Share

November 01, 2016; 87 (18) Article

MRI-negative temporal lobe epilepsy

A network disorder of neocortical connectivity

David N. Vaughan, Genevieve Rayner, Chris Tailby, Graeme D. Jackson
First published September 30, 2016, DOI: https://doi.org/10.1212/WNL.0000000000003289
Full PDF
Citation
Permissions

Make Comment

See Comments

Downloads
617

Share

Abstract

Objective: To define the functional network changes that characterize MRI-negative temporal lobe epilepsy (TLE) and TLE with hippocampal sclerosis (HS-TLE).

Methods: We studied 36 patients with medically refractory unilateral TLE, having either a normal clinical MRI (n = 18) or unilateral hippocampal sclerosis (n = 18). Patients were compared to healthy controls of equivalent age and sex (n = 27). Functional connectivity in 10 minutes of task-free functional MRI was assessed using a voxel-resolution graph theoretic analysis, using the metrics of degree, clustering coefficient, eigenvector, and betweenness centrality. Significant clusters were further explored with a seed-based analysis.

Results: MRI-negative TLE showed decreased connectivity at the ipsilateral superior and middle temporal gyri compared to controls (decreased eigenvector centrality). No functional abnormality was detected within mesial temporal structures. In contrast, HS-TLE showed increased connectivity within the affected hippocampus and anterior thalamus (increased clustering coefficient) and decreased connectivity of the ventromesial prefrontal cortex (decreased betweenness centrality). Using the detected clusters as seed regions revealed decreased connectivity from the sclerotic hippocampus to both the contralateral temporal lobe and regions of the default mode network.

Conclusion: MRI-negative TLE is associated with impaired interictal connectivity of the temporal neocortex, lateralized to the epileptic side. HS-TLE shows a different pattern, with functional segregation of the sclerotic hippocampus and impairment of its long-range connectivity. This suggests that MRI-negative TLE is not merely a subtle version of hippocampal sclerosis, but is rather a separate condition that involves distinct brain networks.

GLOSSARY

fMRI=
functional MRI;
FWE=
family-wise error;
FWHM=
full width at half maximum;
HS=
hippocampal sclerosis;
ROI=
region of interest;
TE=
echo time;
TLE=
temporal lobe epilepsy;
TR=
repetition time;
vmPFC=
ventromesial prefrontal cortex

MRI-negative temporal lobe epilepsy (TLE) comprises an important subgroup of patients with temporal lobe seizures, who have no evidence of hippocampal sclerosis (HS) or of any other epileptogenic lesion on structural MRI. Compared to patients with TLE with HS (HS-TLE), these patients typically have later onset of epilepsy,1 lower prevalence of febrile convulsions,2,3 less impairment of memory function,4 and lower rates of seizure freedom after epilepsy surgery.5 Histopathology may be normal or show relatively mild hippocampal neuronal loss,3 and the temporal neocortex is either normal or shows subtle dyslamination.6 Even though seizures appear similar to those of HS-TLE, these features suggest that different brain networks may be affected in MRI-negative TLE.

Task-free functional MRI (fMRI) has been used to demonstrate abnormal network connectivity in HS-TLE, particularly affecting the hippocampus, contralateral mesial temporal structures, and the default mode network.7,–,11 In contrast, a distinctive abnormality of network connectivity has yet to be identified in MRI-negative TLE. Graph theory methods are ideally suited to describe such network abnormalities. Previous analyses in TLE have used a relatively coarse-grained approach, where the measured signal is first averaged across large territories.12,–,16 A fine-grained approach calculated at the resolution of single voxels is more computationally intensive, but can detect abnormalities of small cortical and subcortical regions.17,18

Here we apply a voxel-wise graph theoretical analysis to task-free fMRI data in patients with TLE to investigate at high spatial resolution the network alterations that are associated with MRI-negative TLE and how these differ from the archetype of HS-TLE.

METHODS

Participants.

Patients with medically refractory unilateral TLE were recruited from the Austin Hospital Comprehensive Epilepsy Program, Melbourne, Australia, between 2009 and 2015. Diagnosis was based on clinical features, scalp video-EEG recording of seizures, and congruent nuclear medicine abnormalities (FDG-PET hypometabolism or ictal-interictal SPECT hyperperfusion). Patients with independent left and right temporal lobe seizures were not included. Routine clinical MRIs were performed with our epilepsy protocol19 and reviewed by a neuroradiologist and an epileptologist. Patients were only included if there were characteristic features of unilateral HS (HS-TLE), or if there was no potentially epileptogenic structural abnormality (MRI-negative TLE).

Twenty-seven patients with HS-TLE and 21 patients with MRI-negative TLE met inclusion criteria, consented to participate, and had fMRI performed. Of these, 6 HS-TLE and 2 MRI-negative TLE participants were excluded because of truncated fMRI acquisitions (<10 minutes) or excess head motion. To balance cohorts for epileptic side and sex, 1 male participant with right MRI-negative TLE and 3 female participants with right TLE-HS were randomly selected and excluded. The final analysis (table 1) included 18 participants each with MRI-negative TLE and HS-TLE. A matching cohort of 27 healthy controls were selected from historical fMRI cohorts, with stratification for MRI scanner and sex. These group sizes are similar to previous fMRI studies that were sufficiently powered to detect significant effects in HS-TLE.7,–,11

View this table:
Table 1

Demographic and clinical characteristics of participants

Standard protocol approvals, registrations, and patient consents.

The protocol was approved by the Austin Hospital Human Research Ethics Committee. All participants, or their legal guardians in the case of minors, gave informed consent.

MRI acquisition and preprocessing.

MRI was performed on 3T Siemens (Erlangen, Germany) Tim Trio or Skyra scanner, with groups balanced for the scanner used. Resting-state functional images were acquired using a whole-brain gradient-echo single shot echoplanar imaging sequence (echo time [TE] 30 ms, repetition time [TR] 3,000 ms, flip angle 85°, matrix 72 × 72, 44 axial slices at angle + 30°, 3 mm isotropic voxels). Participants were asked to keep their eyes closed and not fall asleep. A total of 210 volumes were available in all participants. A T1-weighted magnetization-prepared rapid gradient echo image was also acquired (TR 1,900 ms, inversion time 900 ms, TE 2.6 ms, flip angle 9°, 0.9 mm isotropic voxels). Hippocampal volumes were measured from the T1-weighted images using FreeSurfer 5.3.

Functional data were adjusted for slice timing and head motion, aligned using boundary-based registration to the anatomical image,20 then non-linearly warped to a symmetrical template. Removal of spurious signals due to scanner artefact, head motion, and non-neural physiologic processes was performed with a novel combination of 2 data-driven denoising methods: SOCK21 and CompCor22 (appendix e-1 at Neurology.org). No spatial smoothing was performed before the connectivity graph was calculated.

Patients with right-sided epilepsy, and an equal proportion of randomly selected controls, were flipped right to left to align the epileptic side.

Voxel-wise graph theoretic analysis.

An undirected unweighted (binary) graph was formed for each individual, with a node for every voxel within a symmetrical gray matter mask (51,603 voxels). An edge was added between every voxel pair for which the Pearson correlation coefficient was r > rthreshold > 0. Only the largest connected cluster was retained. The value for rthreshold was calculated adaptively in each individual to maintain a constant total number of edges per graph, ensuring that individual differences are due to graph topology rather than graph size or edge density. We empirically chose a mean of 20 edges per node to keep the majority of voxels within the mask connected. For each individual, voxel-wise maps were calculated for degree, clustering coefficient, eigenvector centrality, and betweenness centrality23 (figure 1), as these metrics are each sensitive to different types of network abnormality.

Figure 1
Figure 1 Mean voxel-wise graph metric values within each group

Mean graph metric values in each group, over the hemisphere ipsilateral to the epileptic focus. Circled regions indicate a significant difference from controls. (A) Degree is the number of connections from each node. (B) Clustering coefficient is the proportion of edges (red lines) that are present between a node's neighbors. (C) Eigenvector centrality measures importance within the graph, where high-scoring nodes are connected to other high-scoring nodes. (D) Betweenness centrality measures the number of shortest paths through a node, emphasizing nodes that form bridges of connectivity between clusters. HS = hippocampal sclerosis; TLE = temporal lobe epilepsy.

Primary analysis: Comparison of voxel-wise graph metrics of TLE groups to controls.

After calculation of the voxel-wise graph metrics, maps were spatially smoothed to improve functional alignment between participants (Gaussian kernel 8 mm full width at half maximum [FWHM]), then rank-transformed to a normal distribution. Each TLE group was compared to controls using a t test contrast within a one-way analysis of covariance model for each graph metric, including the left-to-right flip and MRI scanner as covariates of no interest, and correction for spatial nonstationarity within SPM8.24 The significance threshold was voxel-level p < 0.001, cluster-level p < 0.05, family-wise error (FWE) corrected. At each significant cluster a region of interest (ROI) analysis was performed, by extracting the mean voxel-wise graph metric value for each individual. Values for each participant group were compared using a 2-tailed t test, with p values corrected for false discovery rate over all comparisons.

To explore differences between left-sided and right-sided epilepsy, an additional analysis was performed with groups partitioned as left HS-TLE, right HS-TLE, left MRI-negative TLE, right MRI-negative and controls. An F-contrast modeling left vs right epilepsy (with MRI-negative TLE and HS-TLE specified in separate rows) did not detect any significant clusters for any of the graph metrics (all pFWE > 0.36).

Secondary analysis: Seed-based functional connectivity of detected clusters.

A seed-based analysis was used to further explore connectivity of clusters detected in the primary analysis. Preprocessed fMRI data were spatially smoothed (6 mm FWHM Gaussian kernel). The time course within each seed region was extracted using the first eigenvariate and correlated against activity at all voxels. Correlations were converted using the Fisher r-to-z transformation and SPM groupwise comparison performed as above.

RESULTS

Participant demographics and clinical characteristics.

MRI-negative TLE had epilepsy onset 7.3 years later than HS-TLE (2-tailed t test p = 0.02), with a corresponding shorter duration of epilepsy (table 1). Most patients were taking 2 or 3 antiepileptic medications; carbamazepine, levetiracetam, and lamotrigine were the most common. Patient-reported seizure frequency was not different between groups (p = 0.8). Frequency of interictal epileptiform discharges on clinical scalp EEG, obtained outside of the MRI scanner and counted over a 1-hour period, was not different between groups (MRI-negative TLE median 6/h, range 0–194/h; HS-TLE median 3/h, range 0–49/h; log-rank test p = 0.11).

Ipsilateral hippocampal volumes in MRI-negative TLE were no different from controls (4,776 ± 371 mm3 vs 4,630 ± 272 mm3, p = 0.40, adjusted for intracranial volume and brain side). In HS-TLE, ipsilateral hippocampal volumes (3,579 ± 400 mm3 SD) were smaller than in controls (p < 0.0001) and patients with MRI-negative TLE (p < 0.0001).

In HS-TLE, 15/18 patients had subsequent temporal lobe resection (8 right, 7 left), with histologic confirmation of HS in all. In patients with MRI-negative TLE, 5/18 underwent anterior temporal lobe resection (4 right, 1 left). No histopathologic abnormality was found in 4 of these patients. One patient with MRI-negative TLE showed mild neuronal loss in the cornu ammonis. There was good postoperative seizure outcome in 13/15 patients with HS-TLE and 4/5 patients with MRI-negative TLE (see table e-1 for details).

fMRI acquisition and graph formation.

Average head motion during fMRI acquisition was not significantly different between groups (mean frame-wise displacement ± SD: HS-TLE 0.17 ± 0.09 mm, MRI-negative TLE 0.18 ± 0.09 mm, controls 0.14 ± 0.05 mm, one-way analysis of variance p = 0.11). More high-motion time points (frame-wise displacement >0.5 mm) were censored in patients than controls (HS-TLE 23.0 ± 24.9 volumes, MRI-negative TLE 22.8 ± 26.0, controls 8.7 ± 10.4, p = 0.02). The adaptive correlation threshold produced graphs of 50,610 ± 1,896 SD connected voxels (98.1 ± 3.7% of voxels within the mask; no group-wise difference, p = 0.85). The mean rthreshold was 0.41 ± 0.05, and did not differ between groups (p = 0.33).

Comparison of graph metrics in controls.

The 4 metrics produced similar spatial patterns in controls, with highest connectivity values in parieto-occipital cortex (figure 1), and low values at the brainstem and cerebellum. Correlation between metrics was highest for degree/betweenness centrality (mean r = 0.91 ± 0.13 SD), and lowest for eigenvector centrality/clustering coefficient (0.50 ± 0.12) and eigenvector/betweenness centrality (0.55 ± 0.15) (table e-2).

Primary analysis in MRI-negative TLE: Decreased connectivity of ipsilateral temporal neocortex.

MRI-negative TLE showed decreased eigenvector centrality relative to controls at the anterior part of the superior temporal sulcus, bordering the ipsilateral superior temporal gyrus and middle temporal gyrus (figure 2, cluster pFWE = 0.022). The ROI analysis showed additional decreases in degree and betweenness centrality at this region in MRI-negative TLE.

Figure 2
Figure 2 Primary analysis in MRI-negative temporal lobe epilepsy (TLE): Significant clusters vs controls

(A) Clusters of increased or decreased voxel-wise connectivity in MRI-negative TLE compared to controls (p < 0.05, family-wise error cluster corrected). Comparisons for all other graph metrics were not significant. (B) Individual voxel-average connectivity values within the detected cluster. Group means are marked with “+.” Red bar indicates significant difference between groups (p < 0.05, false discovery rate corrected).

Primary analysis in HS-TLE: Abnormal hippocampal, thalamic, and prefrontal connectivity.

Within the affected hippocampus, HS-TLE showed higher clustering coefficient than controls (figure 3A, cluster pFWE = 0.016). Connectivity was mostly to other hippocampal voxels, indicating an abnormally clique-like segregated network within the sclerotic hippocampus. Voxels in the contralateral thalamus, at the location of the anterior nucleus, also showed increased clustering coefficient (figure 3A, voxel p < 0.0001, cluster pFWE = 0.03).

Figure 3
Figure 3 Primary analysis in hippocampal sclerosis (HS)–temporal lobe epilepsy (TLE): Significant clusters vs controls

(A) Clusters of increased or decreased voxel-wise connectivity in HS-TLE compared to controls (p < 0.05, family-wise error cluster corrected). Other comparisons were not significant. (B) Individual voxel-average connectivity values within the detected clusters. Group means are marked with “+.” Red bar indicates significant difference between groups (p < 0.05, FDR corrected).

The ventromesial prefrontal cortex (vmPFC) showed decreased betweenness centrality in HS-TLE (figure 3B, cluster pFWE = 0.04), with slightly greater spatial extent to the ipsilateral side. The ROI analysis showed a decrease in degree for both HS-TLE and MRI-negative TLE.

Secondary analysis in MRI-negative TLE: Seed-based connectivity of the abnormal temporal neocortex.

Conventional seeded connectivity, from the cluster detected in MRI-negative TLE at the ipsilateral temporal neocortex (figure 4A), showed strongest positive connectivity to bilateral anterior temporal and anterior cingulate regions. Other areas of connectivity included frontal cortex, insula, and subcortical structures; thalamus, caudate, medulla, and pons. The pattern of connectivity was no different from that found in controls.

Figure 4
Figure 4 Secondary analysis: Seed-based connectivity of detected clusters

(A) Seed-based connectivity from the ipsilateral temporal neocortex, using the region detected in the primary analysis of MRI-negative temporal lobe epilepsy (TLE) (figure 2). Regions of significant mean positive connectivity in the MRI-negative TLE group are shown in red/yellow (voxel threshold p < 0.001, uncorrected). (B–D) Seed-based connectivity from the regions detected in the primary analysis of TLE with hippocampal sclerosis (HS) (figure 3). Regions of significant mean connectivity in the HS-TLE group are shown in red/yellow. Right panel (gray) shows regions of relatively decreased seeded connectivity when compared to controls (voxel threshold p < 0.001 uncorrected). DMN = default mode network; vmPFC = ventromesial prefrontal cortex.

Secondary analysis in HS-TLE: Decreased seed-based connectivity between the affected hippocampus and default mode network.

Seeded connectivity from the ipsilateral hippocampus in HS-TLE (figure 4B) showed strong positive connectivity to ipsilateral mesial temporal structures, and also to contralateral mesial temporal structures and ipsilateral cortical regions of the default mode network. Comparison to controls showed decreased connectivity to the contralateral hippocampus and parahippocampal gyrus (cluster pFWE = 0.008), the contralateral temporal neocortex, vmPFC, and posterior cingulate (voxel p < 0.001 uncorrected).

Seeded connectivity from the contralateral thalamic cluster (figure 4C) showed bilateral connectivity to the thalami, caudate, anterior cingulate, cerebellum, and piriform cortex, similar in TLE and controls.

Seeded connectivity from the ventromedial prefrontal cortex (figure 4D) showed positive connectivity to vmPFC, the hippocampi, middle temporal gyri, precuneus/posterior cingulate and lateral parietal lobe (regions of the default mode network), and orbitofrontal cortex. Compared to controls, HS-TLE showed a decrease in connectivity from vmPFC to the contralateral frontal operculum (cluster pFWE = 0.005) and the affected hippocampus (voxel p < 0.001 uncorrected).

Direct comparison of MRI-negative TLE to HS-TLE.

ROI analysis (figure 3B) showed HS-TLE has significantly greater clustering coefficient than MRI-negative TLE at the ipsilateral hippocampus and anterior thalamus. Voxel-wise comparison between MRI-negative TLE and HS-TLE over all gray matter voxels did not detect any significant clusters at the prespecified threshold. However, examination of the effect size map (figure e-1) shows the greatest differences at ipsilateral hippocampus (cluster pFWE = 0.10), thalamus, and temporal neocortex, in a pattern that recapitulates the ROI analysis and the detected differences from controls. Comparison of seed-based connectivity showed clusters of significantly greater connectivity in MRI-negative TLE/lesser connectivity in HS-TLE for each seed, involving the mesial temporal, insular, and frontal operculum regions (figure e-2).

DISCUSSION

Using a voxel-resolution connectomics method, we found that MRI-negative TLE is associated with impaired connectivity of the lateral temporal neocortex, ipsilateral to the side of seizure onset. This is a very different pattern from HS-TLE, where there is increased connectivity within the affected mesial temporal lobe and anterior thalamus. This indicates a fundamental difference in the way interictal networks are affected in each form of TLE.

The specific finding in MRI-negative TLE was of decreased eigenvector centrality, indicating a local change in blood oxygenation level–dependent signal fluctuations that reduces the correlation with other highly connected brain regions. Decreases in degree and betweenness centrality at this region also reflect disturbance of the normal network. Lateral temporal abnormalities in MRI-negative TLE are also described on interictal nuclear medicine studies, with glucose hypometabolism of the inferolateral temporal neocortex,2,25 decreased 5HT1A binding at the lateral temporal lobe,26 and increased GABAA binding in temporal lobe white matter.27 Connectivity from the lateral temporal cortex (using a seed-based approach) was to bilateral anterior temporal, frontal, and subcortical structures, which suggests how these regions could become involved in downstream dysfunction driven from the epileptic network.

The mesial temporal region in MRI-negative TLE was not abnormal on any of the graph theory measures. Neuropsychological testing performed in similar cohorts also shows little impairment in functions relating to mesial temporal structures, with verbal and nonverbal memory function being generally preserved.4 These observations support the concept that the mesial temporal structures are not a primary site of pathology or dysfunction in MRI-negative TLE.

Unlike MRI-negative TLE, HS-TLE does show a significant mesial temporal abnormality, with increased clustering coefficient at voxels within the affected hippocampus. Furthermore, both our seed-based analysis and previous fMRI studies show reduced long-range connectivity from the hippocampus to other brain regions.8,9 Increased connectivity within the hippocampus is a new imaging finding, which cannot be detected using seed-based methods, and is complementary to those previous results. The finding is supported by the physiologic report of increased coherence of EEG signals within the hippocampus in HS-TLE, which is due to both interictal epileptic discharges and intrinsically coherent neuronal oscillations at the alpha and beta frequencies.28

In HS-TLE, the vmPFC showed reduced betweenness centrality. This metric indicates a reduction in the role of the vmPFC as a connector between other network hubs, for example the hippocampi, insular cortex, and other nodes of the default mode network, which were detected in the seed-based analysis (figure 4) and also in other studies.29,30 Subtle frontal lobe dysfunction can be detected clinically in TLE on neuropsychological tests of cognitive flexibility.31,32 Moreover, impairment of theory-of-mind function (for instance, the recognition of a social faux pas) occurs in HS-TLE,33 and has been associated with abnormal function of the vmPFC.34 The current finding suggests that such frontal lobe symptoms may have a basis in interictal disruption of connectivity of the prefrontal cortex. Furthermore, reduced degree at this region in MRI-negative TLE (on the ROI analysis) indicates it may be a more general feature of TLE, albeit of lesser severity in the lesion-negative group.

In HS-TLE, the thalamus showed increased clustering coefficient in the region of the contralateral anterior thalamic nucleus. The thalamus has previously been demonstrated to show atrophy on volumetric MRI,35 hypometabolism on FDG-PET,36 and neuronal loss at postmortem37 in this population. The findings here further support anterior thalamic changes as being a core functional network abnormality of HS-TLE. That this cluster was detected contralateral to the sclerotic hippocampus is interesting but not immediately interpretable, except that some other studies have also shown bilateral36 or even predominantly contralateral thalamic involvement.38 In MRI-negative TLE, there is relative sparing of the thalamus, demonstrated both in this study and in previous work.36,37

We did not detect any increases in connectivity in MRI-negative TLE. Increased connectivity in HS-TLE, particularly in the affected hippocampus, appears to represent the core of the abnormal epileptic network. We speculate that the lack of such areas in the MRI-negative TLE group-level analysis could relate to spatial variability, if a functionally hypersynchronous epileptic focus was located at a different cortical region in each person. Alternatively, MRI-negative TLE may be a truly diffuse process, where decreased connectivity and network instability is present over an extended zone of temporal neocortex in the majority of cases.

The label of MRI-negative TLE is sometimes used to refer to a broad spectrum, from patients with few seizures and little certain localizing evidence to highly characterized patients seen in comprehensive epilepsy surgery centers. Here we focused on the latter and required multimodal concordant localization, to maximize accuracy in determining the seizure focus. Through selecting only medically refractory patients, the seizure frequency in MRI-negative TLE was well-matched to HS-TLE, ensuring that the findings are driven by the underlying pathology, rather than seizure severity or medication burden. Whether any connectivity changes are present in milder forms of MRI-negative TLE remains to be investigated.

A limitation of this study is that no simultaneous EEG was recorded during the fMRI acquisition. Although clinical EEG outside the scanner showed that interictal epileptiform discharges were infrequent in most participants and of similar frequency in both groups, it is not possible to assess here whether discharges might have some influence on the measured functional connectivity. Second, there was a relatively low rate of histologic confirmation in the patients with MRI-negative TLE, which largely reflects clinical concern regarding the lesser effectiveness of temporal lobectomy in this cohort5 and the neuropsychological deficits that may result.4 Finally, the fMRI graph theory metrics used here (at a spatial scale of millimeters) are yet to have specific neuronal correlates identified, and whether these measures have clinical utility at the individual patient level requires further validation.

Taken together, the findings in this study show a pattern of network abnormalities in MRI-negative TLE that are spatially distinct from those found in HS-TLE. MRI-negative TLE has impaired connectivity at the ipsilateral temporal neocortex, with relative sparing of mesial temporal structures. In contrast, HS-TLE shows a hypersynchronously active but functionally segregated hippocampus, with abnormal connectivity extending into the thalamus and ventromedial prefrontal cortex. This reinforces the view that MRI-negative TLE is not simply “mild HS,” but is rather a separate condition that involves distinct brain networks. This is reflected in the patterns of neuropsychological impairment in HS-TLE vs MRI-negative TLE, and the reported difference in clinical responses to temporal lobe surgery.

AUTHOR CONTRIBUTIONS

David N. Vaughan: study concept and design, acquisition of data, statistical analysis and interpretation of data, study coordination, drafting and revising the manuscript. Genevieve Rayner: acquisition of data, revising the manuscript for content. Chris Tailby: study design, data analysis and interpretation, revising the manuscript for content. Graeme D. Jackson: study concept and design, interpretation of data, revising the manuscript for content.

STUDY FUNDING

This study was supported by the National Health and Medical Research Council (NHMRC) of Australia (program grant 628952 and project grant 1081151), a computation grant from the Victorian Life Sciences Computation Initiative, and the Victorian Government Operational Infrastructure Support Program. Some participant scans were funded by an Austin Health Medical Research Foundation Grant awarded to G.R. and Professor Sarah Wilson at the University of Melbourne, and by a Rebecca L. Cooper Medical Research Foundation Grant awarded to C.T. D.N.V. was supported by an NHMRC Postgraduate Scholarship and a Windermere Foundation Doctoral Scholarship. G.D.J. is supported by an NHMRC practitioner fellowship (1060312).

DISCLOSURE

D. Vaughan, G. Rayner, and C. Tailby report no disclosures relevant to the manuscript. G. Jackson has received honoraria from UCB Pharma and receives royalties from Elsevier for the book Magnetic Resonance in Epilepsy. Go to Neurology.org for full disclosures.

ACKNOWLEDGMENT

The authors thank the participants for their involvement in the study, Mira Semmelroch and Susan Palmer for arranging recruitment and MRI scanning, and Richard Masterton for initial development of graph analysis software.

Footnotes

  • Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

  • Supplemental data at Neurology.org

  • Received March 23, 2016.
  • Accepted in final form July 14, 2016.

REFERENCES

  1. 1.↵
    1. LoPinto-Khoury C,
    2. Sperling MR,
    3. Skidmore C, et al
    . Surgical outcome in PET-positive, MRI-negative patients with temporal lobe epilepsy. Epilepsia 2012;53:342–348.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Carne RP,
    2. O'Brien TJ,
    3. Kilpatrick CJ, et al
    . MRI-negative PET-positive temporal lobe epilepsy: a distinct surgically remediable syndrome. Brain 2004;127:2276–2285.
    OpenUrlAbstract/FREE Full Text
  3. 3.↵
    1. Cohen-Gadol AA,
    2. Bradley CC,
    3. Williamson A, et al
    . Normal magnetic resonance imaging and medial temporal lobe epilepsy: the clinical syndrome of paradoxical temporal lobe epilepsy. J Neurosurg 2005;102:902–909.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Helmstaedter C,
    2. Petzold I,
    3. Bien CG
    . The cognitive consequence of resecting nonlesional tissues in epilepsy surgery: results from MRI- and histopathology-negative patients with temporal lobe epilepsy. Epilepsia 2011;52:1402–1408.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Téllez-Zenteno JF,
    2. Hernández Ronquillo L,
    3. Moien-Afshari F,
    4. Wiebe S
    . Surgical outcomes in lesional and non-lesional epilepsy: a systematic review and meta-analysis. Epilepsy Res 2010;89:310–318.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Kuba R,
    2. Tyrlíková I,
    3. Chrastina J, et al
    . “MRI-negative PET-positive” temporal lobe epilepsy: invasive EEG findings, histopathology, and postoperative outcomes. Epilepsy Behav 2011;22:537–541.
    OpenUrlPubMed
  7. 7.↵
    1. Ji GJ,
    2. Zhang Z,
    3. Zhang H, et al
    . Disrupted causal connectivity in mesial temporal lobe epilepsy. PLoS One 2013;8:e63183.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Pereira FR,
    2. Alessio A,
    3. Sercheli MS, et al
    . Asymmetrical hippocampal connectivity in mesial temporal lobe epilepsy: evidence from resting state fMRI. BMC Neurosci 2010;11:66.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Pittau F,
    2. Grova C,
    3. Moeller F,
    4. Dubeau F,
    5. Gotman J
    . Patterns of altered functional connectivity in mesial temporal lobe epilepsy. Epilepsia 2012;53:1013–1023.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Luo C,
    2. Qiu C,
    3. Guo Z, et al
    . Disrupted functional brain connectivity in partial epilepsy: a resting-state fMRI study. PLoS One 2012;7:e28196.
    OpenUrl
  11. 11.↵
    1. Voets NL,
    2. Beckmann CF,
    3. Cole DM,
    4. Hong S,
    5. Bernasconi A,
    6. Bernasconi N
    . Structural substrates for resting network disruption in temporal lobe epilepsy. Brain 2012;135:2350–2357.
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    1. Bettus G,
    2. Bartolomei F,
    3. Confort-Gouny S, et al
    . Role of resting state functional connectivity MRI in presurgical investigation of mesial temporal lobe epilepsy. J Neurol Neurosurg Psychiatry 2010;81:1147–1154.
    OpenUrlAbstract/FREE Full Text
  13. 13.↵
    1. Morgan VL,
    2. Abou-Khalil B,
    3. Rogers B
    . Evolution of functional connectivity of brain networks and their dynamic interaction in temporal lobe epilepsy. Brain Connect 2015;5:35–44.
    OpenUrlPubMed
  14. 14.↵
    1. Liao W,
    2. Zhang Z,
    3. Pan Z, et al
    . Altered functional connectivity and small-world in mesial temporal lobe epilepsy. PLoS One 2010;5:e8525.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Zhang X,
    2. Tokoglu F,
    3. Negishi M, et al
    . Social network theory applied to resting-state fMRI connectivity data in the identification of epilepsy networks with iterative feature selection. J Neurosci Methods 2011;199:129–139.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Doucet GE,
    2. Sharan A,
    3. Pustina D, et al
    . Early and late age of seizure onset have a differential impact on brain resting-state organization in temporal lobe epilepsy. Brain Topogr 2015;28:113–126.
    OpenUrlPubMed
  17. 17.↵
    1. Masterton RA,
    2. Carney PW,
    3. Jackson GD
    . Cortical and thalamic resting-state functional connectivity is altered in childhood absence epilepsy. Epilepsy Res 2012;99:327–334.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Hayasaka S,
    2. Laurienti PJ
    . Comparison of characteristics between region-and voxel-based network analyses in resting-state fMRI data. Neuroimage 2010;50:499.
    OpenUrlCrossRefPubMed
  19. 19.↵
    1. Jackson GD,
    2. Badawy RA
    . Selecting patients for epilepsy surgery: identifying a structural lesion. Epilepsy Behav 2011;20:182–189.
    OpenUrlCrossRefPubMed
  20. 20.↵
    1. Fischl B,
    2. Dale AM
    . Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci USA 2000;97:11050–11055.
    OpenUrlAbstract/FREE Full Text
  21. 21.↵
    1. Bhaganagarapu K,
    2. Jackson GD,
    3. Abbott DF
    . An automated method for identifying artifact in independent component analysis of resting-state fMRI. Front Hum Neurosci 2013;7:343.
    OpenUrlPubMed
  22. 22.↵
    1. Behzadi Y,
    2. Restom K,
    3. Liau J,
    4. Liu TT
    . A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 2007;37:90–101.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Lohmann G,
    2. Margulies DS,
    3. Horstmann A, et al
    . Eigenvector centrality mapping for analyzing connectivity patterns in fMRI data of the human brain. PLoS One 2010;5:e10232.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Hayasaka S,
    2. Phan KL,
    3. Liberzon I,
    4. Worsley KJ,
    5. Nichols TE
    . Nonstationary cluster-size inference with random field and permutation methods. Neuroimage 2004;22:676–687.
    OpenUrlCrossRefPubMed
  25. 25.↵
    1. Carne RP,
    2. Cook MJ,
    3. MacGregor LR,
    4. Kilpatrick CJ,
    5. Hicks RJ,
    6. O'Brien TJ
    . “Magnetic resonance imaging negative positron emission tomography positive” temporal lobe epilepsy: FDG-PET pattern differs from mesial temporal lobe epilepsy. Mol Imaging Biol 2006;9:32–42.
    OpenUrl
  26. 26.↵
    1. Liew CJ,
    2. Lim Y-M,
    3. Bonwetsch R, et al
    . 18F-FCWAY and 18F-FDG PET in MRI-negative temporal lobe epilepsy. Epilepsia 2009;50:234–239.
    OpenUrlPubMed
  27. 27.↵
    1. Hammers A,
    2. Koepp MJ,
    3. Hurlemann R, et al
    . Abnormalities of grey and white matter [11C]flumazenil binding in temporal lobe epilepsy with normal MRI. Brain 2002;125:2257–2271.
    OpenUrlAbstract/FREE Full Text
  28. 28.↵
    1. Bettus G,
    2. Wendling F,
    3. Guye M, et al
    . Enhanced EEG functional connectivity in mesial temporal lobe epilepsy. Epilepsy Res 2008;81:58–68.
    OpenUrlCrossRefPubMed
  29. 29.↵
    1. Haneef Z,
    2. Lenartowicz A,
    3. Yeh HJ,
    4. Engel J,
    5. Stern JM
    . Effect of lateralized temporal lobe epilepsy on the default mode network. Epilepsy Behav 2012;25:350–357.
    OpenUrlPubMed
  30. 30.↵
    1. Liao W,
    2. Zhang Z,
    3. Pan Z, et al
    . Default mode network abnormalities in mesial temporal lobe epilepsy: a study combining fMRI and DTI. Hum Brain Mapp 2011;32:883–895.
    OpenUrlCrossRefPubMed
  31. 31.↵
    1. Kim CH,
    2. Lee SA,
    3. Yoo HJ,
    4. Kang JK,
    5. Lee JK
    . Executive performance on the Wisconsin card Sorting test in mesial temporal lobe epilepsy. Eur Neurol 2007;57:39–46.
    OpenUrlPubMed
  32. 32.↵
    1. Lie CH,
    2. Specht K,
    3. Marshall JC,
    4. Fink GR
    . Using fMRI to decompose the neural processes underlying the Wisconsin Card Sorting Test. Neuroimage 2006;30:1038–1049.
    OpenUrlCrossRefPubMed
  33. 33.↵
    1. Li YH,
    2. Chiu MJ,
    3. Yeh ZT,
    4. Liou HH,
    5. Cheng TW,
    6. Hua MS
    . Theory of mind in patients with temporal lobe epilepsy. J Int Neuropsychol Soc 2013;19:594–600.
    OpenUrlPubMed
  34. 34.↵
    1. Leopold A,
    2. Krueger F,
    3. dal Monte O, et al
    . Damage to the left ventromedial prefrontal cortex impacts affective theory of mind. Soc Cogn Affect Neurosci 2012;7:871–880.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Bonilha L,
    2. Rorden C,
    3. Castellano G,
    4. Cendes F,
    5. Li LM
    . Voxel-based morphometry of the thalamus in patients with refractory medial temporal lobe epilepsy. Neuroimage 2005;25:1016–1021.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Khan N,
    2. Leonhard Leenders K,
    3. Hajek M,
    4. Maguire P,
    5. Missimer J,
    6. Gregor Wieser H
    . Thalamic glucose metabolism in temporal lobe epilepsy measured with 18F-FDG positron emission tomography (PET). Epilepsy Res 1997;28:233–243.
    OpenUrlCrossRefPubMed
  37. 37.↵
    1. Sinjab B,
    2. Martinian L,
    3. Sisodiya SM,
    4. Thom M
    . Regional thalamic neuropathology in patients with hippocampal sclerosis and epilepsy: a postmortem study. Epilepsia 2013;54:2125–2133.
    OpenUrlCrossRefPubMed
  38. 38.↵
    1. Barron DS,
    2. Fox PM,
    3. Laird AR,
    4. Robinson JL,
    5. Fox PT
    . Thalamic medial dorsal nucleus atrophy in medial temporal lobe epilepsy: a VBM meta-analysis. Neuroimage Clin 2013;2:25–32.
    OpenUrlCrossRef
View Abstract

Letters: Rapid online correspondence

No comments have been published for this article.
Comment

REQUIREMENTS

You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.

Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.

If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.

Submission specifications:

  • Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
  • Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
  • Submit only on articles published within 6 months of issue date.
  • Do not be redundant. Read any comments already posted on the article prior to submission.
  • Submitted comments are subject to editing and editor review prior to posting.

More guidelines and information on Disputes & Debates

Compose Comment

More information about text formats

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
NOTE: The first author must also be the corresponding author of the comment.
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g. [email protected]
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Publishing Agreement
NOTE: All authors, besides the first/corresponding author, must complete a separate Publishing Agreement Form and provide via email to the editorial office before comments can be posted.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

You May Also be Interested in

Back to top

More Online

CME Course

Topics Discussed

Alert Me

Recommended articles