Static And Dynamic Abnormalities Of Intrinsic Brain Activity In Temporal Lobe Epilepsy

Background and PurposesTemporal lobe epilepsy(TLE)is the most common form of refractory focal epilepsy that is suitable for surgery.It is of great urgency to correctly understand the disease of epilepsy and to give adequate attention to the patients.It is of great clinical significance to lateralize the epileptic foci and study the origin and transmssion pathways of epileptic activity.Compared with other types of epilepsy,TLE is more likely to reveal cognitive impairment,particularly in memory and language.In addition,TLE patients are also prone to have attention,perception,personality changes,and other social cognitive impairments,accompanied by depression,anxiety,and other psychiatric symptoms.In 2002 and 2004,Spencer et al.and Blumenfeld et al.first proposed the concepts of "temporal lobe epilepsy network" and "epilepsy inhibition network",respectively.It is believed that some of the clinical manifestations and interictal complications of TLE were closely associated with the abnormal changes in the TLE network.The study of the TLE network will contribute to our in-depth understanding of TLE patho-physiological mechanisms,and is expected to be an auxiliary tool for the clinical diagnosis or provide a new direction for the treatment.In recent years,with the rapid development of MRI technology,the exploration of epilepsy networks using resting-state functional magnetic resonance imaging(rs-fMRI)has become a new research focus.Its non-invasiveness,simplicity,and high feasibility have greatly promoted its application in clinical research.Rs-fMRI focuses on the spontaneous fluctuations of BOLD signals in the low-frequency band(<0.1 Hz),also known as intrinsic brain activity(IBA),which helps to keep the brain in a state of readiness.At present,more and more researchers have confirmed the functional importance of IBA.Also,more and more rs-fMRI-based IBA measurements have been proposed,and each method reveals a unique aspect.Nevertheless,the differences and similarities between these approaches need to be fully empirically explored,as the convergence of the findings may suggest the existence of a more fundamental mechanism that is capable of revealing more real abnormal neuronal activity.The most basic IBA measurements focus on the measurement of regional fluctuation,of which the most commonly used are the amplitude of low-frequency fluctuations(ALFF)and fractional amplitude of low-frequency fluctuations(fALFF).The measurements based on inter-regional synchronicity are regional homogeneity(ReHo),voxel-mirrored homotopic connectivity(VMHC),degree centrality(DC),and global signal correlation(GSCorr),which aim to measure the consistency among adjacent areas,consistency between hemispheres,and the degree of network centrality,respectively.Traditional processing methods of fMRI data generally assume that brain function is constant during the scanning period,so the indicators calculated based on this assumption are static indicators.However,the activity of the human brain is environmentally sensitive and activity-dependent,which means that the IBA is highly dynamic and fluctuates over time.Thus,static IBA metrics are insufficient to describe transient changes in IBA.Based on the demand for detecting the dynamic characteristics of IBA,many corresponding dynamic indicators have continuously emerged,such as dynamic ALFF,dynamic fALFF,dynamic ReHo,dynamic DC,dynamic VMHC,dynamic GScorr,etc.The above indicators,which extended from static to dynamic,are mostly based on sliding window technology,not only further reflect rich dynamic information but also proved to have high stability.By calculating the standard deviation(SD)or other time-domain-related variability parameters of these indexes in all the windows,they reflect the dynamic change degree or stability of each index in the time dimension.According to the Chinese and English literature search,we found that most of the IBA-related studies of TLE only used one or two indicators(mainly ALFF,fALFF and ReHo),and the results were partially inconsistent.Studies on VMHC,DC,and GSCorr are rare or even vacant,and comprehensive studies of many indicators are even rarer.In addition,no relevant studies were retrieved on the dynamic changes of IBA in TLE,such as dynamic ALFF and dynamic ReHo,as well as the change in consistency among the indicators.Consequently,there is still a large research space.Therefore,with the purpose of comprehensively understanding the pattern of IBA change in TLE patients and analyzing the consistency and complementarity of various indicators,we comprehensively studied the static and temporal dynamic characteristics of IBA in unilateral TLE patients from the perspectives of local activity characteristics(ALFF,fALFF),local synchronization or connectivity characteristics(ReHo,VMHC)and the whole brain connectivity characteristics(DC,GSCorr).Moreover,the correlations between disease duration,severity,cognitive scores,and each indicator were explored.Materials and Methods1.A total of 57 patients with TLE(27 RTLE and 30 LTLE)and 42 healthy volunteers were enrolled in this study.MRI data were acquired using a 3.0 T Magnetom Prisma MRI scanner(Siemens Healthcare,Erlangen,Germany)with a 64-channel head coil.All subjects underwent resting SMS-BOLD-fMRI and 3D T1WI scanning and were told to lie down flat,close their eyes,remain awake,and not think.Clinical data collection and scale assessment were conducted for all the subjects,including:basic information;seizures duration;national hospital seizure severity scale(NHS3);mini-mental state examination(MMSE);memory and executive screening(MES);Montreal cognitive assessment-basic(MOCA-B);auditory verbal learning test(AVLT);and shape trail making test(STT).2.The DPARSF(http://rfmri.org/DPARSF)was used to preprocess the fMRI data.The following steps were performed:(1)conversion of data formats;(2)removal of the first 10 time points;(3)slice-timing correction;(4)realignment;(5)spatial normalization;(6)regression of covariant.3.Calculation of the static indicators of IB A:the Data Processing and Analysis of Brain Imaging(DPABI)toolbox was used to calculate the following six indicators,including ALFF,fALFF,ReHo,DC,GSCorr,and VMHC.4.Calculation of the dynamic indicators of IBA:the temporal dynamic analysis module in the DPABI toolbox and the sliding window method were used to calculate the dynamic indicators.(1)Sliding to select a rectwin window in the BOLD time series with a window width of 60 TR(60 s)and a step size of 30 TR(30 s).(2)6 rs-fMRI indicators(ALFF,fALFF,ReHo,DC,GSCorr,and VMHC)were calculated in each window.(3)The standard deviation(SD)of the indicator values in all the windows was calculated to reflect the degree of temporal dynamic change of each indicator.5.In order to enlarge the study sample size and facilitate the analysis of IB A changes on the affected side or contralateral side,the images of 27 patients with right TLE were flipped left.Thus,57 left TLE patients were finally enrolled in the statistical analysis.6.Statistical analysis:(1)Comparison of clinical data and cognitive scale scores:If the data were consistent with normal distribution,an independent sample T-test was performed by SPSS software(version 17.0);otherwise,the Kruskal-Wallis test was performed.χ2 test was used for gender comparison.Statistical significance was defined as P<0.05.(2)Differential analysis of the IB A static and dynamic indicators between the two groups:Two-sample T tests were performed between the two groups,and age,sex,education level,and mean FD Jenkinson values were incorporated into the regression model as covariants.Multiple comparison correction was performed using a Gaussian random field(GRF)with thresholds of voxel-wise Pvoxel<0.005 and cluster-level Pcluster<0.05.7.Correlation analysis of static and dynamic IBA indicators with clinical data and cognitive scale scores in the TLE group:With the peak voxels of abnormal regions as spherical centers,spherical ROIs were constructed around these abnormal regions(with a 6 mm radius).To assess the relationship between the scores(including epilepsy duration,NHS3,MMSE,MES,MOCA-B,AVLT,STT-A,and STT-B scores)and metrics in these abnormal regions,we used Spearman rank correlation analysis by SPSS software(version 17.0).Statistical significance is defined as P<0.05.8.Validation analysis:Different parameters(GSR regression,FWHM,sliding window width and size)were used to validate our findings.Results1.Comparison of clinical data and cognitive scale scores between the TLE group and HC group:No significant differences were found in age or gender between the two groups(P>0.05).The 9 cognitive scale scores(including MMSE,MES,MOCA-B,N1,N2,N3,N4,N5,recognition)all decreased in the TLE group,while the time consumption of STT-A and STT-B were longer in the TLE group(P<0.05).2.Abnormalities of static IB A indicators in the TLE group and their correlation with clinical indicators and cognitive scale scores.2.1 Inter-group differences of ALFF:Compared with the healthy volunteers,the ALFF values in the TLE group increased in the left cingulate gyrus,cerebellum,hippocampus,thalamus,supplementary motor area,precuneus,cuneus,and right pallidum,lenticula,amygdala,and parahippocampal gyrus.However,the ALFF values decreased in the right superior cerebellar lobe,anterior and posterior central gyrus,triangular inferior frontal gyrus,middle frontal gyrus,and orbital frontal gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The ALFF values of left cingulate gyrus,cerebellum,hippocampus and right pallidum were positively correlated with the duration of epilepsy,but the ALFF values of the right posterior central gyrus and orbital frontal gyrus were negatively correlated with the duration of epilepsy.The ALFF values of the left cerebellum,hippocampus,and right pallidum were positively correlated with STT.The ALFF values of the left precuneus were positively correlated with MMSE,MOCA,and AVLT(P<0.05).2.2 Inter-group differences of fALFF:Compared with the healthy volunteers,the fALFF values in the TLE group increased in the right hippocampus,amygdala,temporal pole,and left hippocampus and parahippocampal gyrus,thalamus,middle and inferior temporal gyrus,superior marginal gyrus,supplementary motor area,medial superior frontal gyrus,and pons.However,the fALFF values decreased in the left superior and middle temporal gyrus,superior margin gyrus,inferior parietal angular gyrus,angular gyrus,and right anterior and posterior central gyrus,middle and inferior frontal gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The fALFF values of the right hippocampus were positively correlated with the duration of epilepsy,but the fALFF values of right posterior central gyrus were negatively correlated with the duration of epilepsy.The fALFF values of the bilateral hippocampus and left thalamus were positively correlated with STT and negatively correlated with MMSE,MOCA,and AVLT(P<0.05).2.3 Inter-group differences of ReHo:Compared with the healthy volunteers,the ReHo values in the TLE group decreased in the left superior,middle and inferior temporal gyrus,temporal pole,inferior parietal angular gyrus,and superior marginal gyrus.While ReHo values increased in the left supplementary motor area and medial superior frontal gyrus(GRP corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The ReHo values of the left middle temporal gyrus were negatively correlated with the duration of epilepsy and positively correlated with MMSE and MOCA.The ReHo values of the left supplementary motor area were positively correlated with STT(P<0.05).2.4 Inter-group differences of DC:In the TLE group,the DC values decreased in the left superior,middle and inferior temporal gyrus,Heschl’s gyrus,temporal pole,and superior marginal gyrus.While the DC values in the left precuneus,cuneus,calcarine cortex and supplementary motor area were increased(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The DC values of the left supplementary motor area were negatively correlated with STT-A.The DC values of the left precuneus were positively correlated with AVLT(P<0.05).2.5 Inter-group differences of GSCorr:In the TLE group,the GSCorr values decreased in the left superior,middle and inferior temporal gyrus,Heschl’s gyrus,and temporal pole.While in the left middle and inferior temporal gyrus,the GSCorr values increased(GRF corrected,Pvoxel<0.005,pcluster<0.05,T=2.63).The GSCorr values of the left inferior temporal gyrus were positively correlated with AVLT(P<0.05).2.6 Inter-group differences of VMHC:In the TLE group,the VMHC values decreased in the bilateral superior,middle and inferior temporal gyrus,temporal pole,fusiform gyrus,superior cerebellar lobe,and precuneus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The VMHC values of the middle temporal gyrus,superior cerebellar lobe,and precuneus were negatively correlated with the duration of epilepsy.The VMHC values of the middle temporal gyrus were positively correlated with MOCA.The VMHC values of precuneus were negatively correlated with the STT-B(P<0.05).2.7 Overlapping abnormal brain regions of the 6 static IB A indicators:According to the comprehensive analysis of the above results,we found that the abnormal brain regions detected by the above 6 indicators overlapped a lot and were mainly listed as follows:(1)There’s marked overlapped abnormal brain regions of ALFF and fALFF,which increased in the bilateral medial temporal lobe structure,thalamus,and supplementary motor area of the affected side but decreased in the anterior and posterior central gyrus,and lateral frontal regions of the contralateral side.(2)The values of ALFF,fALFF,Reho,and DC both increased in the supplementary motor area of the affected side.(3)ALFF and DC both increased in the cuneus and precuneus of the affected side.And VMHC decreased in the bilateral precuneus.(4)ALFF and VMHC both showed alterations in the cerebellum.(5)The values of fALFF,Reho,DC,GSCorr,and VMHC all decreased in the lateral temporal structure(superior,middle and inferior temporal gyrus,temporal pole)of the affected side.(6)The most static IBA indicators significantly correlated with the duration of epilepsy and several cognitive scores(P<0.05).3.Abnormalities of dynamic IBA indicators in the TLE group and its correlation with clinical indicators and cognitive scale score:3.1 Inter-group differences of SDALFF:Compared with the healthy volunteers,the SDALFF values in the TLE group increased in the bilateral hippocampus,thalamus,olfactory cortex,medial cingulate gyrus,left posterior cingulate gyrus,precuneus,cerebellum,and right amygdala.However,the SDALFF values decreased in the right superior cerebellar lobe,orbital frontal gyrus,superior frontal gyrus,and rectus gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The SDALFF values of left cerebellum and bilateral thalamus were positively correlated with the duration of epilepsy,but the SDALFF values of the right superior cerebellar lobe were negatively correlated with the duration of epilepsy.The SDALFF values of the bilateral thalamus were positively correlated with STT but negatively correlated with MES or MOCA(P<0.05).3.2 Inter-group differences of SDfALFF:Compared with the healthy volunteers,the SDfALFF values in the TLE group increased in bilateral supplementary motor areas,left medial superior frontal gyrus,anterior cingulate gyrus,and right dorsolateral superior frontal gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).SDfALFF values of the right dorsolateral superior frontal gyrus were positively correlated with epilepsy duration,MES,and AVLT but negatively correlated with STT(P<0.05).3.3 Inter-group differences of SDReHo:Compared with the healthy volunteers,the SDReHo values in the TLE group decreased in the left superior and middle temporal gyrus,Heschl’s gyrus,superior marginal gyrus,as well as in the right calcarine cortex and cuneus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The SDReHo values of the right calcarine cortex were negatively correlated with the duration of epilepsy and STT.The SDReHo values of the left superior temporal gyrus were negatively correlated with AVLT but positively correlated with STT(P<0.05).3.4 Inter-group differences of SDDC:Compared with the healthy volunteers,the SDDC values in the TLE group decreased in the bilateral calcarine cortex and cuneus,left superior occipital gyrus,left lingual gyrus,left superior temporal gyrus,right precuneus,while increased in the left middle temporal gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The SDDC values of the left middle temporal gyrus were positively correlated with the duration of epilepsy and AVLT.The SDDC values of the left superior temporal gyrus and bilateral calcarine cortex were negatively correlated with AVLT.The SDDC values of the left calcarine cortex were negatively correlated with the duration of epilepsy(P<0.05).3.5 Inter-group differences of SDGSCorr:Compared with the healthy volunteers,the SDGSCorr values in the TLE group decreased in the bilateral calcarine cortex and cuneus,left superior/middle/inferior occipital gyrus,left superior temporal gyrus,Heschl’s gyrus,and right lingual gyrus(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63).The SDGSCorr values of the left calcarine cortex and superior temporal gyrus were negatively correlated with AVLT.The SDGSCorr values of the left calcarine cortex were negatively correlated with the duration of epilepsy(P<0.05).3.6 Inter-group differences of SDVMHC:Compared with the healthy volunteers,the SDVMHC values in the TLE group decreased in the bilateral calcarine cortex,cuneus,and paracentral lobule.(GRF corrected,Pvoxel<0.005,Pcluster<0.05,T=2.63)3.7 Overlapping abnormal brain regions of the 6 dynamic IB A indicators:The abnormally activated brain regions detected by the above 6 dynamic indicators overlapped a lot and were partially consistent with the static indicators’ result.The findings are listed as follows:① The values of SDReHo,SDDC,SDGSCorr,and SDVMHC all decreased in the visual network areas(bilateral calcarine gyrus and cuneus),while no abnormal changes in the visual network areas were detected in the corresponding static indicators.②SDReHo,SDDC,SDGSCorr,and corresponding static indicators all decreased in the lateral temporal lobe(superior and middle temporal gyrus,Heschl’s gyrus)of the affected side.③SDfALFF and fALFF both increased in the supplementary motor area and medial superior frontal gyrus of the affected side.④There’s a lot of overlapped abnormal brain regions of ALFF and SDALFF,which increased in bilateral medial temporal lobe structure and thalamus,cerebellum,cingulate gyrus,and precuneus of the affected side,and basal ganglia of the contralateral side,while decreased in the superior cerebellar lobe and orbital superior/middle frontal gyrus of the contralateral side.⑤Most of the dynamic IBA indicators were significantly correlated with the duration of epilepsy and several cognitive scores(P<0.05).Conclusions1.For TLE patients,the combined application of static and dynamic IBA indicators(ALFF,SDALFF,fALFF,SDfALFF,ReHo,SDReHo,DC,SDDC,GSCorr,SDGSCorr,VMHC,SDVMHC)revealed significant application value in many aspects,including the detection of abnormal neuronal activities,the lateralization of epileptogenic foci,the exploration of seizure activities’ transmission and inhibition pathways,and the analysis of cognitive impairment or compensatory mechanisms.2.The temporal dynamic IBA indicators contribute to the in-depth verification and comprehensive complement of the static indicators,and their combined use could reveal more true and credible abnormal neuronal activities.