Changes in extratemporal integrity and cognition in temporal lobe epilepsy: A diffusion tensor imaging study
Background: Diffusion tensor imaging (DTI) is a sensitive technique for studying cerebral white matter. Only a few studies have examined the association between changes in extratemporal integrity and cognition in temporal lobe epilepsy (TLE), especially in Chinese patients. Aim: We used DTI to characterize microstructural changes and investigate their associations with cognition in patients with temporal lobe epilepsy. Patients and Methods: We studied 27 adult patients with TLE and 21 healthy controls. A standardized neuropsychological evaluation and diffusion tensor imaging examination were conducted on each subject. Eight patients were excluded because T2-weighted magnetic resonance imaging (MRI) results showed visible lesions. Furthermore, we determined mean diffusivity (MD) and fractional anisotropy (FA) values in the different regions of interest - normal-appearing white matter (NAWM) in the frontal lobe white matter, the occipital lobe white matter, the corpus callosum, the internal capsules, the external capsules; and normal-appearing gray matter (NAGM) in the caudate nucleus head, the putamens and the thalami. These diffusion measurements were compared between the two groups, and we examined the correlations between DTI values and clinical characteristics. In addition, multiple linear regression analysis was used to study the association of DTI values with cognitive function. Results: Compared with normal controls, TLE patients demonstrated decreased FA in NAGM of both thalami and NAWM of the posterior limb of the left internal capsule (P<.01). In patients with temporal lobe epilepsy, right thalamus FA showed a tendency to correlate with age at seizure onset (ß=0.47, P=.045), and left thalamus MD showed a tendency to correlate with the duration of epilepsy (ß=0.54 P=.038). Patients with temporal lobe epilepsy showed significantly poorer performance on nearly all tasks concerning category fluency and other executive functions (P<.01). In patients with TLE, there was a positive correlation between category fluency scores and FA in the white matter of the left frontal lobe (ß=0.429, P=.041) and the right occipital lobe (ß=0.613, P=.001). Conclusions: Our results suggested that the thalamus might be a potentially important extratemporal structure involved in temporal lobe epilepsy. Moreover, a longer duration of epilepsy or an earlier age at onset may result in more abnormalities in the thalamus. Patients with temporal lobe epilepsy showed significantly poorer performance on nearly all tasks concerning category fluency and other executive functions. Our results showed that frontal lobe white matter contributed to category fluency impairment in patients with TLE, but other areas might also contribute to these impairments.
Keywords: Diffusion tensor imaging, executive function, temporal lobe epilepsy, thalamus
Temporal lobe epilepsy (TLE) is the most common form of epilepsy.  Several studies have shown extratemporal abnormalities of white and gray matter in patients with TLE when compared with controls. , In addition, specific cognitive impairments have been found to be associated with temporal lobe epilepsy, , such as deficits in declarative memory, language functions or face/name processing. In recent years, several studies have described executive dysfunction in patients with temporal lobe epilepsy.  These executive functions have been suggested to be a flexible resource related to working memory, cognitive control,  general intelligence  ; or a toolbox of dissociable mechanisms with some degree of neuroanatomical localization within the prefrontal cortex.  Moreover, investigators are uncovering other features of executive functions, which include aspects of emotional regulation, social cognition, memory retrieval, and a complex specification of functions of the rostral prefrontal cortex.  Although the prefrontal cortex is a vital component of the circuitry of executive functions, it is clear that the posterior cortical regions and subcortical structures collaborate with prefrontal cortex to mediate successful executive processing. 
Diffusion tensor imaging (DTI) is a relatively new, noninvasive imaging technique that allows for the detection of orientation and integrity of white matter fiber bundles.  There are several metrics that can be calculated from DTI scans, but the two most common are mean diffusivity (MD) and fractional anisotropy (FA). Mean diffusivity provides a measure of nondirectional diffusion. In contrast to MD, FA provides a measure of directional diffusion. Thus FA and MD can be used as noninvasive measures of the general integrity of the tissue and the organization of remaining white matter fibers. Demyelination and neuronal loss with axonal degeneration would be expected to result in decreased FA and increased MD.
Recently, DTI and tractography have been applied to the study of epilepsy and have demonstrated some diffusion changes in gray and white matter tissue. ,,,, Subcortical structures, such as thalamus,  hippocampus and amygdala,  ipsilateral to the seizure focus have shown a general decrease in FA and an increase in MD. In 2010, Bhardwaj et al. studied 6 pediatric patients with intractable focal epilepsy and found an aberrant tractography pathway traversing through the external capsule connecting two distant foci of epileptiform activity.
Diffusion tensor imaging may provide new insights into the mechanisms behind cognitive dysfunction. Changes in the normal appearance of white matter may also contribute to cognitive function. Huang et al. used the DTI technique to characterize microstructural white matter changes and their associations with cognitive dysfunction in Alzheimer's disease (AD) and mild cognitive impairment (MCI). Compared with normal controls, their results showed that Alzheimer's disease patients had decreased FA and increased radial diffusivity (DR ) in the normal-appearing white matter of the temporal, parietal and frontal lobes, as well as decreased axial diffusivity (DA) in normal-appearing white matter of the temporal lobe. Patients with mild cognitive impairment also showed decreased FA and decreased DA in normal-appearing white matter of the temporal lobe, with decreased FA and increased DR in normal-appearing white matter of the parietal lobe.
There are few studies about the relationship between DTI and cognitive impairment in patients with epilepsy. In 2010, Riley et al. studied 12 temporal lobe epilepsy patients and 10 age-matched healthy controls. They found that the mean FA in the anterior temporal lobe and mesial temporal lobe positively correlated with delayed memory and immediate memory, respectively. There are also few studies about the changes in extratemporal integrity and their association with cognition in temporal lobe epilepsy, especially in Chinese patients. We used the DTI technique to characterize microstructural changes and their associations with cognition in Chinese patients with temporal lobe epilepsy.
This was a prospective study in the Chinese population. We studied 27 adult consecutive outpatients with temporal lobe epilepsy and 21 healthy controls. Patients with TLE were recruited from the Epilepsy Center of PLA General Hospital. A standardized neuropsychological evaluation and diffusion tensor imaging examination were conducted on each subject. Eight patients were excluded because visible lesions were found by T2-weighted MRI. Mean diffusivity and fractional anisotropy values of the white/ gray matter areas of interest were studied in 19 patients of temporal lobe epilepsy, which included 4 patients with unilateral mesial temporal sclerosis.
Temporal lobe epilepsy group
To be included in the study, patients needed to have interictal or ictal electroencephalogram (EEG) evidence that clearly indicated onset in the temporal lobe , and a seizure semiology consistent with onset in the temporal lobe. Patients were excluded if they had an epileptic focus or radiological evidence of dysfunction outside of the temporal region, or a history of a major psychiatric disorder (e.g., depression or psychosis). Background information about the patients was obtained from their clinical records and interviews with their parents, relatives and friends.
With regard to seizure type, eight patients had simple partial seizures, and seven patients had been seizure free for more than one year. The other patients still had occasional seizures (1 or 2 per month), but they had not presented with a seizure in the seven days prior to MRI scan. All but two patients were taking antiepileptic medication at the time of testing. 12 patients were treated with a single drug (carbamazepine [CBZ], phenytoin [PHT], phenobarbital [PB], valproate [VPA] or topiramate [TPM]), whereas five patients received more than one drug. Lamotrigine and topiramate were used most frequently as adjunct drugs in these cases.
Healthy control group
The healthy control group consisted of 21 subjects. The healthy control group and the temporal lobe epilepsy group were matched for age, sex and educational history. The mean age of the patients was 31.45 years (range, 16-56) in the group with temporal lobe epilepsy (TLE group) and 32.33 years (range, 17-54) in the control group (P=.80). 11 patients were male in both the TLE group (58%) and the control group (52%) (P=.76). The mean duration of full-time education was 10.8 years in the TLE group and 11.9 years in the control group (P=.24).
In this study, we used a set of brief neuropsychological tests which were sensitive to various aspects of executive functioning and relatively easy to be performed by patients.
The subject was asked to produce as many words as possible belonging to a particular category (animal names) in 1 minute.
This is a measure that reflects both performance intelligence quotient (PIQ) and executive functions, which were assessed using the Chinese adaptation of the Wechsler adult intelligence scale (WAIS-R ). Each digit (1-9) was ascribed a unique symbol, which was presented in the form of a key. In this test, subjects were presented with a series of digits and were asked to fill in the corresponding symbols. Subjects were asked to fill in as many consecutive spaces as possible in 90 seconds. This test required subjects to switch between rules for each digit; therefore, it required mental flexibility and paralleled other set-shifting tasks.
Digit span task
The digit span task includes both forwards and backwards conditions, where a subject is given a number sequence and asked to repeat it in forward or reverse order until he/ she fails two times at a given level. In the present study, the results of this test were assessed using the Chinese adaptation of the WAIS-R.
Trail-making test (TMT, part A)
In TMT part A, the subject is asked to draw a line to join numbered points scattered randomly over a sheet of paper in numerical order. We evaluated the time taken to perform this test. This test requires cognitive tracking, psychomotor speed, short-term memory and cognitive flexibility.
The Stroop test assesses a more complex ability of response inhibition. The four subtests consist of the following tasks: (1) naming the color of colored dots; (2) naming the color words that are printed in black ink; (3) naming the color words that are written in a different, non-black, color (for example, naming "red" for the word "red" written in green ink); and (4) naming the color in which the color words are written (for example, naming "green" for the word "red" written in green ink). We evaluated the response time and the number of errors of the four subtasks. We used the fourth subtest to measure an executive function to reflect the response inhibition.
Diffusion tensor imaging study
Diffusion tensor imaging was performed in all patients after they had been seizure free for at least seven days. This approach was intended to ensure measurements from a baseline state as much as possible because prior studies indicated that seizures resulting from status epilepticus may affect diffusivity measurements.
MR imaging acquisition parameters
All MR imaging data were obtained with a 3.0-T MR imaging system (Signa Excite, GE Medical Systems, USA). The machine was equipped with a self-shielding gradient (40 mT/m, maximum gradient strength; 150 mT/m/s, slew rate) and manufacturer-supplied 8-channel phase array head coil. Each patient underwent an MR imaging examination that consisted of a spin-echo T1-weighted sequence in the axial plane, a fast spin-echo T2-weighted sequence in the axial plane, and a fluid-attenuated inversion recovery (FLAIR) T2-weighted sequence in the coronal plane. The diffusion tensor imaging was performed in the axial plane parallel to the anterior commissure-posterior commissure (AC-PC) line by using a spin-echo, echo-planar imaging sequence with an array spatial sensitivity encoding technique. The DTI parameters were, repetition time (TR)=8000 ms; echo time (TE)=84 ms (minimum); number of excitations (NEX)=2; field of view (FOV)=240Χ240 mm 2 ; slice thickness=3.0 mm;; matrix=128Χ128; diffusion gradient encoding in 15 directions; 2 diffusion gradient fields (b=0, and b=1,000 s/mm 2 ); total sections, 32-37; and total imaging time, 4 minutes 32 seconds).
After processing, the entire image analysis was performed on a scanner console using a subprogram of the Functool image analysis software (Advantage Windows 4.2, GE Medical Systems, Buc, France). The region of interest (ROI) was placed at b=0 on the image (T2-weighted image) and automatically transferred to a co-registered MD map and FA map constructed from DTI. The MD and FA values were then calculated using the Functool image analysis software. The MD and FA values were measured in normal-appearing white matter (NAWM) in the frontal lobe white matter, the occipital lobe white matter, the corpus callosum, the internal capsules, the external capsules; and normal-appearing gray matter (NAGM) in the caudate nucleus head, the putamens and the thalami. Each ROI was a 20-mm 2 circle (except in the external capsule, which was oval).
Differences in demographic and cognitive functions between the two groups were compared by t tests. The DTI values were compared between the two groups, and significant differences were further examined for correlations with the clinical characteristics using bivariate correlations (Pearson's) and a multiple linear regression model . We also explored the correlations between DTI values and cognitive function in patients with temporal lobe epilepsy using bivariate correlations (Pearson's) and a multiple linear regression model.
MD/FA values for different regions of interest in the temporal lobe epilepsy and healthy control groups
Compared with normal controls, TLE patients demonstrated decreased FA in both thalami normal-appearing gray matter (NAGM) and normal-appearing white matter (NAWM) of the posterior limb of the left internal capsule (right thalamus: t=5.809, P<.001; left thalamus: t=5.893, P<.001; left internal capsule: t=2.836, P=.008). These results are shown in [Table 1] and [Table 2] and [Figure 1] and [Figure 2].
We tested the correlation between DTI values of both thalami and the clinical characteristics (e.g., the duration of epilepsy, age at seizure onset, attack frequency, and seizure-free time before DTI examination). We found a positive correlation between FA values in the right thalamus and seizure-free time (r=0.54, P=.02), and seizure-onset age (r=0.59, P=.01). In addition, we found a negative correlation between FA values in the right thalamus and the duration of epilepsy (r= −0.51, P=.04), which was demonstrated by a bivariate correlation (Pearson's) analysis. A positive correlation between MD values in the left thalamus and the duration of epilepsy (r=0.56, P=.02) was also demonstrated by bivariate correlation analysis.
In patients with temporal lobe epilepsy, increased FA values in the right thalamus were associated with the age at seizure onset (ß=0.47, P=.045), and increased MD values in left thalamus were associated with duration of epilepsy (ß=0.54, P=.038). These variables were tested by the multiple linear regression model, which included duration of epilepsy, age at seizure onset, attack frequency, and seizure-free time before DTI examination.
Results of the neuropsychological tests in the temporal lobe epilepsy and healthy control groups
Patients with temporal lobe epilepsy showed significantly poorer performance on nearly all tasks concerning category fluency and other executive functions. Control group produced significantly more words in 1 minute than the TLE group (P<.01). The scores of the digit span and digit symbol tests for the TLE group were significantly lower than the control group (P<.01). In addition, the TLE group needed significantly more time to accomplish the trail-making test and the Stroop test (P<.01, [Table 3]).
Correlations between DTI values and cognitive function in patients with temporal lobe epilepsy
In patients with TLE, there was a negative correlation between the scores of the category fluency and MD in the left caudate (r= −0.56, P=.04), the right putamen (r= −0.58, P=.04), the left putamen (r= −0.58, P=.04) and the right thalamus (r= −0.64, P=.02). There was a positive correlation between the scores of the category fluency and FA in white matter of the left frontal lobe (r=0.43, P=.04) and the left occipital lobe (r=0.47, P=.02), which was demonstrated by bivariate correlation analysis. Bivariate correlation analysis of patients in the TLE group also revealed a negative correlation between the scores of the digit span task and MD in the splenium of the corpus callosum (r= −0.58, P=.04) and the left putamen (r=−0.59, P=.04). Moreover, there was a positive correlation between the scores of the trail-making test and MD in the posterior limb of the right internal capsule (r=0.65, P=.02) and in the posterior limb of the left internal capsule (r=0.59, P=.03). These results are shown in [Table 4] and [Table 5].
Age at seizure onset, sex, education level, duration of epilepsy, attack frequency and seizure-free time before DTI examination were included in the multiple linear regression analysis. In patients with TLE, increased FA values in NAWM of the left frontal lobe and the right occipital lobe were associated with the scores of the category fluency (the left frontal lobe: ß=0.426, P=.043; the right occipital lobe: ß=0.593, P=.003).
Our results showed significantly reduced FA values in both thalami and the anterior limb of the left internal capsule in temporal lobe epilepsy patients. A longer duration of TLE was associated with higher MD values in the left thalamus, and a younger age at onset of epilepsy was associated with lower FA values in the right thalamus. Thalamofrontal abnormalities have been identified in chronic primary generalized epilepsy.  The role of synchronized thalamocortical oscillations has been well defined in animal models of absence seizures. During partial seizures, the role of such interactions has not been extensively studied. A potentially important extratemporal structure that could be involved in the early stages of temporal lobe epilepsy is the midline thalamus. Quantitative MRI studies have shown a decrease in the volume of the ipsilateral thalamus in patients with TLE. , Positron emission tomography (PET) studies have also demonstrated a reduction in metabolic activity extending from the temporal lobe to the ipsilateral thalamus.  Our results supported that thalamus may be a potentially important extratemporal structure involved in temporal lobe epilepsy, and a longer duration of epilepsy or a younger age at onset may result in more abnormalities in the thalamus.
Patients with temporal lobe epilepsy showed significantly poorer performance on nearly all tasks concerning category fluency and other executive functions. These results confirmed those of previous studies, which have shown that these functions are frequently affected in patients with temporal lobe epilepsy.  The reason for impairment in these executive functions in TLE is unclear, but proposed mechanisms include propagation of seizures from the temporal to the frontal lobes, extratemporal pathology not detected on standard MRI, and/ or an intrinsic role of the hippocampus in executive functioning.
In this study, positive correlations between the scores of the category fluency and FA in white matter of the left frontal lobe and right occipital lobe were demonstrated to be statistically significant by a multiple linear regression model. The results also suggested that frontal lobe white matter contributed to category fluency impairment in patients with TLE; however, more areas might also be involved in category fluency. Previous functional neuroimaging studies have also documented prefrontal cortex (PFC) as a vital component of executive functions. Phelps and colleagues  found left prefrontal cortex activation during a letter fluency task in healthy subjects using functional MRI (fMRI). A similar result was reported in a positron emission tomography study, in which verbal fluency was shown to activate a similar region in healthy middle-aged adults. 
We would like to acknowledge various limitations in the current study. We used DTI methodology because of its increasing availability and use in mainstream clinical settings. However, other more sophisticated methods, such as diffusion spectrum imaging (DSI), are now available, and they can overcome current limitations of DTI. In addition, we only scanned patients who had been seizure free for 7 days, and it is possible that the data would have been different if we had not avoided postictal changes. Nevertheless, postictal diffusivity changes are complex and dynamic, and timing after the seizure may be critical. Additional research is needed to fully evaluate how long postictal changes persist and which DTI measurements they affect. Because the fibers were manually drawn, there was also the possibility of intra-operator error. In addition, the power of our data could be further improved by increasing the number of patients and controls in our study. We intend to study the influence of the antiepileptic medication in our subsequent continuation study.
Our results confirmed those of the other studies, suggesting that the thalamus is a potentially important extratemporal structure involved in temporal lobe epilepsy. Moreover, a longer duration of epilepsy or a younger age at onset may result in more abnormalities in the thalamus. Compared with healthy controls, patients with TLE tended to have poorer executive functioning outcomes. Although multiple brain areas may play a role in category fluency, our results showed that the frontal white matter contributed to category fluency impairment in patients with TLE. We intend to follow up with general cognitive-function tests, increased patient numbers, and longitudinal studies to evaluate the sequential changes in MD/ FA in patients with temporal lobe epilepsy. In addition, future studies will determine the influence of the antiepileptic medication and whether MD/ FA changes precede cognitive deterioration, as well as where these changes occur.
No sources of funding to carry out this study.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]