NeuroImage 60 (2012) 299–304

Contents lists available at SciVerse ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r .com/ locate /yn img

Resting-state fMRI study of treatment-naïve temporal lobe epilepsy patients withdepressive symptoms

Sihan Chen a,b, Xintong Wu b, Su Lui a, Qizhu Wu a, Zhiping Yao b, Qifu Li b, Dongmei Liang c, Dongmei An b,Xiaoyun Zhang b, Jiajia Fang b, Xiaoqi Huang a, Dong Zhou b,⁎, Qi-Yong Gong a,⁎⁎a Huaxi MR Research Center (HMRRC), Department of Radiology, State Key Laboratory of the Biotherapy, West China Hospital of Sichuan University, Chinab Department of Neurology, West China Hospital of Sichuan University, Chengdu, 610041, Chinac Laboratory of Laser Sports Medicine, South China Normal University, Guangzhou, 510006, China

⁎ Correspondence to: D. Zhou, Department of NeuroSichuan University, No. 37 Guo Xue Xiang Chengdu,85422550.⁎⁎ Correspondence to: Q.-Y. Gong, Huaxi MR Research CRadiology, West China Hospital of Sichuan University, N610041, China. Fax: +86 28 85423503.

E-mail addresses: zhoudong66@yahoo.de (D. Zhou),(Q.-Y. Gong).

1053-8119/$ – see front matter © 2011 Published by Eldoi:10.1016/j.neuroimage.2011.11.092

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 12 October 2010Revised 25 November 2011Accepted 30 November 2011Available online 10 December 2011

Keywords:EpilepsyDepressionfMRIFunctional connectivity

Background: Patients with temporal lobe epilepsy are at high risk for comorbid depression, and it is hypoth-esized that these two diseases are share common pathogenic pathways. We aimed to characterize regionalbrain activation in treatment-naïve temporal lobe epilepsy patients with depressive symptoms and comparethe results to epilepsy patients without depressive symptoms and to healthy controls.Subjects and methods: We recruited 23 treatment-naïve patients (including anti-epilepsy drugs (AEDs) andantidepressants) and 17 matched healthy controls for this study. The patients were further divided intotwo groups: patients with depressive symptoms and patients without; the patients then used a self-ratingdepression scale (SDS) to assess their depression. All participants underwent resting functional magnetic res-onance imaging (fMRI) scans using the Trio Tim magnetic resonance (MR) image system (3.0 T). The datawere processed and analyzed using REST and SPSS11.5 software.

Results: The patients with depressive symptoms showed significantly higher activity in the bilateral thala-mus, insula and caudate and right anterior cingulate compared with the two other groups (pb0.05, cor-rected). Brain network connectivity analysis revealed that connectivity decreased in the prefrontal-limbicsystem and increased within the limbic system and angular gyrus in patients with depressive symptoms(pb0.05, corrected).Conclusion: The epilepsy patients with depressive symptoms showed regional brain activity alterations anddisruption of the mood regulation network at the onset of seizures. The present study offers further insightinto the underlying neuropathophysiology of epilepsy with depressive symptoms.

© 2011 Published by Elsevier Inc.

Introduction

Up to 20 to 55% of patients with temporal lobe epilepsy (TLE) havepsychological depressive symptoms (Kanner, 2008; Tellez-Zenteno etal., 2007). This prevalence is significantly higher than in patients withother chronic diseases, such as diabetes and asthma (Ettinger et al.,2004), and the symptoms lead to a decreased quality of life and ahigher risk of suicide (Christensen et al., 2007). Furthermore, whenmajor depression precedes the first epileptic attack, it is associatedwith a 1.7-fold increased risk of developing seizures (Hesdorffer et

logy, West China Hospital of610041, China. Fax: +86 28

enter (HMRRC), Department ofo. 37 Guo Xue Xiang Chengdu,

qiyonggong@hmrrc.org.cn

sevier Inc.

al., 2006). Contrary to traditional beliefs, which proposed that depres-sive symptoms result from a secondary reaction to the epileptic epi-sodes, Kanner (2004)suggested that depressive symptoms mayshare common pathogenic pathways with epilepsy and that thesecommon pathways may facilitate the development of one conditionin the presence of the other. Moreover, depressive disorders may bea predictor of treatment-resistant epilepsy (Kanner, 2008).

To date, animal models and human neuroimaging studies havebeen used to study the bidirectional relationship between the twoconditions (Gilliam et al., 2007; Hasler et al., 2007; Mazarati et al.,2007; Savic et al., 2004). A rat animal model demonstrated that de-pressive behavior developed due to neuronal plasticity that was asso-ciated with repeated electrical kindling (Mazarati et al., 2007). Apositron emission tomography (PET) study found significantly de-creased serotonin (5HT1A) binding potentials (BP) in the epileptogen-ic hippocampus and amygdala and a negative correlation betweendepressive symptoms and the BPs in the anterior cingulate (ACC)(Savic et al., 2004). Using 1H-MRS to detect the creatine/N-acetylas-partate (Cr/NAA) ratio, Gilliam et al. (2007) found abnormalities in

300 S. Chen et al. / NeuroImage 60 (2012) 299–304

the hippocampus that correlated with depressive symptoms in pa-tients with TLE (Gilliam et al., 2007). These findings highlight the in-volvement of the prefrontal-limbic system and paralimbic structuresin these conditions. However, previous neuroimaging experimentswere conducted on patients receiving anti-epilepsy drugs (AEDs)and antidepressants; thus, the potential impact of AEDs and antide-pressants on brain activity could not be ruled out. To the best of ourknowledge, no experiments on treatment-naïve patients with epilep-sy (PWEs) have been carried out, and these studies may be critical toelucidate the underlying mechanisms.

Resting state functional magnetic resonance imaging (fMRI) em-ploys the amplitude of low-frequency (0.01–0.08 Hz) fluctuations(ALFF) of the blood-oxygen-level dependent (BOLD) signal to detectregional activity alterations and is widely used to study neuropsycho-logical illnesses. Resting state fMRI can provide information not onlyon synchronous regional cerebral activity using the ALFF intensity(Goncalves et al., 2006; Shmuel and Leopold, 2008), but it also avoidsperformance confounds in task-design fMRI research (Callicott et al.,2003). Moreover, resting state fMRI allows the integrity of brain net-works to be examined using connectivity analysis.

The present investigation is the first to use resting state fMRI toexamine both regional ALFF intensity and brain network connectivityin treatment-naïve TLE patients with and without depressive symp-toms compared with a cohort of normal controls. We hypothesizedthe following: 1. TLE patients with depressive symptoms show ALFFalterations in the prefrontal lobe-limbic systems; and 2. connectivityanalysis in the patients with depression symptoms (PDSs) group re-veal disruption of the mood regulation network compared with thetwo other groups, while the patients without depression symptoms(nPDSs) group would show epileptic network alteration.

Materials and methods

Participants

A total of 23 treatment naïve (including AEDs and antidepres-sants) TLE patients (all right handed, male/female: 13/10, averageage: 35.0±8.8) and 17 healthy controls (all right handed, male/fe-male: 10/7, average age: 34.9±6.9) were recruited from the out-patient department in the West China Hospital of Sichuan University,China, from July, 2009, to August, 2010. This study was approved bythe local ethics committee, and all participants signed informed con-sent forms. Patients had clinical interviews and physical examina-tions that were conducted by board-certified neurologists. Patientsevaluated their existing depressive symptoms using the self-ratingdepression scale (SDS) (Zung, 1965). The inclusion criteria includedthe following: 1. seizure types and epileptic syndromes as diagnosedaccording to the classification of the International League Against Ep-ilepsy (Anon., 1981, 1989); and 2. TLE diagnosis when continuousinterictal–ictal scalp video electroencephalography showed complexpartial seizures of temporal origin. The exclusion criteria were as fol-lows: 1. patients with structural lesions other than medial temporalsclerosis; 2. patients with a history of psychological diseases, or useof anti-depressive drugs, or other chronic diseases, such as diabetesand asthma; 3. pregnant and nursing women; and 4. patients who ex-perienced seizures within 15 days prior to fMRI scanning. All controlswere also screened using the SCID (non-patient version) to confirmthe lifetime absence of psychiatric illness.

Data acquisition

MR images detecting BOLD signal were obtained using a Trio Tim(3 T) magnetic resonance (MR) imaging system (Siemens; Erlangen)with a gradient-echo echo-planar imaging sequence: repetition time/echo time (TR/TE), 2000/30 ms; voxel size, 3.75×3.75×5 mm3; flipangle, 90°; slice thickness, 5 mm (no gap); matrix, 64×64; and FOV,

240×240 mm2. Each brain volume comprised 30 axial slices, andeach functional run contained 200 volumes following 5 dummy vol-umes, with a total scan time of 416 s. All participants were instructedto keep their eyes closed and not to think of anything in particularduring the resting-state MR scan.

Data processing

Functional image preprocessing was carried out using DPARSF(State Key Laboratory of Cognitive Neuroscience and Learning at Bei-jing Normal University; http://resting-fmri.sourceforge.net/) soft-ware following these steps: conversion of the DICOM data to NIFTIimages; removal of the first 10 time points from each patient's data;slice timing correction; realignment to the middle image; spatial nor-malization to the Montreal Neurological Institute (MNI) template;and resampling of each voxel to 3×3×3 mm3 without spatial filter-ing. ALFF maps were calculated using REST (State Key Laboratory ofCognitive Neuroscience and Learning at Beijing Normal University;http://resting-fmri.sourceforge.net). Following band pass filtering(0.01–0.08 Hz) and linear-trend removal, the time series was trans-formed to the frequency domain using a fast Fourier transform(FFT). The power spectrum obtained by FFT was square root-transformed and averaged across the frequency of 0.01–0.08 Hz. Theaveraged square root of activity comprised the ALFF. Functional con-nectivity (FC) was examined by analyzing a seed area provided byREST. Five areas were selected as seed regions based on ALFF findingsand former neuroimaging findings (Hermann et al., 2008) and includ-ed the right ACC, bilateral insula and thalamus.

Statistical analysis

ALFF maps from the patient and control groups were compared ona voxel-wise basis with an ANCOVA in REST, covarying for disease du-ration and number of seizures (when comparing with normal con-trols we didn't use disease duration and number of seizures ascovarying factors). Whole-brain analysis was conducted with statisti-cal threshold of Pb0.05 (FDR corrected). The FC map was comparedacross the three groups using a statistical threshold of Pb0.05(Alpha sim corrected).

Results

We acquired resting fMRI data from 23 patients (male/female: 13/10, average age: 35.0±8.8) and 17 normal controls (male/female:10/7, average age: 34.9±6.9) on a 3.0 T MR system (Siemens, TrioTim, Erlangen). The 23 PWE were divided into two groups based ontheir SDS score. The patients who scored over 50 were included inthe PDS group (male/female: 4/3, average age: 37.2±11.1) and theremaining patients were placed into the nPDS group (male/female:9/7, average age: 34.0±7.8). The disease duration, localization of ep-ileptic discharge and the number of seizures did not differ betweenthese two groups (P>0.05). Age, gender and years of education didnot differ between the three groups (Table 1).

We used the DPARSF (State Key Laboratory of Cognitive Neurosci-ence and Learning in Beijing Normal University; http://resting-fmri.sourceforge.net) software to explore alterations in regional ALFF(0.01–0.08 Hz) signals. By comparing the whole brain ALFF map be-tween the three groups, using ANCOVA, we found that PDSs, com-pared with nPDSs, showed a significantly elevated ALFF value in theright ACC, bilateral thalamus and insula and decreased ALFF in theleft amygdala and left medial prefrontal cortex (mPFC). When thePDS group was compared with healthy controls, the ALFF showed in-creased values in the right ACC, bilateral caudate and insula. Therewere no significantly decreased areas in the PDS group. The nPDSgroup showed increased regional activity in the right ACC and left an-terior insula compared with the healthy control group and decreased

Table 1Demographic information for the PDS/nPDS and control groups.

Characteristics Healthy control±SD(N=17)

nPDS±SD(N=16)

PDS±SD(N=7)

P

Male:Female 10:7 9:7 4:3 0.390Mean age, years 34.9±6.9 34.0±7.8 37.2±11.1 0.671Years of education 13.3±3.6 13.4±3.2 13.7±3.9 0.965SDS score – 34±6.5 62±13.3 b0.001Epilepsy duration/month

2.9±1.5 3.4±1.7 0.493

Number of seizures /n 3.0±0.9 3.6±0.9 0.247Focus

Right 7 3 0.563★Left 5 3Bilateral 4 1

Mesial temporalsclerosis

2 1

nPDS: Patients without depressive symptoms;PDS: Patients with depressive symptoms;n: number; ★:chi-square fisher exact test;SDS: self-rating depression scale;SD: standard deviation.

301S. Chen et al. / NeuroImage 60 (2012) 299–304

ALFF in the bilateral thalamus and right posterior insula ( Pb0.05, FDRcorrected, Table 2, Figs. 1.1–6).

Functional connectivity analysis was employed to explore alter-ations in the brain network. The selection of seed regions was basedon findings from the whole brain analysis. Connectivity maps werecompared across the three groups using an ANCOVA. We found a de-creased brain network between the limbic system, temporal lobe andprefrontal lobe, and increased connectivity between the limbic sys-tem and angular gyrus when comparing the PDS group with thenPDS group. When comparing the PDS group with the healthy controlgroup, we found decreased connectivity among limbic system

Table 2The location of ALFF value altered areas in whole brain analysis.

Location Talairach (Peak voxel) Voxelsize

Pvalue

x y z

PDS>nPDSRight insula 33 18 6 39 *Left insula −33 9 9 11 *Right thalamus 12 −9 9 9 *Left thalamus −9 −12 3 19 *Right ACC 12 36 21 8 *

nPDS>PDSLeft MPFC −6 21 −18 19 *Left amygdala −15 0 −15 25 *

PDS>HCRight insula 39 24 −12 44 *Left insula −36 30 3 22 *Right caudate 12 18 3 7 *Left caudate −9 6 9 11 *Right ACC 9 42 9 30 *

nPDS>HCRight ACC 15 42 12 7 *Left insula −39 33 3 52 *

HC>nPDSRight insula 33 −21 3 19 *Left thalamus −15 −12 9 18 *Right thalamus 12 −9 6 21 *

*Pb0.05,HC: healthy control;nPDS: Patients without depressive symptoms;PDS: Patients with depressive symptoms;ACC: anterior cingulate cortex;MPFC: medial pre-frontal cortex;ALFF: amplitude of low-frequency fluctuatio.

components and the prefrontal lobe (Brodmann area 10) and in-creased connectivity among the limbic system, angular gyrus andprefrontal lobe (Brodmann area 8). When comparing the nPDSgroup with the HC group, we found increased connectivity amongthe precentral gyrus, superior temporal gyrus and limbic system(Pb0.05, alpha sim corrected, Table 3).

Discussion

We are the first to demonstrate that treatment naïve patientsshow altered regional brain activity in the prefrontal-limbic systemand a disruption in the mood regulation network. We found signifi-cantly increased ALFF values in the bilateral thalamus, insula andright ACC and decreased ALFF values in the left amygdala and leftmPFC for the PDS group compared with the nPDS group. Using con-nectivity analysis, we found a decreased brain network between thelimbic system, temporal lobe and inferior prefrontal lobe, and in-creased connectivity between the limbic system and the angulargyrus when comparing the PDS and nPDS groups. When comparingthe PDS group with the healthy control, we found decreased connec-tivity between the limbic system and prefrontal lobe (Brodmann area10) and increased connectivity between the limbic system, angulargyrus and prefrontal lobe (Brodmann area 8).

The pathogenic mechanism underlying depression in epilepsy pa-tients has received much attention in the last decade. Kanner (2004)suggested that epilepsy and psychiatric disorders may share commonpathogenic pathways. This proposal was supported by previous ex-periments on a kindling-induced animal model of depression and 5-HT1A receptor research in PWE with depressive symptoms(Mazarati et al., 2007; Savic et al., 2004). However, previous neuroim-aging experiments have focused on PDSs receiving AED treatment(Hasler et al., 2007). Our work demonstrates regional brain activityand connectivity alterations in treatment-naïve patients with shortseizure duration, thus ruled out this potential confounding factor.

The altered ALFF value may represent the abnormal regional activ-ity of the prefrontal-limbic system in the PDS group and nPDS group.The prefrontal-limbic system, including the amygdala, striatum,insula, ACC and thalamus, is implicated in mood regulation and thepathogenic pathways of epilepsy. The amygdala is a classic compo-nent of mood regulation and also of TLE seizures. Epileptic micewith the Bassoon gene mutation demonstrate pathological forms ofplasticity and an imbalance in brain-derived neurotrophic factor(BNDF) distribution in fast-spiking (FS) interneurons and mediumspiny (MS) neurons in the striatum with concomitant behavioral al-teration (Ghiglieri et al., 2010). The mPFC has been suggested to reg-ulate the limbic system, especially the amygdala (Foland et al., 2008).Post-mortem analysis has revealed reduced Really Interesting NewGene (RING) Finger 41 (Rnf41) expression in the prefrontal lobethat has proven critical in depressive anxiety-like behavior andbeta-carboline-induced seizures in animal models (Kim et al., 2009).In cross-sectional and longitudinal MR imaging research on corticalthickness, the cingulate cortex has been found to be thinner in TLE pa-tients, and atrophy of the insula has been associated with residual sei-zures in post-operation patients (Bernhardt et al., 2010). Our results,which demonstrate increased regional activity in the mood regulationsystem, support our first hypothesis. Considering the short durationof our patients' epilepsy, this alteration may not be due to the recur-rent chronic stress. Our results suggest a common etiology betweenthese two conditions (epilepsy and depression).

The epileptiform discharges affected the ALFF values and the con-nectivity (Bettus et al., 2011). Our patients' EEGs showed that theinterictal discharge localizations were matched between these twopatient groups. Thus, the lateral interictal discharges didn't introducea bias in the functional connectivity comparisons (Bettus et al., 2009).

In the PDS group, we found decreased connectivity between thelimbic system, temporal lobe and prefrontal lobe (inferior frontal

Fig. 1. The ALFF value alterations between each group.1.1: The PDS group showed higher ALFF values in the right anterior cingulate (ACC), bilateral thalamus and insula comparedwith the nPDS group; (Pb0.05).1.2: The nPDS group showed higher ALFF values in the left amygdala and left medial prefrontal cortex (mPFC) compared with the PDS group;(Pb0.05).1.3: The PDS group showed higher ALFF values in the right ACC, bilateral caudate and insula compared with the healthy controls; (Pb0.05).1.4–5: The nPDS group showedincreased regional activity in the right ACC and left anterior insula compared with healthy controls; (Pb0.05).1.6: The healthy control group showed higher ALFF values in the bi-lateral thalamus and right posterior insula compared with the nPDS group; (Pb0.05).

302 S. Chen et al. / NeuroImage 60 (2012) 299–304

gyrus and Brodmann area 10) compared with the nPDS group. Thesedata are consistent with the classical top-down mood regulation net-work that implicates the pre-frontal lobe and limbic system (Folandet al., 2008). Kushnir noted activation of the angular gyrus in acuteabstinent nicotine-dependent individuals when facing cigarettes

Table 3The alteration of brain network in PDS/nPDS groups.

Groups Seeds Connectivity network

R-Ins L-Ins R-PCC R-Para L-Para R-Tha L-Tha L-Ca

PDS vs nPDS R-ACCR-ThaL-ThaR-CauL-Cau

PDS vs HC R-ACCR-ThaL-ThaR-CauL-Cau

R-Ins L-Ins R-PCC R-Para L-Para R-Tha L-Tha L-CaR-ACC

nPDS vs HC R-ThaL-ThaR-CauL-Cau

: Red dot represents increased functional connectivity.: Blue dot for the decreased functional connectivity (Pb0.05 Alpha sim corrected).

L: left; R: right; Ins: Insula; PCC: posterior cingulate; Para: Parahippocampus gyrus; Amy: AFron: Frontal lobe, 1,4,6,8 stand for inferior frontal and brodmann area 10(B10), 9, 10, 11 fTem: Temporal lobe, 2, 3, 5, 7 for medial temporal gyrus; 12, 14, 16, 19, 21 for superior tem§: for increased connnectivity among superior frontal gyrus (brodmann area 8) and decrea

cues (Kushnir et al., 2010). The angular gyrus is part of the visuospa-tial attention circuit. Our result, showing increased connectivity be-tween the limbic system and angular gyrus, may indicate theheightened sensitivity of the PDS group to their surrounding environ-ment (Kushnir et al., 2010). A recent review noted the damaged brain

u R-Ang L-Ang R-Fron L-Fron L-ACC R-Amy L-Amy R-Tem L-Tem

§ §

u R-Ang L-Ang R-Fron L-Fron L-ACC R-Amy L-Amy R-Tem L-Tem

mygdala; Tha: Thalamus; Cau: Caudate; ACC: anterior cingulate; Ang: Angular gyrus;or B8; 13, 15, 17, 18, 20 for precentral gyrus;poral gyrus;sed connectivity among B10.

303S. Chen et al. / NeuroImage 60 (2012) 299–304

network present in TLE patients (Vlooswijk et al., 2010). We founddecreased connectivity within the limbic system and medial temporalgyrus that might reflect disruption in the network due to the patho-logical epileptic condition, as reported in previous research on epilep-sy (Auer, 2008; Liao et al., 2010).

When comparing the PDS group with HCs, the network may re-flect both epileptic and depressive characteristics. We found in-creased connectivity within the superior frontal gyrus (Brodmannarea 8-B8) and decreased connectivity within Brodmann area 10(B10). B10 is involved in strategic processing for memory retrievaland executive functioning. The impairment in the connectivity inB10 may lead to cognitive impairment, which is common in epilepsypatients and patients with depressive symptoms (Monkul et al., 2011;Shelton et al., 2011). The B8 area is involved in management of uncer-tainty, which is critical in depressive conditions.

We also found increased connectivity between the limbic systemand the precentral gyrus in the nPDS group that may reflect the ongo-ing epileptic condition in the nPDS group. Research has shown motorsystem hyperconnectivity in juvenile myoclonic epilepsy, and the ef-fect is more robust in patients with ongoing epilepsy compared withhealthy controls (Vollmar et al., 2011). We also observed increasedconnectivity between the limbic system and superior temporalgyrus. Similarly, Liao et al. (2010) reported increased connectivity be-tween the temporal lobe and other areas in TLE patients. These latterdata suggest elevated synchronization between the temporal lobeand the default-mode network that may indicate a heightened mon-itoring of internal thoughts and feelings (Mason et al., 2007).

Our finding of increased connectivity between the ACC and limbicsystem in epilepsy patients has not previously been reported, al-though one publication noted increased connectivity in the PCC(Zhang et al., 2010). Our finding may be due to compensatory actionin the ACC due to the disruption in the prefrontal lobe-limbic systemconnection (Holmes and Pizzagalli, 2008). However, it may also bedue to the limited number of subjects. Taken together, the findingson connectivity alterations reported here may indicate an excitato-ry–inhibitory imbalance in brain activity and an epileptic network al-teration in epilepsy patients (Waites et al., 2006).

There were two issues that should be considered when interpret-ing our results: first, the relatively small number of participants maylimit the statistical power, and an enlargement of the database maysolve this problem; second, the experiment cannot per se providelongitudinal alteration data on the participants because our studywas cross sectional. We cannot determine whether the characteristicalteration in the PDS group predicts the poor treatment outcomesthat were the concern of Kanner (2008), although a retrospectiveanalysis in patients who received surgical operations revealed thatthe Beck depression index predicted postoperative seizure frequencyand seizure freedom (Metternich et al., 2009). In the future, longitu-dinal research may answer these questions and provide evidence onthe natural history of these two conditions (Hermann et al., 2008).

Conclusion

Taken together, the current results demonstrate that treatment-naïve TLE patients show hyperactivity in the prefrontal-limbic andstriatal brain areas and in the ACC with decreased functional connec-tivity within the prefrontal-limbic system in the PDS group. Longitu-dinal studies of these patients may illustrate whether the alteration inthe PDS group predicts treatment outcomes, and these studies mayhelp understand the lifetime disease course in PDS patients(Hermann et al., 2008).

Acknowledgments

This studywas supported by the National Basic Research Program(973 Program No. 2007CB512305) and National Natural Science

Foundation of China (Grant No. 81030027), and Dr. Qiyong Gong ac-knowledges the support from his CMB Distinguished ProfessorshipAward administered by the Institute of International Education, USA.

References

Anon., 1981. Proposal for revised clinical and electroencephalographic classification ofepileptic seizures. From the Commission on Classification and Terminology of theInternational League Against Epilepsy. Epilepsia 22, 489–501.

Anon., 1989. Proposal for revised classification of epilepsies and epileptic syndromes.Commission on Classification and Terminology of the International League AgainstEpilepsy. Epilepsia 30, 389–399.

Auer, D.P., 2008. Spontaneous low-frequency blood oxygenation level-dependent fluc-tuations and functional connectivity analysis of the ‘resting’ brain. Magn. Reson.Imaging 26, 1055–1064.

Bernhardt, B.C., Bernasconi, N., Concha, L., Bernasconi, A., 2010. Cortical thickness anal-ysis in temporal lobe epilepsy: reproducibility and relation to outcome. Neurology74, 1776–1784.

Bettus, G., Guedj, E., Joyeux, F., Confort-Gouny, S., Soulier, E., Laguitton, V., Cozzone, P.J.,Chauvel, P., Ranjeva, J.P., Bartolomei, F., Guye, M., 2009. Decreased basal fMRI func-tional connectivity in epileptogenic networks and contralateral compensatorymechanisms. Hum. Brain Mapp. 30, 1580–1591.

Bettus, G., Ranjeva, J.P., Wendling, F., Benar, C.G., Confort-Gouny, S., Regis, J., Chauvel,P., Cozzone, P.J., Lemieux, L., Bartolomei, F., Guye, M., 2011. Interictal functionalconnectivity of human epileptic networks assessed by intracerebral EEG andBOLD signal fluctuations. PLoS One 6, e20071.

Callicott, J.H., Mattay, V.S., Verchinski, B.A., Marenco, S., Egan, M.F., Weinberger, D.R.,2003. Complexity of prefrontal cortical dysfunction in schizophrenia: more thanup or down. Am. J. Psychiatry 160, 2209–2215.

Christensen, J., Vestergaard, M., Mortensen, P.B., Sidenius, P., Agerbo, E., 2007. Epilepsyand risk of suicide: a population-based case–control study. Lancet Neurol. 6,693–698.

Ettinger, A., Reed, M., Cramer, J., 2004. Depression and comorbidity in community-based patients with epilepsy or asthma. Neurology 63, 1008–1014.

Foland, L.C., Altshuler, L.L., Bookheimer, S.Y., Eisenberger, N., Townsend, J., Thompson,P.M., 2008. Evidence for deficient modulation of amygdala response by prefrontalcortex in bipolar mania. Psychiatry Res. 162, 27–37.

Ghiglieri, V., Sgobio, C., Patassini, S., Bagetta, V., Fejtova, A., Giampa, C., Marinucci, S.,Heyden, A., Gundelfinger, E.D., Fusco, F.R., Calabresi, P., Picconi, B., 2010. TrkB/BDNF-dependent striatal plasticity and behavior in a genetic model of epilepsy:modulation by valproic acid. Neuropsychopharmacology 35, 1531–1540.

Gilliam, F.G., Maton, B.M., Martin, R.C., Sawrie, S.M., Faught, R.E., Hugg, J.W., Viikinsalo,M., Kuzniecky, R.I., 2007. Hippocampal 1H-MRSI correlates with severity of depres-sion symptoms in temporal lobe epilepsy. Neurology 68, 364–368.

Goncalves, S.I., de Munck, J.C., Pouwels, P.J., Schoonhoven, R., Kuijer, J.P., Maurits, N.M.,Hoogduin, J.M., Van Someren, E.J., Heethaar, R.M., Lopes da Silva, F.H., 2006. Corre-lating the alpha rhythm to BOLD using simultaneous EEG/fMRI: inter-subject var-iability. Neuroimage 30, 203–213.

Hasler, G., Bonwetsch, R., Giovacchini, G., Toczek, M.T., Bagic, A., Luckenbaugh, D.A.,Drevets, W.C., Theodore, W.H., 2007. 5-HT1A receptor binding in temporal lobe ep-ilepsy patients with and without major depression. Biol. Psychiatry 62, 1258–1264.

Hermann, B., Seidenberg, M., Jones, J., 2008. The neurobehavioural comorbidities of ep-ilepsy: can a natural history be developed? Lancet Neurol. 7, 151–160.

Hesdorffer, D.C., Hauser, W.A., Olafsson, E., Ludvigsson, P., Kjartansson, O., 2006. De-pression and suicide attempt as risk factors for incident unprovoked seizures.Ann. Neurol. 59, 35–41.

Holmes, A.J., Pizzagalli, D.A., 2008. Spatiotemporal dynamics of error processing dys-functions in major depressive disorder. Arch. Gen. Psychiatry 65, 179–188.

Kanner, A., 2004. Do epilepsy and psychiatric disorders share common pathogenicmechanisms? A look at depression and epilepsy. Clin. Neurosci. Res. 4, 31–37.

Kanner, A.M., 2008. Depression in epilepsy: a complex relation with unexpected conse-quences. Curr. Opin. Neurol. 21, 190–194.

Kim, S., Zhang, S., Choi, K.H., Reister, R., Do, C., Baykiz, A.F., Gershenfeld, H.K., 2009. AnE3 ubiquitin ligase, really interesting new gene (RING) finger 41, is a candidategene for anxiety-like behavior and beta-carboline-induced seizures. Biol. Psychia-try 65, 425–431.

Kushnir, V., Menon, M., Balducci, X.L., Selby, P., Busto, U., Zawertailo, L., 2010. Enhancedsmoking cue salience associated with depression severity in nicotine-dependentindividuals: a preliminary fMRI study. Int. J. Neuropsychopharmacol. 1–12.

Liao, W., Zhang, Z., Pan, Z., Mantini, D., Ding, J., Duan, X., Luo, C., Lu, G., Chen, H., Valdes-Sosa, P.A., 2010. Altered functional connectivity and small-world in mesial tempo-ral lobe epilepsy. PLoS One 5, e8525.

Mason, M.F., Norton, M.I., Van Horn, J.D., Wegner, D.M., Grafton, S.T., Macrae, C.N.,2007. Wandering minds: the default network and stimulus-independent thought.Science 315, 393–395.

Mazarati, A., Shin, D., Auvin, S., Caplan, R., Sankar, R., 2007. Kindling epileptogenesis inimmature rats leads to persistent depressive behavior. Epilepsy Behav. 10,377–383.

Metternich, B., Wagner, K., Brandt, A., Kraemer, R., Buschmann, F., Zentner, J.,Schulze-Bonhage, A., 2009. Preoperative depressive symptoms predict postoper-ative seizure outcome in temporal and frontal lobe epilepsy. Epilepsy Behav. 16,622–628.

Monkul, E.S., Silva, L.A., Narayana, S., Peluso, M.A., Zamarripa, F., Nery, F.G., Najt, P., Li, J.,Lancaster, J.L., Fox, P.T., Lafer, B., Soares, J.C., 2011. Abnormal resting state corticolimbicblood flow in depressed unmedicated patients withmajor depression: a (15) O-H(2) O

304 S. Chen et al. / NeuroImage 60 (2012) 299–304

PET study. Hum. Brain Mapp. doi:10.1002/hbm.21212 (Electronic publication aheadof print).

Savic, I., Lindstrom, P., Gulyas, B., Halldin, C., Andree, B., Farde, L., 2004. Limbic reduc-tions of 5-HT1A receptor binding in human temporal lobe epilepsy. Neurology62, 1343–1351.

Shelton, R.C., Claiborne, J., Sidoryk-Wegrzynowicz, M., Reddy, R., Aschner, M., Lewis,D.A., Mirnics, K., 2011. Altered expression of genes involved in inflammationand apoptosis in frontal cortex in major depression. Mol. Psychiatry 16,751–762.

Shmuel, A., Leopold, D.A., 2008. Neuronal correlates of spontaneous fluctuations infMRI signals in monkey visual cortex: Implications for functional connectivity atrest. Hum. Brain Mapp. 29, 751–761.

Tellez-Zenteno, J.F., Patten, S.B., Jette, N., Williams, J., Wiebe, S., 2007. Psychiatric co-morbidity in epilepsy: a population-based analysis. Epilepsia 48, 2336–2344.

Vlooswijk, M.C., Jansen, J.F., de Krom, M.C., Majoie, H.M., Hofman, P.A., Backes, W.H.,Aldenkamp, A.P., 2010. Functional MRI in chronic epilepsy: associations with cog-nitive impairment. Lancet Neurol. 9 (10), 1018–1027.

Vollmar, C., O'Muircheartaigh, J., Barker, G.J., Symms, M.R., Thompson, P., Kumari, V.,Duncan, J.S., Janz, D., Richardson, M.P., Koepp, M.J., 2011. Motor system hypercon-nectivity in juvenile myoclonic epilepsy: a cognitive functional magnetic reso-nance imaging study. Brain 134, 1710–1719.

Waites, A.B., Briellmann, R.S., Saling, M.M., Abbott, D.F., Jackson, G.D., 2006. Functionalconnectivity networks are disrupted in left temporal lobe epilepsy. Ann. Neurol.59, 335–343.

Zhang, Z., Lu, G., Zhong, Y., Tan, Q., Liao, W., Wang, Z., Li, K., Chen, H., Liu, Y., 2010. Al-tered spontaneous neuronal activity of the default-mode network in mesial tem-poral lobe epilepsy. Brain Res. 1323, 152–160.

Zung, W.W., 1965. A self-rating depression scale. Arch. Gen. Psychiatry 12, 63–70.