If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
First Department of Neurology, St. Anne’s University Hospital and Medical Faculty of Masaryk University, Brno, Czech RepublicDepartment of Neurology, University Hospital Ostrava, Czech Republic
First Department of Neurology, St. Anne’s University Hospital and Medical Faculty of Masaryk University, Brno, Czech RepublicUniversity of Minnesota: Department of Neurology, School of Medicine, University of Minnesota, Minneapolis, MN, USA
First Department of Neurology, St. Anne’s University Hospital and Medical Faculty of Masaryk University, Brno, Czech RepublicBehavioral and Social Neuroscience Research Group, CEITEC (Central European Institute of Technology), Masaryk University, Brno, Czech Republic
The character of changes differs significantly within the cerebellar substructures.
•
Clinical correlations of cerebellar atrophy were examined.
•
Cerebellar atrophy is associated with worse outcome of temporal lobe resection.
Abstract
Purpose
To evaluate cerebellar volume changes in temporal lobe epilepsy (TLE) patients in greater detail. We aimed to determine which discrete substructures significantly differ in patients with TLE compared to controls and the nature of this difference. Correlations with age at epilepsy onset, epilepsy duration, seizure frequency, and total number of antiepileptic drugs (AED) in the patient’s history were studied. We analyzed the potential association between cerebellar atrophy and epilepsy surgery outcome.
Methods
Study participants were 36 TLE patients; 22 hippocampal sclerosis (HS) only and 38 healthy controls. All patients later underwent temporal lobe resection. All subjects were examined using 1.5T MRI. Cerebellar volume was adjusted for total intracranial volume, age, and gender, and measured using voxel-based morphometry. Cerebellar substructures were defined using the AAL atlas. Data processing was performed automatically. Separate analyses for HS only subset were performed.
Results
Total cerebellar gray matter volume (GMV) appeared non-significantly smaller in epilepsy patients. Within the substructures, the GMV of the selected vermian segments were significantly larger in patients. The GMV of the whole cerebellum and of all individual cerebellar substructures non-significantly decreased with increasing complex partial seizure frequency and total number of AEDs in the patient’s history. Total cerebellar GMV was significantly smaller in patients with persistent seizures after epilepsy surgery than in seizure-free patients.
Conclusion
Cerebellar atrophy is a complex phenomenon, the character of changes differs significantly within the cerebellar substructures.
Total cerebellar GMV reduction is associated with worse outcome of temporal lobe resection.
The role of the cerebellum goes beyond its effect on motor functions. The cerebellum is an important part of the networks involved in cognitive processing, including memory, language, and emotions [
]. The role of the cerebellum in the genesis of seizures is also a subject of discussion. Cerebellar dysfunction has been associated with disinhibition of seizure activity in the cerebral cortex [
Voxel based morphometry of grey matter abnormalities in patients with medically intractable temporal lobe epilepsy: effects of side of seizure onset and epilepsy duration.
] Segmented gray and white matter volumes within the cerebellar hemispheres or specific cerebellar subregions were studied with inconsistent findings (McDonald: [
] non-differentiated cerebellar volume (gray and white matter together) divided into the midsagittal vermian area, the anterior lobule area, the posterior superior lobule area, and the posterior inferior lobule area; and Oyegbile: [
] gray and white matter volumes of the right and left anterior lobes, the superior posterior lobes, the inferior posterior lobes, and the corpus medullare.)
There is conformity in GMV reduction of the whole cerebellum and the posterior cerebellar lobe. There were varying findings for the anterior cerebellar lobe: Oyegbile found an even larger volume of GMV; [
In the present study, we evaluate cerebellar volume changes in TLE patients and healthy controls in greater detail using the Automated Anatomical Labeling (AAL) atlas (Fig. 1). We aim to determine the character of changes within cerebellar substructures. With regard to previous findings, we hypothesized that whereas total cerebellar volume is reduced in TLE patients, the character of changes might differ within the cerebellar substructures. We expected the difference to depend on the patient age at epilepsy onset, epilepsy duration, number of anti-epileptic drugs (AEDs) in the patient’s history, and complex partial seizure (CPS) frequency. Detailed knowledge of the nature of morphological changes could help determine their underlying mechanism.
Fig. 1Cerebellar division according to Automated Anatomical Labeling atlas.
We focus on the association between cerebellar atrophy and epilepsy surgery outcome as it could contribute to the optimal selection of patients for resective and palliative epilepsy surgery.
2. Methods
2.1 Subjects
Our sample comprised 36 consecutive patients, all with medically intractable focal epilepsies, (25 females, 11 males) ranging in age from 17.7 to 56.9 years (mean age of 36.9 years; median 39 years). All the patients fulfilled the diagnostic criteria for TLE according the ILAE criteria [
]. All patients had a history of at least three AEDs (mean 5.8, max. 13); detailed figures are in Table 1. All the patients had been routinely investigated prior to epilepsy surgery, including long-term semi-invasive video-EEG monitoring (using sphenoidal electrodes), high resolution magnetic resonance imaging (MRI), and neuropsychological testing. The diagnosis of TLE was based on a concordance of history data, ictal and interictal EEG findings, stereoelectroencephalography in selected cases, ictal semiology, neuropsychology, and neuroimaging findings. All patients had been seizure free for ≥24 h before the preoperative MRI investigation. All the patients underwent anteromedial temporal resection (AMTR). Hippocampal sclerosis was found in 24 cases (in two patients combined with focal cortical dysplasia), focal cortical dysplasia alone in five patients and seven patients were histopathologically negative. The follow-up interval after epilepsy surgery was at least 12 months (mean 61.1 months, max. 117 months).
Table 1Main demographic characteristics of patients. Differences between Engel I and Engel II–IV were tested using Fisher’s exact test and Mann-Whitney There should be a space between words "demographic characteristics" in Table 1 caption.U test, as appropriate. Starred figures (*) are statistically significant (p < 0.05).
Patients (n = 36)
Engel I (n = 27)
Engel II-IV (n = 9)
p-value
Gender: Female/Male
25/11
19/8
6/3
1.000
Age, median (min; max) [years]
39.0 (17.7; 56.9)
39.2 (17.7; 56.9)
35.0 (19.3; 46.9)
0.387
Age at epilepsy onset, median (min; max) [years]
14.5 (0.8; 43.0)
10.0 (0.8; 43.0)
17.0 (12.0; 27.0)
0.043*
Epilepsy duration, median (min; max) [years]
19.6 (1.0; 41.1)
28.5 (4.1; 41.1)
12.6 (1.0; 31.3)
0.032*
Localization of seizure onset zone: left-sided/right-sided/bitemporal
22/12/2
18/8/1
4/4/1
0.340
Follow-up period, median (min; max) [months]
62.5 (12.0; 117.0)
64.0 (12.0; 110.0)
51.0 (20.0; 117.0)
0.914
Total number of AED in medical history, median (min; max)
5.0 (3.0; 13.0)
5.0 (3.0; 13.0)
5.0 (4.0; 11.0)
0.615
Complex partial seizure frequency preoperatively, median (min; max) [monthly]
After surgical resection, 27 patients (75%) were rated as Engel I (seizure free), nine patients (25%) were Engel II–IV (non-seizure free). In the hippocampal sclerosis only (HS) patients, 21 were rated as Engel I, one patient was Engel II–IV. There was no change in Engel classification into group Engel I or group Engel II–IV during the overall follow-up period compared to outcome in the first year after surgery. Significant differences between the Engel groups were detected in epilepsy duration (Mann-Whitney U test, p = 0.032), which was longer in Engel I group; and in patient age at epilepsy onset (Mann-Whitney U test, p = 0.043). There was no difference in patient age (Mann-Whitney U test, p = 0.387), gender (Fisher’s exact test, p = 1), localization of seizure onset zone (Fisher’s exact test, p = 0.34), number of AED in medical history (Mann-Whitney U test, p = 0.615), duration of postoperative follow-up (Mann-Whitney U test, p = 0.914), or in CPS frequency (Mann-Whitney U test, p = 0.14). Table 1 provides detailed demographic and medical history data.
The control group consisted of 38 healthy subjects (24 females, 14 males) ranging in age from 18.3 to 56.4 years (mean age of 36.6 years; median 38.2 years). No significant differences in age (Mann-Whitney U test, p = 0.953) or sex (Fisher’s exact test, p = 0.628) between the subgroups of patients and controls was detected. The majority of the healthy subjects in the control group were volunteers from the professional sector; no history of neurological or psychiatric disease was presented in any controls.
Informed consent was obtained from each participant after all of the procedures were fully explained. This study was approved by the St. Anne’s University Hospital Research Ethics Committee. All patients signed an informed consent form.
2.2 MR image acquisition, preprocessing, and statistical analyses
MR examinations were performed on a 1.5T scanner (Siemens Magnetom Symphony, Erlangen, Germany) using a multichannel head coil. The MRI protocol for voxel-based morphometry included 3D T1-weighted magnetization prepared rapid gradient echo (MPRAGE) sequence with TR = 1.7 s, TE = 3.93 ms, TI = 1.1 s, FA = 15°, 512 × 512 matrix size, FOV 246 × 246 mm, and 160 sagittal slices with slice thickness = 1.17 mm. MRI studies were visually evaluated and those with any artifacts affecting cerebellum were excluded from further analyses (41 MRI studies of patients and 41 MRI studies of controls were initially available). Anatomical MRI data were analyzed using SPM12 (Wellcome Department of Cognitive Neurology, www.fil.ion.ucl.ac.uk) with its internal toolbox DARTEL running in MATLAB 8.4 (The MathWorks, Natick, MA, USA). Data of all the participants were segmented into six tissue types (gray matter segments, white matter segments, cerebrospinal fluid (CSF) segment, and three other segments for the rest of the head). The gray/white/CSF matter segments were used to estimate total intracranial volume (TIV). The gray and white matter segments were used to precisely register all subjects’ data into the MNI standard stereotactic space using DARTEL. The gray matter segments were modulated by the Jacobian of the nonlinear transformation of coordinates and smoothed by a spatial filter with a Gaussian kernel and full-width-half-maxima of 8 mm. The resulting gray matter images have the character of local GMV distribution. Before ensuing analyses, the data were corrected for variability in TIV (normalization of the data by TIV) and corrected for nuisance effects of age and gender using linear regression. To compare average volumes of cerebellar substructures between groups, the cerebellum was parcellated according to the Automated Anatomical Labelling (AAL) atlas provided with SPM12. The volume of a specific substructure was computed as the sum of the voxels belonging to that substructure. No corrections were made to the automated delineation of the cerebellar substructures.
Statistical analyses were performed separately for a non-differentiated subset of TLE patients and for a subset of patients with hippocampal sclerosis only. The significance of differences in volumes between patients and controls was assessed using the Mann-Whitney U Test; the level of statistical significance was set to p < 0.0052 after False Discovery Rate (FDR) correction for multiple testing for non-differentiated group of TLE patients and to p < 0.0185 for a subset of patients with HS. (FDR was used to eliminate the risk of obtaining p values less than 0.05 purely by chance.) The Mann-Whitney nonparametric test was used because the data did not follow the normal distribution (according to Shapiro-Wilk test).
The local GMV data were correlated with the patient age at onset, disease duration, the number of AED, and CPS frequency to test for the relationship between local GMV alterations and disease characteristics and development for both subsets separately. We computed Spearman’s correlation coefficients and evaluated statistical significance by setting the threshold to p < 0.035 (p < 0.005 for HS subset) for correlation with the patient age at onset and p < 0.02 (p < 0.002 for HS subset) otherwise after FDR correction. The significance of the difference in volumes in non-differentiated TLE subset between the group without seizures (Engel I patients) and the group with persistent seizures (Engel II–IV patients) after epilepsy surgery was assessed using the Mann-Whitney U Test; the level of statistical significance was set to p < 0.0054 after FDR correction for multiple comparisons. Evaluation of the difference in the HS subset was not possible as there was only one Engel II–IV patient.
3. Results
3.1 Total cerebellar GMV and the GMV of cerebellar substructures
Total cerebellar GMV appears marginally and non-significantly smaller in non-differentiated subset of TLE patients (p = 0.6246) and non-significantly larger in HS subset (p = 0.348) when compared to healthy controls. An average total cerebellar GMV was calculated to be 53.1 ml (median 53.8 ml, min. 35.3 ml, max. 65.5 ml) in TLE patients, 55.5 ml (median 54.4 ml, min. 47.4 ml, max. 64.6 ml) in HS patients and 53.9 ml (median 54.4 ml, min. 45.2 ml, max. 61.7 ml) in healthy controls.
The GMV of cerebellar substructures in patients differs heterogeneously from the identical substructures in controls. In all the vermian substructures (vermis 1–10) of non-differentiated TLE patients, the average volumes of gray matter of the substructures in patients are larger than that of the corresponding substructures in controls. Individual proportions are given in Fig. 2. Reflecting FDR correction for multiple comparisons, there were significant differences in the GMVs of only two frontmost (vermis 1,2; vermis 3) and one rearmost structure (vermis 10) of the cerebellar vermis according to the AAL atlas. Similarly, the remaining part of the anterior lobe vermis was larger (vermis 4,5), above the level of statistical significance after FDR (p = 0.016). P-values of all comparisons are listed in Table 2. The site of maximal difference in volumes between non-differentiated TLE patients and controls is depicted in Fig. 3. The highlighted areas are of larger volumes in patients with epilepsy.
Fig 2Percent difference in gray matter volumes of individual cerebellar substructures of non-differentiated temporal lobe epilepsy patients and hippocampal sclerosis only patients compared to controls. Asterisks indicate significant difference.
Table 2p-values of comparisons of cerebellar gray matter volumes [median (min.; max.)] in ml of individual cerebellar substructures between patients (All TLE patients and Hippocampal sclerosis only patients) and controls. Starred figures (*) are statistically significant after FDR correction for multiple testing.
Fig. 3The site of maximal difference in volumes between non-differentiated temporal lobe epilepsy patients and controls. Vermis of anterior lobe outlined. Highlighted areas are of larger volumes in patients with epilepsy.
Compared to non-differentiated TLE patients, there was higher amount of significantly different substructures in the HS subset; substructures were of larger volume in epileptics. Individual proportions are given in Fig. 2. P-values of all comparisons are listed in Table 2.
3.2 Correlation of gray matter volume with epilepsy duration and age at epilepsy onset
In our data, there is no significant finding regarding a correlation of GMV with epilepsy duration in the group of all TLE patients. However, in case of HS only patients, there is significant association between cerebellum 3 and epilepsy duration. Specifically, cerebellum 3 is significantly larger for patients with larger epilepsy duration.
For the age at epilepsy onset, the GMV of the whole cerebellum and of all individual cerebellar substructures (except of vermis 9 and 10) statistically significantly decreases with increasing age at onset of epilepsy in the group of all TLE patients. Correlation coefficients are listed in Supporting Table 1. In the HS subgroup, the correlations show similar trend but they are statistically insignificant after FDR correction (see Supporting Table 2).
3.3 Correlation between gray matter volume and number of AEDs in the medical history
We did not find a significant correlation between the GMV of the whole cerebellum or of individual cerebellar substructures and the number of AEDs in the patient’s medical history. There was a trend to decreasing volume in the whole cerebellar GMV as well as in the majority of individual cerebellar substructures. Correlation coefficients are listed in Supporting Table 1.
We did not find a significant correlation in HS subset as well. Correlation coefficients are listed in Supporting Table 2.
3.4 Correlation between gray matter volume and complex partial seizure frequency
We found a uniform decrease of GMV of all individual cerebellar substructures as well as of the whole cerebellum with the increasing number of CPS. There was no significant correlation after FDR. Correlation coefficients are listed in Supporting Table 1.
We did not find a significant correlation in HS subset as well. Correlation coefficients are listed in Supporting Table 2.
3.5 Cerebellar atrophy and epilepsy surgery outcome
The total cerebellar GMV is significantly smaller in patients who are not seizure free after surgery (Engel II–IV, 9 patients) than in patients who are seizure free after epilepsy surgery (Engel I, 27 patients). The average total cerebellar GMV was calculated to be 47.9 ml (median 48.4 ml) in the Engel II–IV group and 54.9 ml (median 54.3 ml) in the Engel I group. The P-value of this comparison is 0.0046. For individual cerebellar substructures, there was a statistically significant difference in cerebellum 8 (p = 0.0053) according to the AAL atlas. This observation is illustrated in Supporting Fig. 1. P-values of all comparisons are listed in Table 3.
Table 3p-values of comparisons of cerebellar gray matter volumes (GMV) [ml] of individual cerebellar substructures between patients with persistent seizures after epilepsy surgery (Engel II–IV) and seizure-free patients (Engel I). Starred figures (*) are statistically significant after FDR correction for multiple testing.
]. The lower density of Purkinje cells was confirmed as a result of seizure activity in an animal experiment. There was no relation to type or number of seizures [
]. A similar finding was made in the cerebellar biopsies of patients with epilepsy at the time of cerebellar electrode installation. The specimens from patients with epilepsy showed gliosis and significantly lower Purkinje cell densities than the specimens from nonepileptic control patients [
]. Some authors have suggested that cerebellar volume reduction is present at epilepsy onset, and might not be a result of epilepsy but a preceding factor [
MRI studies explored patterns of cerebellar changes in a variety of clinical frames. Coan et al. described more pronounced cerebellar gray matter atrophy in TLE patients with hippocampal sclerosis than in patients with MRI non-lesional TLE [
Voxel based morphometry of grey matter abnormalities in patients with medically intractable temporal lobe epilepsy: effects of side of seizure onset and epilepsy duration.
] found the reduction in the left cerebellum in patients with left-sided seizure onset while right-sided seizure onset patients showed a bilateral cerebellar reduction. Oyegbile et al. [
] recently explored the cerebellum divided into three parts − the anterior, superior, and inferior posterior lobe. As in previous studies, they found total cerebellar volume reduction and no lateralized cerebellar abnormalities when comparing left TLE and right TLE groups. They also described significant anterior lobe GMV increase, while volumes of the superior and inferior posterior lobes were reduced. This phenomenon was evident in a subgroup of earlier onset and longer epilepsy duration.
The main benefit of our study is in the more detailed exploration of the morphological changes of cerebellar gray matter in TLE patients. As far as we are aware, this is the first study to explore cerebellar changes within 17 distinct substructures according to the standard AAL atlas. We observed a heterogeneous character of volume changes. An average total cerebellar GMV was only non-significantly smaller in our limited set of patients and even non-significantly larger in the HS subset (however medians in HS subset and controls are the same, so there is no real difference between them). A possible explanation for this non-significance is that there was just a slight volume difference between TLE patients and controls. In a larger group of patients, even so subtle difference could be significant.
We are not exclusively addressing different degrees of hypotrophy, but we have detected some structures of even larger volumes in TLE patients compared to controls. Oyegbile et al. [
] observed significantly increased GMVs of the anterior lobe; we proved significantly larger GMVs in two out of three substructures of the anterior lobe vermis in non-diferentiated TLE patients and even more significantly larger substructures, mostly of vermis in HS subset; see Table 3. Vermian hypertrophy in TLE probably reflects a distinct pathogenetic mechanism, when compared to volume decreases observed repeatedly in the posterior lobes [
], the role of which is still unclear. One recent study suggests that cerebellar vermis targeted optogenetic intervention inhibits spontaneous hippocampal seizures in a mouse model of TLE [
] observed that the stimulation of the vermis during the early stages of focal seizure usually resulted in spike inhibition or reduced spike frequency and that ablation of the vermis facilitated seizure activity. However, seizure activity suppression might also be achieved by the stimulation of other cerebellar structures. Electrical stimulation of the anterior lobe appears to be more efficient than stimulation of the posterior lobe [
]. There is increasing knowledge of complex cerebro-cerebellar connectivity. Tracers in primates proved the existence of topographically arranged closed cerebellar loops with the motor and prefrontal cortex [
]. There is a dynamic interplay between the cerebellum and the hippocampus during temporal lobe seizures. Hippocampal epileptiform activity modulates cerebellar activity and vice versa; excitation or inhibition of the cerebellum could significantly decrease hippocampal seizure duration [
] revealed diverse cerebro-verminal connections, targeting the motor, sensory, and association areas of the cortex. However, precise information on a temporo-vermian connection is missing.
In relation to epilepsy duration, we found no significant changes. For the age at onset of epilepsy, the volume of all individual cerebellar substructures except vermis 9 and 10 as well as of the whole cerebellum significantly decreases with increasing age at onset of epilepsy in the group of all TLE patients. In the HS subset the significance was not expressed, however the trend was the same. Most previous authors [
], found an increase of cerebellar atrophy in connection with epilepsy duration and the age at onset of epilepsy. We were dealing with patients who underwent epilepsy surgery, not the general TLE population. Three of the referenced studies of patients who underwent epilepsy surgery also concerned the correlation between cerebellar volume and epilepsy duration: Specht et al. [
Voxel based morphometry of grey matter abnormalities in patients with medically intractable temporal lobe epilepsy: effects of side of seizure onset and epilepsy duration.
We have limited information about the exact lifetime frequency of generalized tonic-clonic seizures, which are relatively infrequent in our patient group. CPS are significantly more frequent and better documented. Therefore, we focus on evaluating the relationship between the GMV of cerebellar substructures and CPS frequency. The tendency to decreasing cerebellar GMV with increasing seizure frequency and the same but less pronounced trend for the increasing number of AEDs in medical history is consistent with the expectations and literature [
Significantly smaller total cerebellar GMV was found in our group of patients with less favorable epilepsy surgery outcome. The association between the presence of cerebellar atrophy and temporal lobe resection outcome in patients with epilepsy was examined with differing conclusions, perhaps because of the different methods of cerebellar atrophy assessment and different etiology. Sandok et al. [
] did not prove an association between epilepsy surgery outcome and cerebellar atrophy. Cerebellar atrophy was determined there as being present when the mean measured volume was smaller than two standard deviations from the mean value found in control subjects. In contrast, Specht et al. [
] found a significantly greater incidence of cerebellar atrophy in TLE patients with worse temporal lobe resection outcome. Even then, cerebellar atrophy was determined as a delimited category, but the determination was based on a visual analysis. Finally, Bohnen et al. [
] found a trend for patients without excellent outcomes to have smaller relative cerebellar volumes than the patients with excellent outcome. Our results provide a dissolution of previous contradictions with potentially great practical importance.
The detection of significantly worse epilepsy surgery outcome in patients with mesial TLE and more pronounced cerebellar GMV reduction might be helpful in temporal lobe resection outcome prediction.
Future studies on a larger and more homogeneous sample of patients could better examine individual changes of the cerebellar substructures. Further research should also explore the interconnections of the most differing regions of cerebellum with the cerebral cortical and subcortical regions. This knowledge might create a background for determining new efficacious targets for cerebellar stimulation.
5. Limitations
We studied a limited number of subjects and the patient group was not perfectly homogeneous. The patients varied particularly in their anti-epileptic treatment, using different combinations of AEDs with varying lengths of drug administration and lifetime sum and severity of generalized seizures.
Acknowledgement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Appendix A. Supplementary data
The following are Supplementary data to this article:
The total cerebellar GMV is significantly smaller in patients who are not seizure free after surgery (Engel II-IV) than in patients who are seizure free after epilepsy surgery (Engel I).
Voxel based morphometry of grey matter abnormalities in patients with medically intractable temporal lobe epilepsy: effects of side of seizure onset and epilepsy duration.
30655-6&id=gr1.jpg){kind=link}
30655-6&id=gr2.jpg){kind=link}
30655-6&id=gr3.jpg){kind=link}