) Letter fluency test
25.9 7.2 23: 26 14.5 2.1 JME (31), GTCS (17), JAE (1) 15.5 3.8 (range, 9–26) 10.4 6.1 (range, 1–27) 1.2 0.8 (range, 0–3.5)
29.3 28.6 8.9 7.2 67.9 13.2 14.2 18.5 36.1
n = 44 0.8 12.9 2.4 2.9 43.0 2.7 2.5 3.3 12.9
Control subjects (n = 42)
p-Value
27.0 5.9 20: 22 14.8 1.6
0.421 0.948 0.351
n = 38 29.6 0.7 20.0 5.0 11.2 1.6 10.3 2.6 43.1 18.5 12.2 2.7 13.0 3.0 16.2 3.3 48.6 9.9
0.125
Altered thalamocortical functional connectivity in idiopathic generalized epilepsy *Jung Bin Kim, †Sang-il Suh, *Woo-Keun Seo, *Kyungmi Oh, *Seong-Beom Koh, and *Ji Hyun Kim Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
SUMMARY
Dr. Ji Hyun Kim is an associate professor in Neurology at Korea University Guro Hospital, Seoul, Republic of Korea.
Objective: Aberrant thalamocortical network has been hypothesized to play a crucial role in the fundamental pathogenesis underlying idiopathic generalized epilepsy (IGE). We aimed to investigate alterations of thalamocortical functional network in patients with IGE using thalamic seed-based functional connectivity (FC) analysis, and their relationships with frontal cognitive functions and clinical characteristics. Methods: Forty-nine IGE patients (31 with juvenile myoclonic epilepsy, 17 with IGE with generalized tonic–clonic seizures only, one with juvenile absence epilepsy) and 42 control subjects were prospectively recruited. Voxel-based morphometry (VBM) was first performed to detect thalamic region of gray matter (GM) reduction in patients compared to controls. Between-group comparison of thalamocortical FC was then carried out using resting-state functional magnetic resonance imaging (MRI) analysis seeding at thalamic region of volume difference. In addition, thalamocortical FC was correlated with frontal cognitive performance and clinical variables. Results: Neuropsychological assessment revealed that patients with IGE had poorer performance than controls on most of the frontal cognitive functions. VBM detected a reduction in GM in the anteromedial thalamus in patients relative to controls. FC analysis seeding at the anteromedial thalamus revealed a reduction of thalamocortical FC in the bilateral medial prefrontal cortex and precuneus/posterior cingulate cortex in patients with IGE compared to controls. Thalamocortical FC strength of bilateral medial prefrontal cortex correlated negatively with disease duration, but did not correlate with seizure frequency or frontal cognitive functions in patients with IGE. Significance: Our results indicate that IGE is associated with decreased thalamocortical FC between anteromedial thalamus and medial prefrontal cortex and precuneus/ posterior cingulate cortex. Our finding of greater reduction of medial prefrontal FC in relation to increasing disease duration suggests that thalamoprefrontal network abnormality, the proposed pathophysiologic mechanism underlying IGE, may be the consequence of the long-standing burden of the disease. KEY WORDS: Idiopathic generalized epilepsy, Resting-state functional connectivity, Thalamocortical network, Frontal cognitive dysfunction.
Idiopathic generalized epilepsy (IGE) constitutes a heterogeneous group of epilepsy syndromes with a nonfocal mechanism of seizure onset and no identifiable cause
other than a genetic predisposition. Childhood absence epilepsy (CAE), juvenile absence epilepsy, juvenile myoclonic epilepsy (JME), and IGE with generalized tonic–clonic seizures only (GTCS) are the well-recognized subsyndromes of IGE, according to predominant seizure semiology and age of seizure onset. Based on genetic traits, similar seizure semiology, and typical electroencephalography (EEG) features of 3- to 5-Hz generalized spike-wave discharges (GSWDs), these IGE subsyndromes are considered to share a common pathogenetic mechanism.
Accepted January 27, 2014. Departments of *Neurology and † Radiology, Korea University Guro Hospital, Korea University College of Medicine, Seoul, Korea Address correspondence to Ji Hyun Kim, Department of Neurology, Korea University Guro Hospital, Korea University College of Medicine, 152-703, Guro-dong Ro 148, Guro-gu, Seoul, Republic of Korea. E-mail: [email protected] Wiley Periodicals, Inc. © 2014 International League Against Epilepsy
1
2 J. B. Kim et al. The fundamental pathogenesis that underlies IGE is not fully elucidated; however, converging evidence has suggested a critical role of abnormal thalamocortical circuit in the generation of GSWDs. Recent advances in computational analysis of multimodal neuroimaging, such as voxel-based morphometry (VBM), magnetic resonance spectroscopy (MRS), diffusion tensor imaging (DTI), and functional magnetic resonance imaging (fMRI) have contributed greatly to the understanding of structural and functional changes of the thalamus and frontal cortex in IGE.1,2 In particular, contribution of the thalamocortical network to the generation of GSWDs has been highlighted in simultaneous EEG-fMRI studies that showed both thalamic activation and widespread cortical deactivation in IGE.3,4 Recently, resting-state functional connectivity MRI (fcMRI) has been used increasingly to investigate largescale functional networks at a whole-brain level, based on the temporal correlation of spontaneous fluctuations of blood oxygen level–dependent (BOLD) signals in a very low frequency range (0.01–0.1 Hz).5,6 Because fcMRI reliably identifies a variety of intrinsic corticocortical and corticosubcortical networks,5 it is widely used to noninvasively evaluate changes in functional brain networks in diverse neurologic and psychiatric disorders as well as in the normal aging process.6 Previous fcMRI studies have consistently shown decreased functional connectivity (FC) within the default mode network (DMN) in IGE, suggesting a disruption of functional integration of DMN in association with impaired consciousness and cognitive dysfunctions.7–11 Given the hypothesis of thalamocortical network abnormality in the fundamental pathomechanism underlying IGE, investigation of the thalamocortical FC seems more likely to provide valuable information regarding functional changes of this network. The purpose of the current study was to investigate FC alterations within the thalamocortical network in IGE, by using the thalamic seed derived from the VBM of group difference between patients with IGE and controls. We predicted that FC between thalamus and frontal cortex may be altered in patients relative to controls. Correlation analyses were additionally performed between thalamocortical FC and neuropsychological measures and clinical variables to explore the possible influence of the clinical and cognitive factors on FC changes.
Methods Subjects We prospectively recruited 54 right-handed adult patients with IGE (33 patients with JME, 20 patients with GTCS, one patient with juvenile absence epilepsy), who were followed at least 1 year in the epilepsy clinic of Korea University Guro Hospital. This study was an extension of our previous works.12,13 Diagnosis of IGE was based on Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
electroclinical criteria according to the International League Against Epilepsy (ILAE) classification. Inclusion criteria we used were as follows: (1) unequivocal seizure semiology of IGE—typical absence seizure, myoclonic seizure involving the bilateral upper extremities preferentially occurring early in the morning, or GTCS; (2) no evidence of developmental and neurologic abnormalities, and global cognitive impairment on Mini-Mental State Examination (MMSE score ≥ 28/30); (3) at least one EEG examination demonstrating typical GSWDs on a normal background; and (4) neither abnormal nor unusual findings on clinical MRI. Patients with comorbid neurologic, psychiatric, or chronic systemic disorders were excluded. None of the patients were taking medications other than antiepileptic drugs (AEDs) at the time of study inclusion. Demographic data and clinical information such as seizure semiology, age of seizure onset, duration of epilepsy, seizure frequency, and current AEDs were obtained through interviews with the patients and their parents and reviews of medical records. For group comparison, 45 right-handed healthy volunteers matched for age and gender were recruited to serve as controls. All control subjects underwent neurologic examination as well as a detailed interview to ensure that they had (1) no neurologic abnormality and global cognitive impairment (MMSE score ≥ 28/30); (2) no history of neurologic, psychiatric, or systemic disorders; (3) no family history of epilepsy; and (4) no history of alcohol or drug abuse. Control subjects with abnormal or unusual MRI findings were also excluded. The local ethics committee approved the study protocol, and all participants provided written informed consent prior to study inclusion. Neuropsychological assessment Neuropsychological assessment was carried out by an experienced neuropsychologist, who was blinded to the clinical diagnosis, on the same day of MRI scanning. Because IGE patients are known to have frontal cognitive dysfunctions,14,15 the neuropsychological battery was more weighted on the frontal cognitive functions. Assessed domains and the tests were as follows: (1) global cognitive function—Korean version of MMSE; (2) attention and working memory—Trail-Making Test part A, Digit Span Test (forward and backward); and (3) executive function – Trail-Making Test part B, Stroop Test (word, color, wordcolor), and Letter Fluency Test (words beginning with the three Korean letters). Group comparisons of demographic data and neuropsychological measures were undertaken with use of the student t-test and chi-square test, where appropriate (p < 0.05). z-Score was calculated for each subject for later correlation analysis, with the mean and standard deviation of data from control subjects used to define z-scores for both controls and patients. Domain scores for attention/working memory and executive function were computed by averaging the z-scores of each test.
3 Altered Thalamocortical FC in IGE Magnetic resonance imaging acquisition MR images were acquired on a Siemens Trio 3T scanner (Erlangen, Germany) with a 12-channel phased array head coil. For identification of structural abnormalities, the following conventional MR images were acquired simultaneously: axial T2-weighted and fluid-attenuated inversion recovery (FLAIR) images (4 mm thickness), oblique coronal T2-weighted and FLAIR images perpendicular to the long axis of hippocampus (3 mm thickness), and gadolinium-enhanced axial T1-weighted images (5 mm thickness). The MR images of all participants were reviewed by a board-certified neuroradiologist (S.I.S.) for any structural abnormalities and reported as normal. For VBM analysis, high-resolution three-dimensional (3D) magnetization prepared rapid gradient echo (MPRAGE) sequence was acquired with the following parameters: repetition time (TR) = 1,780 msec, echo time (TE) = 2.34 msec, voxel size = 1 mm3. For fcMRI analysis, 245 volumes of echo planar imaging (EPI) were acquired (TR = 2,000 msec, TE = 30 msec, voxel dimensions = 3.4 9 3.4 9 3.75 mm3). All participants were instructed to relax, think of nothing in particular, and keep their eyes closed but remain awake during scanning. No sedative drug was given and all participants reported no sleep and seizures during the scanning. Voxel-based morphometry Data were processed and analyzed using SPM8 (http:// www.fil.ion.ucl.ac.uk/spm), where we applied VBM implemented in the VBM8 toolbox (http://dbm.neuro.uni-jena. de/vbm.html) with default parameters. Images were bias corrected, tissue classified, and registered using linear (12parameter affine) and nonlinear transformations within a unified segmentation model. Subsequently, analyses were performed on the volumes of the gray matter (GM) segments, which were multiplied by nonlinear components derived from the normalization matrix to compensate for possible volume changes during the nonlinear spatial normalization procedure. These modulated GM images were then smoothed with an 8-mm full-width half-maximum (FWHM) isotropic Gaussian kernel. Regionally specific differences in GM volume between patients and controls were assessed using an analysis of covariance (ANCOVA) with total intracranial volume, age, and gender as nuisance variables. Statistical significance was set at a height threshold of p < 0.001 and an extent threshold of cluster-level p < 0.05, corrected for multiple comparisons using family-wise error (FWE). Seed-based functional connectivity analysis The fcMRI data were processed and analyzed using SPM8 and REST software (http://rfmri.org). The first five volumes were discarded to ensure steady-state longitudinal magnetization. Data preprocessing included slice-timing correction, head motion correction, coregistration of EPI
to high-resolution T1-weighted image, and spatial normalization by applying transformation matrix obtained during the segmentation of T1-weighted images. Five patients and three control subjects, who had excessive head motion (>1 mm of maximum displacement in the x, y, or z directions and 1 degree of angular rotation about each axis), were excluded from the statistical analysis. Spatially normalized images were then resampled to an isotropic voxel size of 3 9 3 9 3 mm3 and smoothed with a 6-mm FWHM Gaussian kernel. Linear detrending and temporal band pass (0.01–0.08 Hz) filtering were made to remove low-frequency drifts and physiologic high-frequency noise. In addition, global mean signal, six motion parameters, cerebrospinal fluid signal, and white matter signal were regressed out to reduce the effects of head motion and nonneuronal BOLD fluctuations.16 For seed-based FC analysis, thalamic region of GM volume reduction based on the VBM result was used as the seed. The mean time series of the thalamic seed was calculated by averaging the time series of all voxels within the seed. The time series of the thalamic seed was then correlated with the time series of the remaining voxels in the whole brain, resulting in a correlation map for each participant that contained correlation coefficient for each voxel with that of the thalamic seed. The resulting correlation coefficients for each voxel (r values) were converted to z scores by using Fisher’s r-to-z transformation to improve the gaussianity of their distribution. The z-transformed FC maps were used as the FC strength for between-group comparison and correlation analysis. Because there is debate concerning the physiologic significance of negative connectivity,16,17 only positive connectivity maps were analyzed in our study. Group comparison of FC maps between whole IGE patients and controls was assessed using two-sample t-test, and statistical significance was set at a height threshold of p < 0.001 and an extent threshold of cluster-level p < 0.05, corrected for multiple comparisons using FWE. Subgroup analyses were further performed to determine similarities and differences in thalamocortical FC between two subsyndromes of IGE, JME, and GTCS. To this end, one-way analysis of variance (ANOVA) was first performed to identify overall differences in thalamocortical FC among the three groups (42 controls, 31 JME patients, and 17 GTCS patients) using an F contrast (cluster-level FWE-corrected p < 0.05). Post hoc pairwise comparisons were then made between the three groups (control vs. JME, control vs. GTCS, and JME vs. GTCS) using T contrasts with Bonferroni correction (two-sample t-test, cluster-level FWE-corrected p < 0.017 [0.05/3]). Within-group correlation analysis was performed to delineate possible relationships between thalamocortical FC strength and neuropsychological measures and clinical variables in patients. The z-scaled domain scores for both attention/working memory and executive function were Epilepsia, **(*):1–9, 2014 doi: 10.1111/epi.12580
4 J. B. Kim et al. correlated with z-scaled FC strength of thalamocortical alterations (partial correlation controlling for the effect of age followed by Bonferroni correction, p < 0.05). In addition, z-scaled FC strength was correlated with clinical variables such as disease duration (partial correlation controlling for the effect of age followed by Bonferroni correction, p < 0.05) and number of generalized tonic–clonic seizures per year (Pearson correlation followed by Bonferroni correction, p < 0.05) in the patient group. Multiple stepwise linear regression analysis was then performed to assess the influence of variables of clinical importance (duration of epilepsy, frequency of GTCS, attention/working memory domain score, and executive domain score) on the FC strength of altered thalamocortical network. Bonferroni correction was further applied to correct for multiple comparisons. Statistical analyses were performed by using the Statistical Package for Social Sciences (SPSS, Version 19.0; IBM, Armonk, NY, U.S.A.).
Results Clinical characteristics Forty-nine patients (23 women; mean age 26 [standard deviation] 7 years) and 42 control subjects (20 women; mean age 27 6 years) were finally included for statistical analysis. Two groups did not differ in age, gender, and education years (all p > 0.05, Table 1). Patient group consisted of 31 JME, 17 GTCS, and one juvenile absence epilepsy. Mean age of seizure onset was 15.5 3.8 years (range 9– 26 years), and mean duration of epilepsy was 10.4
6.1 years (range 1–27 years). Mean number of GTCS per year was 1.2 0.8 (range 0–3.5). AEDs at the time of study consisted of valproate monotherapy in 30 (61%), lamotrigine monotherapy in 6 (12%), levetiracetam monotherapy in 6 (12%), topiramate monotherapy in 2 (4%), valproate/lamotrigine polytherapy in 2 (4%), and valproate/ levetiracetam polytherapy in 3 (6%). Eighteen patients (37%) had positive family history of epilepsy in their firstdegree relatives. Neuropsychological assessment Forty-four patients and 38 controls completed neuropsychological assessment (Table 1). There was no difference in MMSE score between the groups. Results of attention and working memory tests showed that patients had poorer performance than controls in Trail-Making Test part A (p < 0.001), Digit Span forward score (p < 0.001), and Digit Span backward score (p < 0.001). In addition, patients performed worse than controls in executive function tests, including Trail-Making Test part B (p = 0.001), Stroop word-color test (p = 0.002), and Letter fluency test (p < 0.001). Voxel-based morphometry Compared to controls, patients with IGE had a significant regional GM volume reduction in the anteromedial thalamus (FWE-corrected p = 0.036; peak z score = 4.5; Montreal Neurological Institute [MNI] coordinates of local maxima = 5/ 7/6) (Fig. 1). No area of increased GM volume was found in patients compared to controls at the same threshold.
Table 1. Clinical characteristics and neuropsychological measures in patients with idiopathic generalized epilepsy and control subjects IGE patients (n = 49) Demographic and clinical data Age (years) Gender (F:M) Education years IGE subsyndrome (no. of patients) Age of seizure onset (years) Duration of epilepsy (years) Seizure frequency (no. of generalized tonic–clonic seizures per year) Neuropsychological data Mini-Mental State Examination score Trail-Making Part A (s) Digit Span Forward scaled score Digit Span Backward scaled score Trail-Making Part B (s) Stroop I (word, s) Stroop II (color, s) Stroop III (word-color, s) Letter fluency test
25.9 7.2 23: 26 14.5 2.1 JME (31), GTCS (17), JAE (1) 15.5 3.8 (range, 9–26) 10.4 6.1 (range, 1–27) 1.2 0.8 (range, 0–3.5)
29.3 28.6 8.9 7.2 67.9 13.2 14.2 18.5 36.1
n = 44 0.8 12.9 2.4 2.9 43.0 2.7 2.5 3.3 12.9
Control subjects (n = 42)
p-Value
27.0 5.9 20: 22 14.8 1.6
0.421 0.948 0.351
n = 38 29.6 0.7 20.0 5.0 11.2 1.6 10.3 2.6 43.1 18.5 12.2 2.7 13.0 3.0 16.2 3.3 48.6 9.9
0.125
