Epilepsy Research (2014) 108, 853—860
journal homepage: www.elsevier.com/locate/epilepsyres
Frequency-dependent amplitude alterations of resting-state spontaneous fluctuations in idiopathic generalized epilepsy Zhengge Wang a,b,e, Zhiqiang Zhang a,e,∗, Wei Liao a,c, Qiang Xu a,e, Jie Zhang d,e, Wenlian Lu d,e, Qing Jiao a,e, Guanghui Chen f, Jianfeng Feng d,e, Guangming Lu a,e a
Department of Medical Imaging, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China b Department of Medical Imaging, Nanjing Drum Tower Hospital, the Affiliated Hospital of Nanjing University Medical School, Nanjing 210008, PR China c Center for Cognition and Brain Disorders and the Affiliated Hospital, Hangzhou Normal University, Hangzhou 310036, China d Center for Computational Systems Biology, Fudan University, Shanghai 200433, PR China e Jinling Hospital-Fudan University Computational Translational Medicine Center, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China f Department of Neurology, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China Received 14 July 2013; received in revised form 24 January 2014; accepted 17 March 2014 Available online 26 March 2014
KEYWORDS Idiopathic generalized epilepsy; Resting-state fMRI; Frequency; Amplitude; Spontaneous fluctuation
Summary Purpose: Amplitude of low-frequency fluctuation (ALFF) of blood-oxygenation level-dependent (BOLD) has proven a promising way to detect disease-related local brain activity. However, routine approach employs an arbitrary frequency band of 0.01—0.08 Hz, which lacks frequency specificity and blinds to the information contained in other frequency bands. This study investigated the amplitude of fluctuations in full BOLD frequency bands, and addressed how amplitudes of fluctuations change in each specific frequency range in idiopathic generalized epilepsy (IGE). Methods: Thirty-four IGE patients with generalized tonic—clonic seizure and the same number of age- and sex-matched healthy controls were included. Functional MRI data were acquired using a 2 s repetition time. Routine amplitude of low-frequency fluctuation analysis was first performed. The regions showing group difference were set as Region-of-interest for analysis of amplitudes of full-frequency. The amplitudes of BOLD fluctuations were consecutively performed at each
∗ Corresponding author at: Department of Medical Imaging, Nanjing Jinling Hospital, 305# Eastern Zhongshan Rd., Nanjing 210002, China. Tel.: +86 25 80860187. E-mail address: [email protected] (Z. Zhang).
http://dx.doi.org/10.1016/j.eplepsyres.2014.03.003 0920-1211/© 2014 Elsevier B.V. All rights reserved.
854
Z. Wang et al. frequency bin of 0.002 Hz, and specific frequency amplitude analyses were performed in five different frequency ranges (0—0.01 Hz, 0.01—0.027 Hz, 0.027—0.073 Hz, 0.073—0.198 Hz, and 0.198—0.25 Hz). Key findings: The thalamus and prefrontal cortex showed significant group differences in routine amplitude analysis. For amplitude of full-frequency analysis, a reverse pattern was found in the dynamic changes between the thalamus and prefrontal cortex in IGE. Moreover, the prefrontal cortex showed amplitude difference in the 0.01—0.027 Hz band, while the thalamus showed amplitude difference in the 0.027—0.073 Hz band. Both these two regions showed amplitude differences in 0.198—0.25 Hz band. Significance: We demonstrated the characteristic alterations of amplitude of BOLD fluctuations in IGE in frequency domain. The amplitude analysis of full frequency may potentially help to select specific frequency range for detecting epilepsy-related brain activity, and provide insights into the pathophysiological mechanism of IGE. © 2014 Elsevier B.V. All rights reserved.
Introduction Resting-state fMRI has recently proven a promising tool for interrogating human brain function. This technique is frequently used to map functional connectivity by measuring temporal coherence of spontaneous low frequency (0.01—0.08 Hz) blood-oxygenation level-dependent (BOLD) fluctuations. Moreover, amplitude of low-frequency fluctuation (ALFF) has been considered as another useful tool for depicting local brain activity. ALFF reflects brain activity level during a period of time, and may be somewhat similar to PET measurement (Zang et al., 2007; Zuo et al., 2010). Thus this technique has been used to detect brain activity abnormalities in a spectrum of brain diseases, such as mild cognitive impairment, attention deficit hyperactivity disorder and posttraumatic stress disorder (Han et al., 2011; Yin et al., 2011; Zang et al., 2007). Specifically, a recent fMRI study has applied ALFF to mesial temporal lobe epilepsy and found that increased ALFF related with epileptogenic foci, and decreased ALFF related with functional impairments in the default brain regions (Zhang et al., 2010). These findings suggested that ALFF can be a feasible way for epilepsy research. However, routine ALFF only considers BOLD fluctuations at low frequency band (0.01—0.08 Hz), which may lead to two limitations. Firstly, the arbitrary selection of frequency band causes information loss of the other frequency realms. It has been proposed that there is also physiological significance in lower- or higher-frequency BOLD signals (Baria et al., 2011; Niazy et al., 2011; Wu et al., 2008). Secondly, the band of 0.01—0.08 Hz covers a rather broader realm of spontaneous fluctuations, and would mingle any physiological fluctuations with potentially specific frequencies. Recently, Zuo et al. (2010) have addressed the spatial patterns of ALFF distributions at four different frequency bands. Interestingly, they found that stronger ALFF at lower frequency band (0.01—0.027 Hz) tended to be present in the cortical structures, and that at higher frequency (0.027—0.073 Hz) tended to be present in the subcortical structures. Thus, in relative to the routine ALFF, it is more rational to depict brain activity by measuring amplitude of spontaneous fluctuations at specific frequencies, and also in broader frequency bands.
Idiopathic generalized epilepsy (IGE) is a group of gene-related seizure syndromes charactering by widespread generalized discharges in electroencephalography (EEG) and normal routine MRI examinations. Generalized tonic—clonic seizures (GTCS) (Mattson, 2003) is the most frequent and severest type of IGE, clinically presents with rigid stiffening of the limbs, followed by violent bilateral spasms, and loss of consciousness. Evidence from neuroimaging studies has implicated that both the cortical and subcortical structures are involved in IGE (Bernhardt et al., 2009; Pulsipher et al., 2011). In addition to generalized-linear-model-based simultaneous EEG and fMRI studies (Gotman et al., 2005; Moeller et al., 2008), a battery of resting-state fMRI studies have depicted abnormalities in the regional and connectivity properties in the regions associating thalamocortical circuits in IGE (Luo et al., 2011; Wang et al., 2011, 2012; Zhong et al., 2011). As comment above, however, there still lacks information about the spontaneous brain activity in IGE measured in different specific frequencies. In this study, we first performed routine ALFF analyses in the low frequency band within 0.01—0.08 Hz, and several specific frequency bands of 0—0.01 Hz, 0.01—0.027 Hz, 0.027—0.073 Hz, 0.073—0.198 Hz, and 0.198—0.25 Hz for patients with IGE-GTCS. Specifically, we observed the amplitude of BOLD fluctuations with a full frequency realm from 0 to 0.25 Hz in IGE, and address the characteristic alteration of amplitudes along with the evolution of frequencies. This work provided an overall description about alterations of amplitudes of BOLD fluctuations in frequency domain, and might be potentially helpful for selection of specific frequency for measuring abnormal brain activity in epilepsy.
Methods Subjects Thirty-four patients with idiopathic generalized epilepsy characterized by generalized tonic—clonic seizures (IGEGTCS) [age (mean ± SD): 23.44 ± 4.96 years, age of seizure onset: 17.01 ± 6.19 years, duration: 6.42 ± 5.49 years] were recruited in this study. They were all right handed. All patients were diagnosed as IGE with only GTCS according to the criteria of international league against epilepsy
Frequency-dependent amplitude alterations in IGE Table 1 Demographic and clinical information data of the subjects. Characteristics
Gender (male/female) Age (year) Handedness (right/left) Seizure onset age Duration (year)
GTCS (n = 34)
HC (n = 34)
Mean ± SD
Mean ± SD
21/13 23.44 ± 4.96 34/0 17.01 ± 6.19 6.42 ± 5.49
19/15 24.08 ± 4.48 34/0 \ \
(ILAE) classification: (1) with typical clinical symptoms of generalized tonic—clonic seizures, including tonic extension of the limbs, followed by a clonic phase of rhythmic jerking of the extremities, loss of consciousness during seizures; (2) presence of generalized polyspike-wave in their interictal scalp EEG; and (3) no focal abnormality in the structural MRI. All patients were diagnosed as IGE with only GTCS according to the international league against epilepsy (ILAE) classification. All demographic and clinical information were detailed in Table 1. Age- and sex-matched right-handed healthy subjects (n = 34, age (mean ± SD): 24.08 ± 4.48 years) were recruited as controls. There was no significant difference in age and gender between the two groups. None of them had history of neurological or psychiatric disorder. The study was approved by the local medical ethics committee at Jinling Hospital, Nanjing University School of Medicine.
Data acquisition All data were collected on a 3 T Siemens Trio scanner with an eight-channel phased array head coil. Resting-state fMRI data were acquired using gradient echo type echo planar imaging (GRE-EPI) sequence with following parameters: TR/TE = 2000 ms/30 ms, FA = 90◦ , matrix = 64 × 64, FOV = 24 × 24 cm2 , and slice thickness = 4 mm, slice gap = 0.4 mm. A total of 30 slices were used to cover the whole brain. Each section contained 250 volumes. All patients and healthy subjects were instructed to relax, hold still, keep their eyes closed but stay awake during the resting-state functional MRI examination. All patients were absence of seizures symptoms, and were in interictal state during MRI scanning. Routine MRI examination images were also collected for excluding anatomic abnormality, such as T1, T2, DWI and T2-FLAIR images.
Image preprocessing Resting-state fMRI data were preprocessed using the toolkit of SPM8 (http://www.fil.ion.ucl.ac.uk/spm). The first 10 volumes of each participant were discarded. The remaining 240 scans were slice-time corrected and then realigned to the first volume to correct head motions. There is no data whose motion exceeded 1.5 mm or rotation exceeded 1◦ . All realigned images were spatially normalized to the MNI template and resampled to the voxel size of 3 mm × 3 mm × 3 mm. At last, the data were smoothed with an isotropic 8 mm FWHM (full width at half maximum) and
855 the systematic drift of the baseline signal was removed. No temporal filtering was implemented during preprocessing, which enable the signal analyses within full frequency realm (0—0.25 Hz).
Routine ALFF analysis Routine ALFF was calculated using REST software (http://www.restfmri.net/forum/REST V1.7). Briefly, for a given voxel, the time course was first converted to the frequency domain by using fast Fourier transform and the power spectrum was then obtained. For routine ALFF analysis, the square root of the power was computed and then averaged across 0.01—0.08 Hz. This averaged square root was taken as the ALFF (Zang et al., 2007). Moreover, we used a fractional ALFF for effectively suppressing the physiological noises. Fractional ALFF was calculated by dividing the ALFF in a given frequency band to the ALFF over the entire frequency range (Zou et al., 2008).
Amplitude of BOLD fluctuations in full frequency realm The full frequency range (0—0.25 Hz, the upper bound was restricted by the repetition time of the BOLD acquisition) of BOLD fluctuation was divided into 120 (this number was defined as the used volumes of each fMRI session) narrow band bins, i.e., each narrow frequency band covered 0.002 Hz. In line with the routine ALFF analysis, the amplitude of each narrow frequency band fluctuation was calculated. A total of 120 amplitude maps with narrow frequency band bin were produced in each subject. Secondly, according to the suggestion by the previous studies, we calculated amplitude in five specific frequency bands: slow-6 (0—0.01 Hz), slow-5 (0.01—0.027 Hz), slow4 (0.027—0.073 Hz), slow-3 (0.073—0.198 Hz), and slow-2 (0.198—0.25 Hz) (Han et al., 2011, 2012; Zuo et al., 2010).
Group analyses We first performed comparison of routine ALFF (i.e., 0.01—0.08 Hz) values between patients and controls using voxel-wise two sample t-test. The brain regions showing increased (i.e., the thalamus) and decreased ALFF (i.e., the prefrontal cortex) in the patients in comparison results were set as regions-of-interest (ROIs) for the subsequent full frequency amplitude analyses (p < 0.05 by using AlphaSim [combined p < 0.01 and voxel size >79]). The amplitude values of each ROI were extracted from 120 amplitude maps with narrow frequency band. Two-sample t-tests group comparisons were performed in each narrow frequency band (p < 0.05, Bonferroni corrections). Moreover, consistent with previous studies, we performed group comparisons of amplitude values in five specific frequency bands mentioned-above. Moreover, we produced maps to demonstrate the brain topology of maximal group difference of amplitude over all frequency bands, and specific frequency range corresponding to the maximal group difference of amplitude. We used spatial smooth
856
Z. Wang et al.
Figure 1 ALFF (0.01—0.08 Hz) group difference between patients and healthy controls. Compared to the controls, the GTCS patients showed increased ALFF in warm colored brain regions including anterior cingulated cortex, bilateral thalamus, putamen and cerebellum, whereas decreased ALFF in cool colored regions including mesial prefrontal cortex, bilateral dorsolateral prefrontal cortex, and bilateral orbital frontal cortex. Map threshold of multiple comparisons were set at p < 0.05 using AlphaSim correction. The mesial prefrontal cortex and bilateral thalamus were selected as ROIs.
(averaging the values of a voxel and its neighboring 27 voxels) and frequency smooth (averaging the values of a frequency band and two neighboring frequency bands) for map producing. These maps were anticipated to give a panorama showing maximally altered amplitude (MAA) in patients across wide spatial and frequencies realms. In addition, correlation analyses between clinical variables (including epilepsy durations, seizure frequency) and amplitudes of fluctuations in routine ALFF analysis and specific frequencies analysis were performed (FDR correction, p < 0.05).
Results Routine ALFF analysis We first report ALFF results from the routine frequency band (0.01—0.08 Hz) (Fig. 1). Compared with healthy controls using two-sample t-test, the patients exhibited increased ALFF in anterior cingulated cortex, bilateral thalamus, putamen and cerebellum, and decreased ALFF in medial prefrontal cortex (MPFC), bilateral orbit frontal cortex, and bilateral dorsolateral frontal cortex. Here the MPFC and bilateral thalamus were selected as ROIs for full frequency band analyses.
Full and specific frequency bands analyses ROI-based analyses demonstrated dynamic alteration of amplitude of fluctuations in the thalamus and medial prefrontal cortex along with frequency adaption (Fig. 2). Overall, in IGE patients, the thalamus showed increased amplitude in the lower frequency realm (0.08 Hz). Conversely, the MPFC showed decreased amplitude in the lower frequency realm (0.08 Hz). These two regions shared a transition point at about 0.085 Hz for the alteration of amplitude. Five specific frequency bands analyses showed spatial patterns of altered amplitudes along with frequency transition from low to high realm. In line with the above results, the thalamus showed significant increase in amplitude at slow-4 (0.027—0.073 Hz), and decrease at slow-2 (0.198—0.25 Hz). Conversely, the MPFC showed decrease in amplitude at slow-5 (0.01—0.027 Hz), and increase at slow-2 (0.198—0.25 Hz). Voxel-based analyses demonstrated the topology of the maximally altered amplitude (MAA) over all frequency bands between patients and healthy controls. Specifically, the thalamus and the prefrontal cortex showed higher positive MAA. The frequency corresponding to the thalamus positive MAA was at lower realm, whereas the frequency corresponding to the prefrontal cortex was at higher realm. For negative MAA, the MPFC was at lower frequency realm, and the thalamus was at higher frequency realm. Furthermore, the amplitude alteration of every brain region and corresponding frequency band could be observed (Fig. 3). Voxel-wise correlation analysis was carried out between ALFF (within traditional low frequency range 0.01—0.08 Hz and five frequency ranges mentioned above) and clinical indices (seizure frequency and seizure duration). The anterior cingulated cortex showed significant positive correlation with seizure duration and seizure frequency in routine 0.01—0.08 Hz band, including slow-5 (0.01—0.027 Hz) and slow-4 (0.027—0.073 Hz) bands. The negative correlation
Frequency-dependent amplitude alterations in IGE
857
Figure 2 The amplitude spectrum changes (0—0.25 Hz) in mesial prefrontal cortex and thalamus. These curves indicated the amplitude trend across full frequency band in medial prefrontal cortex and bilateral thalamus. The red and blue lines represented GTCS patients and healthy controls respectively. The full frequency band (0—0.25 Hz) was divided into five bands. They were displayed with different color background [slow-6 (0—0.01 Hz, pink), slow-5 (0.01—0.027 Hz, yellow), slow-4 (0.027—0.073 Hz, green), slow-3 (0.073—0.198 Hz, blue), and slow-2 (0.198—0.25 Hz, violet)]. The brain maps in the middle row showed group differences of amplitude between the two groups in five frequency bands. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
with seizure frequency was present at posterior cingulated cortex. Interestingly, both positive and negative correlations were contrary in slow-3 (0.073—0.198 Hz) band. The slow6 (0—0.01 Hz) and slow-2 (0.198—0.25 Hz) bands showed no significant correlation (Fig. 4).
Discussion Summary of results The present study investigated the amplitude of spontaneous BOLD fluctuations in patients with IGE, and especially addressed the features of altered amplitude in
different frequency bands. Routine amplitude of lowfrequency (0.01—0.08 Hz) fluctuation analysis revealed increased ALFF in the thalamus and decreased ALFF in the prefrontal cortex in IGE-GTCS. Importantly, we demonstrated BOLD fluctuation amplitudes in the whole frequency bands (0—0.25 Hz), and found a reversed pattern of changes in frequency realm in the thalamus and prefrontal cortex in IGE. We also mapped the topology of the maximally altered amplitude over all frequency bands, and identified the corresponding frequency. These findings gave a full view of amplitudes of BOLD fluctuations in frequency domain, and were potentially helpful to the selection of specific frequency for detecting epilepsy related brain activity.
858
Z. Wang et al.
Figure 3 The topology of the maximally altered amplitude (MAA) over all frequency bands between patients and healthy controls. (A) ALFF difference between patients and healthy controls. (B) Maximal positive group difference of amplitude over all frequency bands. (C) Corresponding frequency realm of maximal positive difference of amplitude. (D) Maximal negative group difference of amplitude over all frequency bands. (E) Corresponding frequency realm of maximal negative difference of amplitude.
Different ALFF in cortical and subcortical structures in IGE-GTCS Using routine ALFF analysis, the patients showed increased ALFF mainly in the bilateral thalamus, anterior cingulated cortex and putamen, and decreased ALFF in the MPFC,
which is concordant with the findings revealed by previous studies using other fMRI approaches. Generalized epileptic discharges related activation in the thalamus along with deactivation in the default mode areas have been reported in a large body of literatures, suggesting seizure generation and the suspension of brain default function (Bai et al.,
Figure 4 Correlation analysis between amplitude of specific frequency and seizure duration and seizure frequency. The warm color represents that the brain regions showed positive correlation between amplitude of specific frequency and seizure duration and seizure frequency. The cool color represents the regions showed negative correlation.
Frequency-dependent amplitude alterations in IGE 2010; Gotman et al., 2005; Moeller et al., 2011). This study for the first time showed the amplitude change of restingstate BOLD fluctuations in IGE. In line with the previous studies (Zhang et al., 2010), we considered that increased ALFF in bilateral thalamus reflected higher level of brain activity correlating with epileptic activity in IGE; decreased ALFF in the frontal cortex including MPFC might implicate impairment of the brain default function in IGE (Gotman et al., 2005; Song et al., 2011; Wang et al., 2011). On the whole, our results support the concept of thalamocortical circuitry abnormalities as the underlying pathophysiological substrate of IGE, and suggest that ALFF-based resting-state fMRI could be an alternative approach for fMRI investigation on epilepsy.
Frequency dependent amplitude changes in IGE-GTCS In line with the previous studies (Baria et al., 2011; Zuo et al., 2010), we investigated band-specific changes of resting-state BOLD fluctuations. We further demonstrated the overall properties of amplitude changes in full frequency bands (0—0.25 Hz) in IGE. The thalamus and the MPFC showed robust changes across the five frequency bands. Specific amplitude changes of resting-state brain fluctuations at different frequency bands have been observed in the intact, e.g., personality traits (Wei et al., 2014) and the diseased brain, e.g., Parkinson’s disease (Esposito et al., 2013), schizophrenia (Yu et al., 2012) and amnestic mild cognitive impairment (Han et al., 2011). This strategy has been considered to have advantage by providing more overall information about changed BOLD signals in the patients. We further showed band-specific correlations with clinical variables (durations and seizure frequency) of amplitudes of BOLD fluctuations. The findings suggest that different frequency bands may be underlain by specific pathological significances. Interestingly, the thalamus and MPFC both showed reciprocally alternating activity over the whole frequency band. Compared with the HC, we found increased amplitude in the lower range of low frequency band (0.08 Hz) in the thalamus. Conversely, the patients showed decreased amplitude in the lower range of low frequency realm (0.08 Hz) in the MPFC. These two regions shared a transition point at about 0.085 Hz for the alteration of amplitude. Consensus has been reached that lower frequency fluctuations possess higher magnitude power, coordinates long-distance neural activity; while higher frequency fluctuations have lower magnitude power, and coordinates more local neural activity (Baria et al., 2011; Buzsaki, 2004). Thus the current findings may suggest that the IGE patients have reduced/increased coordinating capability of long-distance/local neural activity in the frontal lobe/thalamus. This assumption could also be supported by the previous findings that the frontal lobe showed decreased functional connectivity (Song et al., 2011; Wang et al., 2011), and the thalamus showed increased local coherence of resting-state brain activity in IGE (Zhong et al., 2011).
859 Noteworthy, we further observed the topology of the maximally altered amplitude over all frequency bands for the first time. Distinct with the earlier studies only focusing on amplitude alterations at slow-5 and slow-4 band (Han et al., 2011; Wei et al., 2014; Yu et al., 2012), our results showed amplitude alterations of all regions at full frequency bands. In contrast to the routine ALFF that employs an arbitrary frequency band of 0.01—0.08 Hz, this strategy can select the most specific frequency range corresponding to the maximal group difference of amplitude. Thus this approach is potentially helpful to select the optimal frequency range for detecting epilepsy-related amplitude changes of BOLD fluctuations.
Limitation There are several limitations in our study. Firstly, the AEDs may affect brain function and low-frequency oscillations, although the AEDs therapy information were collected, because of the variability of the different combinations across the study population, an analysis of contribution for each single AED on the ALFF was not feasible. Secondly, the interictal epileptic discharges may lead to the increased ALFF, however the data were acquired without simultaneous EEG, and we could not evaluate the possible effects of interictal epileptic discharges. In order to attain whole brain coverage, we used a TR = 2 s for our resting-state fMRI data. A faster TR can extend the frequency range.
Conclusion In conclusion, we find specific alterations of amplitude of BOLD fluctuations of the cortical and subcortical structures in GTCS patients. The full frequency amplitude analysis demonstrated frequency-dependent alterations and may potentially help to select specific frequency range for detecting epilepsy-related brain activity. The amplitude in higher range of low frequency band (0.198—0.25 Hz) can also be used for discriminating GTCS patients from healthy controls. In the near future, we expect to investigate the high frequency oscillations in IGE patients.
Source of funding Supported by National Natural Science Foundation of China (Grant nos. 81271553, 81201078, 61131003, 81171328, 81301198, and 81201155), Key Program (Grant nos. 10Z026, BWS11J063) and grants for Young Scholars in Jinling Hospital (Grant no. 2011060), and the China Postdoctoral Science Foundation (Grant no. 2013M532229).
Ethical approval We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
860
Conflict of interest None of the authors has any conflict of interest to disclose.
Acknowledgments We thank the patients and volunteers for participating in this study. We thank the anonymous reviewers for their constructive suggestions to improve this work. We thank Mingquan Li, Nanjing University School of Medicine, for collection of data.
References Bai, X., Vestal, M., Berman, R., Negishi, M., Spann, M., Vega, C., Desalvo, M., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2010. Dynamic time course of typical childhood absence seizures: EEG, behavior, and functional magnetic resonance imaging. J. Neurosci. 30, 5884—5893. Baria, A.T., Baliki, M.N., Parrish, T., Apkarian, A.V., 2011. Anatomical and functional assemblies of brain BOLD oscillations. J. Neurosci. 31, 7910—7919. Bernhardt, B.C., Rozen, D.A., Worsley, K.J., Evans, A.C., Bernasconi, N., Bernasconi, A., 2009. Thalamo-cortical network pathology in idiopathic generalized epilepsy: insights from MRI-based morphometric correlation analysis. NeuroImage 46, 373—381. Buzsaki, G., 2004. Neuronal oscillations in cortical networks. Science 304, 1926—1929. Esposito, F., Tessitore, A., Giordano, A., De Micco, R., Paccone, A., Conforti, R., Pignataro, G., Annunziato, L., Tedeschi, G., 2013. Rhythm-specific modulation of the sensorimotor network in drug-naive patients with Parkinson’s disease by levodopa. Brain 136, 710—725. Gotman, J., Grova, C., Bagshaw, A., Kobayashi, E., Aghakhani, Y., Dubeau, F., 2005. Generalized epileptic discharges show thalamocortical activation and suspension of the default state of the brain. Proc. Natl. Acad. Sci. U. S. A. 102, 15236—15240. Han, Y., Lui, S., Kuang, W., Lang, Q., Zou, L., Jia, J., 2012. Anatomical and functional deficits in patients with amnestic mild cognitive impairment. PLoS ONE 7, e28664. Han, Y., Wang, J., Zhao, Z., Min, B., Lu, J., Li, K., He, Y., Jia, J., 2011. Frequency-dependent changes in the amplitude of lowfrequency fluctuations in amnestic mild cognitive impairment: a resting-state fMRI study. Neuroimage 55, 287—295. Luo, C., Li, Q., Xia, Y., Lei, X., Xue, K., Yao, Z., Lai, Y., Marti Nez-Montes, E., Liao, W., Zhou, D., Valdes-Sosa, P.A., Gong, Q., Yao, D., 2011. Resting state basal ganglia network in idiopathic generalized epilepsy. Hum Brain Mapp. 36, 1279—1294. Mattson, R.H., 2003. Overview: idiopathic generalized epilepsies. Epilepsia 44 (Suppl. 2), 2—6. Moeller, F., Maneshi, M., Pittau, F., Gholipour, T., Bellec, P., Dubeau, F., Grova, C., Gotman, J., 2011. Functional connectivity in patients with idiopathic generalized epilepsy. Epilepsia 52, 515—522. Moeller, F., Siebner, H., Wolff, S., Muhle, H., Boor, R., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008. Changes in
Z. Wang et al. activity of striato-thalamo-cortical network precede generalized spike wave discharges. NeuroImage 39, 1839—1849. Niazy, R.K., Xie, J., Miller, K., Beckmann, C.F., Smith, S.M., 2011. Spectral characteristics of resting state networks. Prog. Brain Res. 193, 259—276. Pulsipher, D.T., Dabbs, K., Tuchsherer, V., Sheth, R.D., Koehn, M.A., Hermann, B.P., Seidenberg, M., 2011. Thalamofrontal neurodevelopment in new-onset pediatric idiopathic generalized epilepsy. Neurology 76, 28—33. Song, M., Du, H., Wu, N., Hou, B., Wu, G., Wang, J., Feng, H., Jiang, T., 2011. Impaired resting-state functional integrations within default mode network of generalized tonic—clonic seizures epilepsy. PLoS ONE 6, e17294. Wang, Z., Lu, G., Zhang, Z., Zhong, Y., Jiao, Q., Tan, Q., Tian, L., Chen, G., Liao, W., Li, K., Liu, Y., 2011. Altered resting state networks in epileptic patients with generalized tonic—clonic seizures. Brain Res. 1374, 134—141. Wang, Z., Zhang, Z., Jiao, Q., Liao, W., Chen, G., Sun, K., Shen, L., Wang, M., Li, K., Liu, Y., Lu, G., 2012. Impairments of thalamic nuclei in idiopathic generalized epilepsy revealed by a study combining morphological and functional connectivity MRI. PLoS ONE 7, e39701. Wei, L., Duan, X., Zheng, C., Wang, S., Gao, Q., Zhang, Z., Lu, G., Chen, H., 2014. Specific frequency bands of amplitude lowfrequency oscillation encodes personality. Hum Brain Mapp. 35, 331—339. Wu, C.W., Gu, H., Lu, H., Stein, E.A., Chen, J.-H., Yang, Y., 2008. Frequency specificity of functional connectivity in brain networks. NeuroImage 42, 1047—1055. Yin, Y., Li, L., Jin, C., Hu, X., Duan, L., Eyler, L.T., Gong, Q., Song, M., Jiang, T., Liao, M., Zhang, Y., Li, W., 2011. Abnormal baseline brain activity in posttraumatic stress disorder: a resting-state functional magnetic resonance imaging study. Neurosci. Lett. 498, 185—189. Yu, R., Chien, Y.L., Wang, H.L., Liu, C.M., Liu, C.C., Hwang, T.J., Hsieh, M.H., Hwu, H.G., Tseng, W.Y., 2012. Frequency-specific alternations in the amplitude of low-frequency fluctuations in schizophrenia. Hum Brain Mapp. 35, 627—637. Zang, Y.F., He, Y., Zhu, C.Z., Cao, Q.J., Sui, M.Q., Liang, M., Tian, L.X., Jiang, T.Z., Wang, Y.F., 2007. Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain Dev. 29, 83—91. Zhang, Z., Lu, G., Zhong, Y., Tan, Q., Chen, H., Liao, W., Tian, L., Li, Z., Shi, J., Liu, Y., 2010. fMRI study of mesial temporal lobe epilepsy using amplitude of low-frequency fluctuation analysis. Hum Brain Mapp. 31, 1851—1861. Zhong, Y., Lu, G., Zhang, Z., Jiao, Q., Li, K., Liu, Y., 2011. Altered regional synchronization in epileptic patients with generalized tonic—clonic seizures. Epilepsy Res. 97, 83—91. Zou, Q.H., Zhu, C.Z., Yang, Y., Zuo, X.N., Long, X.Y., Cao, Q.J., Wang, Y.F., Zang, Y.F., 2008. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J. Neurosci. Methods 172, 137—141. Zuo, X.-N., Di Martino, A., Kelly, C., Shehzad, Z.E., Gee, D.G., Klein, D.F., Castellanos, F.X., Biswal, B.B., Milham, M.P., 2010. The oscillating brain: complex and reliable. NeuroImage 49, 1432—1445.
journal homepage: www.elsevier.com/locate/epilepsyres
Frequency-dependent amplitude alterations of resting-state spontaneous fluctuations in idiopathic generalized epilepsy Zhengge Wang a,b,e, Zhiqiang Zhang a,e,∗, Wei Liao a,c, Qiang Xu a,e, Jie Zhang d,e, Wenlian Lu d,e, Qing Jiao a,e, Guanghui Chen f, Jianfeng Feng d,e, Guangming Lu a,e a
Department of Medical Imaging, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China b Department of Medical Imaging, Nanjing Drum Tower Hospital, the Affiliated Hospital of Nanjing University Medical School, Nanjing 210008, PR China c Center for Cognition and Brain Disorders and the Affiliated Hospital, Hangzhou Normal University, Hangzhou 310036, China d Center for Computational Systems Biology, Fudan University, Shanghai 200433, PR China e Jinling Hospital-Fudan University Computational Translational Medicine Center, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China f Department of Neurology, Jinling Hospital, Nanjing University School of Medicine, Nanjing 210002, PR China Received 14 July 2013; received in revised form 24 January 2014; accepted 17 March 2014 Available online 26 March 2014
KEYWORDS Idiopathic generalized epilepsy; Resting-state fMRI; Frequency; Amplitude; Spontaneous fluctuation
Summary Purpose: Amplitude of low-frequency fluctuation (ALFF) of blood-oxygenation level-dependent (BOLD) has proven a promising way to detect disease-related local brain activity. However, routine approach employs an arbitrary frequency band of 0.01—0.08 Hz, which lacks frequency specificity and blinds to the information contained in other frequency bands. This study investigated the amplitude of fluctuations in full BOLD frequency bands, and addressed how amplitudes of fluctuations change in each specific frequency range in idiopathic generalized epilepsy (IGE). Methods: Thirty-four IGE patients with generalized tonic—clonic seizure and the same number of age- and sex-matched healthy controls were included. Functional MRI data were acquired using a 2 s repetition time. Routine amplitude of low-frequency fluctuation analysis was first performed. The regions showing group difference were set as Region-of-interest for analysis of amplitudes of full-frequency. The amplitudes of BOLD fluctuations were consecutively performed at each
∗ Corresponding author at: Department of Medical Imaging, Nanjing Jinling Hospital, 305# Eastern Zhongshan Rd., Nanjing 210002, China. Tel.: +86 25 80860187. E-mail address: [email protected] (Z. Zhang).
http://dx.doi.org/10.1016/j.eplepsyres.2014.03.003 0920-1211/© 2014 Elsevier B.V. All rights reserved.
854
Z. Wang et al. frequency bin of 0.002 Hz, and specific frequency amplitude analyses were performed in five different frequency ranges (0—0.01 Hz, 0.01—0.027 Hz, 0.027—0.073 Hz, 0.073—0.198 Hz, and 0.198—0.25 Hz). Key findings: The thalamus and prefrontal cortex showed significant group differences in routine amplitude analysis. For amplitude of full-frequency analysis, a reverse pattern was found in the dynamic changes between the thalamus and prefrontal cortex in IGE. Moreover, the prefrontal cortex showed amplitude difference in the 0.01—0.027 Hz band, while the thalamus showed amplitude difference in the 0.027—0.073 Hz band. Both these two regions showed amplitude differences in 0.198—0.25 Hz band. Significance: We demonstrated the characteristic alterations of amplitude of BOLD fluctuations in IGE in frequency domain. The amplitude analysis of full frequency may potentially help to select specific frequency range for detecting epilepsy-related brain activity, and provide insights into the pathophysiological mechanism of IGE. © 2014 Elsevier B.V. All rights reserved.
Introduction Resting-state fMRI has recently proven a promising tool for interrogating human brain function. This technique is frequently used to map functional connectivity by measuring temporal coherence of spontaneous low frequency (0.01—0.08 Hz) blood-oxygenation level-dependent (BOLD) fluctuations. Moreover, amplitude of low-frequency fluctuation (ALFF) has been considered as another useful tool for depicting local brain activity. ALFF reflects brain activity level during a period of time, and may be somewhat similar to PET measurement (Zang et al., 2007; Zuo et al., 2010). Thus this technique has been used to detect brain activity abnormalities in a spectrum of brain diseases, such as mild cognitive impairment, attention deficit hyperactivity disorder and posttraumatic stress disorder (Han et al., 2011; Yin et al., 2011; Zang et al., 2007). Specifically, a recent fMRI study has applied ALFF to mesial temporal lobe epilepsy and found that increased ALFF related with epileptogenic foci, and decreased ALFF related with functional impairments in the default brain regions (Zhang et al., 2010). These findings suggested that ALFF can be a feasible way for epilepsy research. However, routine ALFF only considers BOLD fluctuations at low frequency band (0.01—0.08 Hz), which may lead to two limitations. Firstly, the arbitrary selection of frequency band causes information loss of the other frequency realms. It has been proposed that there is also physiological significance in lower- or higher-frequency BOLD signals (Baria et al., 2011; Niazy et al., 2011; Wu et al., 2008). Secondly, the band of 0.01—0.08 Hz covers a rather broader realm of spontaneous fluctuations, and would mingle any physiological fluctuations with potentially specific frequencies. Recently, Zuo et al. (2010) have addressed the spatial patterns of ALFF distributions at four different frequency bands. Interestingly, they found that stronger ALFF at lower frequency band (0.01—0.027 Hz) tended to be present in the cortical structures, and that at higher frequency (0.027—0.073 Hz) tended to be present in the subcortical structures. Thus, in relative to the routine ALFF, it is more rational to depict brain activity by measuring amplitude of spontaneous fluctuations at specific frequencies, and also in broader frequency bands.
Idiopathic generalized epilepsy (IGE) is a group of gene-related seizure syndromes charactering by widespread generalized discharges in electroencephalography (EEG) and normal routine MRI examinations. Generalized tonic—clonic seizures (GTCS) (Mattson, 2003) is the most frequent and severest type of IGE, clinically presents with rigid stiffening of the limbs, followed by violent bilateral spasms, and loss of consciousness. Evidence from neuroimaging studies has implicated that both the cortical and subcortical structures are involved in IGE (Bernhardt et al., 2009; Pulsipher et al., 2011). In addition to generalized-linear-model-based simultaneous EEG and fMRI studies (Gotman et al., 2005; Moeller et al., 2008), a battery of resting-state fMRI studies have depicted abnormalities in the regional and connectivity properties in the regions associating thalamocortical circuits in IGE (Luo et al., 2011; Wang et al., 2011, 2012; Zhong et al., 2011). As comment above, however, there still lacks information about the spontaneous brain activity in IGE measured in different specific frequencies. In this study, we first performed routine ALFF analyses in the low frequency band within 0.01—0.08 Hz, and several specific frequency bands of 0—0.01 Hz, 0.01—0.027 Hz, 0.027—0.073 Hz, 0.073—0.198 Hz, and 0.198—0.25 Hz for patients with IGE-GTCS. Specifically, we observed the amplitude of BOLD fluctuations with a full frequency realm from 0 to 0.25 Hz in IGE, and address the characteristic alteration of amplitudes along with the evolution of frequencies. This work provided an overall description about alterations of amplitudes of BOLD fluctuations in frequency domain, and might be potentially helpful for selection of specific frequency for measuring abnormal brain activity in epilepsy.
Methods Subjects Thirty-four patients with idiopathic generalized epilepsy characterized by generalized tonic—clonic seizures (IGEGTCS) [age (mean ± SD): 23.44 ± 4.96 years, age of seizure onset: 17.01 ± 6.19 years, duration: 6.42 ± 5.49 years] were recruited in this study. They were all right handed. All patients were diagnosed as IGE with only GTCS according to the criteria of international league against epilepsy
Frequency-dependent amplitude alterations in IGE Table 1 Demographic and clinical information data of the subjects. Characteristics
Gender (male/female) Age (year) Handedness (right/left) Seizure onset age Duration (year)
GTCS (n = 34)
HC (n = 34)
Mean ± SD
Mean ± SD
21/13 23.44 ± 4.96 34/0 17.01 ± 6.19 6.42 ± 5.49
19/15 24.08 ± 4.48 34/0 \ \
(ILAE) classification: (1) with typical clinical symptoms of generalized tonic—clonic seizures, including tonic extension of the limbs, followed by a clonic phase of rhythmic jerking of the extremities, loss of consciousness during seizures; (2) presence of generalized polyspike-wave in their interictal scalp EEG; and (3) no focal abnormality in the structural MRI. All patients were diagnosed as IGE with only GTCS according to the international league against epilepsy (ILAE) classification. All demographic and clinical information were detailed in Table 1. Age- and sex-matched right-handed healthy subjects (n = 34, age (mean ± SD): 24.08 ± 4.48 years) were recruited as controls. There was no significant difference in age and gender between the two groups. None of them had history of neurological or psychiatric disorder. The study was approved by the local medical ethics committee at Jinling Hospital, Nanjing University School of Medicine.
Data acquisition All data were collected on a 3 T Siemens Trio scanner with an eight-channel phased array head coil. Resting-state fMRI data were acquired using gradient echo type echo planar imaging (GRE-EPI) sequence with following parameters: TR/TE = 2000 ms/30 ms, FA = 90◦ , matrix = 64 × 64, FOV = 24 × 24 cm2 , and slice thickness = 4 mm, slice gap = 0.4 mm. A total of 30 slices were used to cover the whole brain. Each section contained 250 volumes. All patients and healthy subjects were instructed to relax, hold still, keep their eyes closed but stay awake during the resting-state functional MRI examination. All patients were absence of seizures symptoms, and were in interictal state during MRI scanning. Routine MRI examination images were also collected for excluding anatomic abnormality, such as T1, T2, DWI and T2-FLAIR images.
Image preprocessing Resting-state fMRI data were preprocessed using the toolkit of SPM8 (http://www.fil.ion.ucl.ac.uk/spm). The first 10 volumes of each participant were discarded. The remaining 240 scans were slice-time corrected and then realigned to the first volume to correct head motions. There is no data whose motion exceeded 1.5 mm or rotation exceeded 1◦ . All realigned images were spatially normalized to the MNI template and resampled to the voxel size of 3 mm × 3 mm × 3 mm. At last, the data were smoothed with an isotropic 8 mm FWHM (full width at half maximum) and
855 the systematic drift of the baseline signal was removed. No temporal filtering was implemented during preprocessing, which enable the signal analyses within full frequency realm (0—0.25 Hz).
Routine ALFF analysis Routine ALFF was calculated using REST software (http://www.restfmri.net/forum/REST V1.7). Briefly, for a given voxel, the time course was first converted to the frequency domain by using fast Fourier transform and the power spectrum was then obtained. For routine ALFF analysis, the square root of the power was computed and then averaged across 0.01—0.08 Hz. This averaged square root was taken as the ALFF (Zang et al., 2007). Moreover, we used a fractional ALFF for effectively suppressing the physiological noises. Fractional ALFF was calculated by dividing the ALFF in a given frequency band to the ALFF over the entire frequency range (Zou et al., 2008).
Amplitude of BOLD fluctuations in full frequency realm The full frequency range (0—0.25 Hz, the upper bound was restricted by the repetition time of the BOLD acquisition) of BOLD fluctuation was divided into 120 (this number was defined as the used volumes of each fMRI session) narrow band bins, i.e., each narrow frequency band covered 0.002 Hz. In line with the routine ALFF analysis, the amplitude of each narrow frequency band fluctuation was calculated. A total of 120 amplitude maps with narrow frequency band bin were produced in each subject. Secondly, according to the suggestion by the previous studies, we calculated amplitude in five specific frequency bands: slow-6 (0—0.01 Hz), slow-5 (0.01—0.027 Hz), slow4 (0.027—0.073 Hz), slow-3 (0.073—0.198 Hz), and slow-2 (0.198—0.25 Hz) (Han et al., 2011, 2012; Zuo et al., 2010).
Group analyses We first performed comparison of routine ALFF (i.e., 0.01—0.08 Hz) values between patients and controls using voxel-wise two sample t-test. The brain regions showing increased (i.e., the thalamus) and decreased ALFF (i.e., the prefrontal cortex) in the patients in comparison results were set as regions-of-interest (ROIs) for the subsequent full frequency amplitude analyses (p < 0.05 by using AlphaSim [combined p < 0.01 and voxel size >79]). The amplitude values of each ROI were extracted from 120 amplitude maps with narrow frequency band. Two-sample t-tests group comparisons were performed in each narrow frequency band (p < 0.05, Bonferroni corrections). Moreover, consistent with previous studies, we performed group comparisons of amplitude values in five specific frequency bands mentioned-above. Moreover, we produced maps to demonstrate the brain topology of maximal group difference of amplitude over all frequency bands, and specific frequency range corresponding to the maximal group difference of amplitude. We used spatial smooth
856
Z. Wang et al.
Figure 1 ALFF (0.01—0.08 Hz) group difference between patients and healthy controls. Compared to the controls, the GTCS patients showed increased ALFF in warm colored brain regions including anterior cingulated cortex, bilateral thalamus, putamen and cerebellum, whereas decreased ALFF in cool colored regions including mesial prefrontal cortex, bilateral dorsolateral prefrontal cortex, and bilateral orbital frontal cortex. Map threshold of multiple comparisons were set at p < 0.05 using AlphaSim correction. The mesial prefrontal cortex and bilateral thalamus were selected as ROIs.
(averaging the values of a voxel and its neighboring 27 voxels) and frequency smooth (averaging the values of a frequency band and two neighboring frequency bands) for map producing. These maps were anticipated to give a panorama showing maximally altered amplitude (MAA) in patients across wide spatial and frequencies realms. In addition, correlation analyses between clinical variables (including epilepsy durations, seizure frequency) and amplitudes of fluctuations in routine ALFF analysis and specific frequencies analysis were performed (FDR correction, p < 0.05).
Results Routine ALFF analysis We first report ALFF results from the routine frequency band (0.01—0.08 Hz) (Fig. 1). Compared with healthy controls using two-sample t-test, the patients exhibited increased ALFF in anterior cingulated cortex, bilateral thalamus, putamen and cerebellum, and decreased ALFF in medial prefrontal cortex (MPFC), bilateral orbit frontal cortex, and bilateral dorsolateral frontal cortex. Here the MPFC and bilateral thalamus were selected as ROIs for full frequency band analyses.
Full and specific frequency bands analyses ROI-based analyses demonstrated dynamic alteration of amplitude of fluctuations in the thalamus and medial prefrontal cortex along with frequency adaption (Fig. 2). Overall, in IGE patients, the thalamus showed increased amplitude in the lower frequency realm (0.08 Hz). Conversely, the MPFC showed decreased amplitude in the lower frequency realm (0.08 Hz). These two regions shared a transition point at about 0.085 Hz for the alteration of amplitude. Five specific frequency bands analyses showed spatial patterns of altered amplitudes along with frequency transition from low to high realm. In line with the above results, the thalamus showed significant increase in amplitude at slow-4 (0.027—0.073 Hz), and decrease at slow-2 (0.198—0.25 Hz). Conversely, the MPFC showed decrease in amplitude at slow-5 (0.01—0.027 Hz), and increase at slow-2 (0.198—0.25 Hz). Voxel-based analyses demonstrated the topology of the maximally altered amplitude (MAA) over all frequency bands between patients and healthy controls. Specifically, the thalamus and the prefrontal cortex showed higher positive MAA. The frequency corresponding to the thalamus positive MAA was at lower realm, whereas the frequency corresponding to the prefrontal cortex was at higher realm. For negative MAA, the MPFC was at lower frequency realm, and the thalamus was at higher frequency realm. Furthermore, the amplitude alteration of every brain region and corresponding frequency band could be observed (Fig. 3). Voxel-wise correlation analysis was carried out between ALFF (within traditional low frequency range 0.01—0.08 Hz and five frequency ranges mentioned above) and clinical indices (seizure frequency and seizure duration). The anterior cingulated cortex showed significant positive correlation with seizure duration and seizure frequency in routine 0.01—0.08 Hz band, including slow-5 (0.01—0.027 Hz) and slow-4 (0.027—0.073 Hz) bands. The negative correlation
Frequency-dependent amplitude alterations in IGE
857
Figure 2 The amplitude spectrum changes (0—0.25 Hz) in mesial prefrontal cortex and thalamus. These curves indicated the amplitude trend across full frequency band in medial prefrontal cortex and bilateral thalamus. The red and blue lines represented GTCS patients and healthy controls respectively. The full frequency band (0—0.25 Hz) was divided into five bands. They were displayed with different color background [slow-6 (0—0.01 Hz, pink), slow-5 (0.01—0.027 Hz, yellow), slow-4 (0.027—0.073 Hz, green), slow-3 (0.073—0.198 Hz, blue), and slow-2 (0.198—0.25 Hz, violet)]. The brain maps in the middle row showed group differences of amplitude between the two groups in five frequency bands. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
with seizure frequency was present at posterior cingulated cortex. Interestingly, both positive and negative correlations were contrary in slow-3 (0.073—0.198 Hz) band. The slow6 (0—0.01 Hz) and slow-2 (0.198—0.25 Hz) bands showed no significant correlation (Fig. 4).
Discussion Summary of results The present study investigated the amplitude of spontaneous BOLD fluctuations in patients with IGE, and especially addressed the features of altered amplitude in
different frequency bands. Routine amplitude of lowfrequency (0.01—0.08 Hz) fluctuation analysis revealed increased ALFF in the thalamus and decreased ALFF in the prefrontal cortex in IGE-GTCS. Importantly, we demonstrated BOLD fluctuation amplitudes in the whole frequency bands (0—0.25 Hz), and found a reversed pattern of changes in frequency realm in the thalamus and prefrontal cortex in IGE. We also mapped the topology of the maximally altered amplitude over all frequency bands, and identified the corresponding frequency. These findings gave a full view of amplitudes of BOLD fluctuations in frequency domain, and were potentially helpful to the selection of specific frequency for detecting epilepsy related brain activity.
858
Z. Wang et al.
Figure 3 The topology of the maximally altered amplitude (MAA) over all frequency bands between patients and healthy controls. (A) ALFF difference between patients and healthy controls. (B) Maximal positive group difference of amplitude over all frequency bands. (C) Corresponding frequency realm of maximal positive difference of amplitude. (D) Maximal negative group difference of amplitude over all frequency bands. (E) Corresponding frequency realm of maximal negative difference of amplitude.
Different ALFF in cortical and subcortical structures in IGE-GTCS Using routine ALFF analysis, the patients showed increased ALFF mainly in the bilateral thalamus, anterior cingulated cortex and putamen, and decreased ALFF in the MPFC,
which is concordant with the findings revealed by previous studies using other fMRI approaches. Generalized epileptic discharges related activation in the thalamus along with deactivation in the default mode areas have been reported in a large body of literatures, suggesting seizure generation and the suspension of brain default function (Bai et al.,
Figure 4 Correlation analysis between amplitude of specific frequency and seizure duration and seizure frequency. The warm color represents that the brain regions showed positive correlation between amplitude of specific frequency and seizure duration and seizure frequency. The cool color represents the regions showed negative correlation.
Frequency-dependent amplitude alterations in IGE 2010; Gotman et al., 2005; Moeller et al., 2011). This study for the first time showed the amplitude change of restingstate BOLD fluctuations in IGE. In line with the previous studies (Zhang et al., 2010), we considered that increased ALFF in bilateral thalamus reflected higher level of brain activity correlating with epileptic activity in IGE; decreased ALFF in the frontal cortex including MPFC might implicate impairment of the brain default function in IGE (Gotman et al., 2005; Song et al., 2011; Wang et al., 2011). On the whole, our results support the concept of thalamocortical circuitry abnormalities as the underlying pathophysiological substrate of IGE, and suggest that ALFF-based resting-state fMRI could be an alternative approach for fMRI investigation on epilepsy.
Frequency dependent amplitude changes in IGE-GTCS In line with the previous studies (Baria et al., 2011; Zuo et al., 2010), we investigated band-specific changes of resting-state BOLD fluctuations. We further demonstrated the overall properties of amplitude changes in full frequency bands (0—0.25 Hz) in IGE. The thalamus and the MPFC showed robust changes across the five frequency bands. Specific amplitude changes of resting-state brain fluctuations at different frequency bands have been observed in the intact, e.g., personality traits (Wei et al., 2014) and the diseased brain, e.g., Parkinson’s disease (Esposito et al., 2013), schizophrenia (Yu et al., 2012) and amnestic mild cognitive impairment (Han et al., 2011). This strategy has been considered to have advantage by providing more overall information about changed BOLD signals in the patients. We further showed band-specific correlations with clinical variables (durations and seizure frequency) of amplitudes of BOLD fluctuations. The findings suggest that different frequency bands may be underlain by specific pathological significances. Interestingly, the thalamus and MPFC both showed reciprocally alternating activity over the whole frequency band. Compared with the HC, we found increased amplitude in the lower range of low frequency band (0.08 Hz) in the thalamus. Conversely, the patients showed decreased amplitude in the lower range of low frequency realm (0.08 Hz) in the MPFC. These two regions shared a transition point at about 0.085 Hz for the alteration of amplitude. Consensus has been reached that lower frequency fluctuations possess higher magnitude power, coordinates long-distance neural activity; while higher frequency fluctuations have lower magnitude power, and coordinates more local neural activity (Baria et al., 2011; Buzsaki, 2004). Thus the current findings may suggest that the IGE patients have reduced/increased coordinating capability of long-distance/local neural activity in the frontal lobe/thalamus. This assumption could also be supported by the previous findings that the frontal lobe showed decreased functional connectivity (Song et al., 2011; Wang et al., 2011), and the thalamus showed increased local coherence of resting-state brain activity in IGE (Zhong et al., 2011).
859 Noteworthy, we further observed the topology of the maximally altered amplitude over all frequency bands for the first time. Distinct with the earlier studies only focusing on amplitude alterations at slow-5 and slow-4 band (Han et al., 2011; Wei et al., 2014; Yu et al., 2012), our results showed amplitude alterations of all regions at full frequency bands. In contrast to the routine ALFF that employs an arbitrary frequency band of 0.01—0.08 Hz, this strategy can select the most specific frequency range corresponding to the maximal group difference of amplitude. Thus this approach is potentially helpful to select the optimal frequency range for detecting epilepsy-related amplitude changes of BOLD fluctuations.
Limitation There are several limitations in our study. Firstly, the AEDs may affect brain function and low-frequency oscillations, although the AEDs therapy information were collected, because of the variability of the different combinations across the study population, an analysis of contribution for each single AED on the ALFF was not feasible. Secondly, the interictal epileptic discharges may lead to the increased ALFF, however the data were acquired without simultaneous EEG, and we could not evaluate the possible effects of interictal epileptic discharges. In order to attain whole brain coverage, we used a TR = 2 s for our resting-state fMRI data. A faster TR can extend the frequency range.
Conclusion In conclusion, we find specific alterations of amplitude of BOLD fluctuations of the cortical and subcortical structures in GTCS patients. The full frequency amplitude analysis demonstrated frequency-dependent alterations and may potentially help to select specific frequency range for detecting epilepsy-related brain activity. The amplitude in higher range of low frequency band (0.198—0.25 Hz) can also be used for discriminating GTCS patients from healthy controls. In the near future, we expect to investigate the high frequency oscillations in IGE patients.
Source of funding Supported by National Natural Science Foundation of China (Grant nos. 81271553, 81201078, 61131003, 81171328, 81301198, and 81201155), Key Program (Grant nos. 10Z026, BWS11J063) and grants for Young Scholars in Jinling Hospital (Grant no. 2011060), and the China Postdoctoral Science Foundation (Grant no. 2013M532229).
Ethical approval We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
860
Conflict of interest None of the authors has any conflict of interest to disclose.
Acknowledgments We thank the patients and volunteers for participating in this study. We thank the anonymous reviewers for their constructive suggestions to improve this work. We thank Mingquan Li, Nanjing University School of Medicine, for collection of data.
References Bai, X., Vestal, M., Berman, R., Negishi, M., Spann, M., Vega, C., Desalvo, M., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2010. Dynamic time course of typical childhood absence seizures: EEG, behavior, and functional magnetic resonance imaging. J. Neurosci. 30, 5884—5893. Baria, A.T., Baliki, M.N., Parrish, T., Apkarian, A.V., 2011. Anatomical and functional assemblies of brain BOLD oscillations. J. Neurosci. 31, 7910—7919. Bernhardt, B.C., Rozen, D.A., Worsley, K.J., Evans, A.C., Bernasconi, N., Bernasconi, A., 2009. Thalamo-cortical network pathology in idiopathic generalized epilepsy: insights from MRI-based morphometric correlation analysis. NeuroImage 46, 373—381. Buzsaki, G., 2004. Neuronal oscillations in cortical networks. Science 304, 1926—1929. Esposito, F., Tessitore, A., Giordano, A., De Micco, R., Paccone, A., Conforti, R., Pignataro, G., Annunziato, L., Tedeschi, G., 2013. Rhythm-specific modulation of the sensorimotor network in drug-naive patients with Parkinson’s disease by levodopa. Brain 136, 710—725. Gotman, J., Grova, C., Bagshaw, A., Kobayashi, E., Aghakhani, Y., Dubeau, F., 2005. Generalized epileptic discharges show thalamocortical activation and suspension of the default state of the brain. Proc. Natl. Acad. Sci. U. S. A. 102, 15236—15240. Han, Y., Lui, S., Kuang, W., Lang, Q., Zou, L., Jia, J., 2012. Anatomical and functional deficits in patients with amnestic mild cognitive impairment. PLoS ONE 7, e28664. Han, Y., Wang, J., Zhao, Z., Min, B., Lu, J., Li, K., He, Y., Jia, J., 2011. Frequency-dependent changes in the amplitude of lowfrequency fluctuations in amnestic mild cognitive impairment: a resting-state fMRI study. Neuroimage 55, 287—295. Luo, C., Li, Q., Xia, Y., Lei, X., Xue, K., Yao, Z., Lai, Y., Marti Nez-Montes, E., Liao, W., Zhou, D., Valdes-Sosa, P.A., Gong, Q., Yao, D., 2011. Resting state basal ganglia network in idiopathic generalized epilepsy. Hum Brain Mapp. 36, 1279—1294. Mattson, R.H., 2003. Overview: idiopathic generalized epilepsies. Epilepsia 44 (Suppl. 2), 2—6. Moeller, F., Maneshi, M., Pittau, F., Gholipour, T., Bellec, P., Dubeau, F., Grova, C., Gotman, J., 2011. Functional connectivity in patients with idiopathic generalized epilepsy. Epilepsia 52, 515—522. Moeller, F., Siebner, H., Wolff, S., Muhle, H., Boor, R., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008. Changes in
Z. Wang et al. activity of striato-thalamo-cortical network precede generalized spike wave discharges. NeuroImage 39, 1839—1849. Niazy, R.K., Xie, J., Miller, K., Beckmann, C.F., Smith, S.M., 2011. Spectral characteristics of resting state networks. Prog. Brain Res. 193, 259—276. Pulsipher, D.T., Dabbs, K., Tuchsherer, V., Sheth, R.D., Koehn, M.A., Hermann, B.P., Seidenberg, M., 2011. Thalamofrontal neurodevelopment in new-onset pediatric idiopathic generalized epilepsy. Neurology 76, 28—33. Song, M., Du, H., Wu, N., Hou, B., Wu, G., Wang, J., Feng, H., Jiang, T., 2011. Impaired resting-state functional integrations within default mode network of generalized tonic—clonic seizures epilepsy. PLoS ONE 6, e17294. Wang, Z., Lu, G., Zhang, Z., Zhong, Y., Jiao, Q., Tan, Q., Tian, L., Chen, G., Liao, W., Li, K., Liu, Y., 2011. Altered resting state networks in epileptic patients with generalized tonic—clonic seizures. Brain Res. 1374, 134—141. Wang, Z., Zhang, Z., Jiao, Q., Liao, W., Chen, G., Sun, K., Shen, L., Wang, M., Li, K., Liu, Y., Lu, G., 2012. Impairments of thalamic nuclei in idiopathic generalized epilepsy revealed by a study combining morphological and functional connectivity MRI. PLoS ONE 7, e39701. Wei, L., Duan, X., Zheng, C., Wang, S., Gao, Q., Zhang, Z., Lu, G., Chen, H., 2014. Specific frequency bands of amplitude lowfrequency oscillation encodes personality. Hum Brain Mapp. 35, 331—339. Wu, C.W., Gu, H., Lu, H., Stein, E.A., Chen, J.-H., Yang, Y., 2008. Frequency specificity of functional connectivity in brain networks. NeuroImage 42, 1047—1055. Yin, Y., Li, L., Jin, C., Hu, X., Duan, L., Eyler, L.T., Gong, Q., Song, M., Jiang, T., Liao, M., Zhang, Y., Li, W., 2011. Abnormal baseline brain activity in posttraumatic stress disorder: a resting-state functional magnetic resonance imaging study. Neurosci. Lett. 498, 185—189. Yu, R., Chien, Y.L., Wang, H.L., Liu, C.M., Liu, C.C., Hwang, T.J., Hsieh, M.H., Hwu, H.G., Tseng, W.Y., 2012. Frequency-specific alternations in the amplitude of low-frequency fluctuations in schizophrenia. Hum Brain Mapp. 35, 627—637. Zang, Y.F., He, Y., Zhu, C.Z., Cao, Q.J., Sui, M.Q., Liang, M., Tian, L.X., Jiang, T.Z., Wang, Y.F., 2007. Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain Dev. 29, 83—91. Zhang, Z., Lu, G., Zhong, Y., Tan, Q., Chen, H., Liao, W., Tian, L., Li, Z., Shi, J., Liu, Y., 2010. fMRI study of mesial temporal lobe epilepsy using amplitude of low-frequency fluctuation analysis. Hum Brain Mapp. 31, 1851—1861. Zhong, Y., Lu, G., Zhang, Z., Jiao, Q., Li, K., Liu, Y., 2011. Altered regional synchronization in epileptic patients with generalized tonic—clonic seizures. Epilepsy Res. 97, 83—91. Zou, Q.H., Zhu, C.Z., Yang, Y., Zuo, X.N., Long, X.Y., Cao, Q.J., Wang, Y.F., Zang, Y.F., 2008. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J. Neurosci. Methods 172, 137—141. Zuo, X.-N., Di Martino, A., Kelly, C., Shehzad, Z.E., Gee, D.G., Klein, D.F., Castellanos, F.X., Biswal, B.B., Milham, M.P., 2010. The oscillating brain: complex and reliable. NeuroImage 49, 1432—1445.
