|Home | About | Journals | Submit | Contact Us | Français|
Depression in patients with temporal lobe epilepsy (TLE) is highly prevalent and carries significant morbidity and mortality. Its neural basis is poorly understood. We used quantitative, surface-based MRI analysis to correlate brain morphometry with severity of depressive symptoms in 38 TLE patients and 45 controls. Increasing severity of depressive symptoms was associated with orbitofrontal cortex (OFC) thinning in controls, but with OFC thickening in TLE patients. These results demonstrate distinct neuroanatomical substrates for depression with and without TLE, and suggest a unique role for OFC – a limbic region for emotional processing strongly interconnected with medial temporal structures – in TLE-related depressive symptoms.
Depression is a significant yet underappreciated problem affecting up to 55% of patients with treatment resistant epilepsy . Mood more powerfully predicts quality of life than seizure severity , and epilepsy patients have a five-fold increased risk for suicide . Although depressed mood may result partly from an expected psychological reaction to lifestyle limitations and stigma, there is evidence that depression in people with epilepsy differs phenomenologically and perhaps neurobiologically from depression in people without epilepsy.
Phenomenologically, depression in people with versus without epilepsy has been characterized by greater temporal variability in symptoms related to peri-ictal mood changes , prominent anhedonia , and decreased sadness . These differences mean that depressive symptoms in epilepsy, while significant and requiring treatment, may not qualify for a diagnosis of major depression according to standard DSM-IV  criteria [6–8].
Neurobiologically, potential differences underlying depression in people with versus without epilepsy has not to our knowledge been investigated directly; the small number of prior studies examining the neuroanatomical correlates of depression in epilepsy have not included normal controls. These studies have shown that in patients with temporal lobe epilepsy (TLE) – the most common treatment resistant adult epilepsy – major depression as well as subsyndromic depressive symptoms are associated with hippocampal atrophy [9, 10] and relative amygdalar enlargement [11–13]. In patients with depression but without epilepsy, a similar neuroanatomic pattern of hippocampal atrophy and relative amygdalar enlargement has been demonstrated [14, 15]. However, the majority of recent research attention has been focused on depression-related abnormalities in frontal regions involved in emotional regulation such as orbitofrontal cortex (OFC) [16, 17], anterior cingulate cortex  and dorsolateral prefrontal cortex . Because studies of epilepsy-related depression have focused primarily on the temporal lobe, it is not known whether these frontal or other extra-temporal regions are also involved in epilepsy-related depression.
To address these issues, we utilized unbiased whole-cortex MRI morphometric assessment to identify brain regions associated with depressive symptoms in patients with TLE and control subjects without TLE. To assess depressive symptoms, we used the Beck Depression InventoryII . The BDI-II is self-report inventory assessing symptoms of depression which has been validated for use in medically healthy people as well as patients with epilepsy .
36 patients (mean age 37; 22 female) with mesial TLE were recruited from the Epilepsy Centers at New York University (NYU; n=16) and the University of California, San Diego (UCSD; n=20). Based on clinical information (MRI, video-EEG and intracranial EEG when available) 22 patients were categorized as having left TLE, 12 as having right TLE, and 2 as having bilateral TLE.
45 subjects (mean age: 40; 19 female) were recruited through online advertisement at both institutions (NYU n=24; UCSD n=21). Controls were medically healthy with no history of epilepsy. Their responses to a questionnaire indicated no prior psychiatric diagnosis or treatment.
Participants provided informed consent to participate in this study, which was approved by each institution’s IRB.
The BDI-II  was administered as part of a neuropsychological assessment. The BDI-II is a 21-item self-report inventory used to assess affective, somatic, and cognitive symptoms of depression that has been validated for use in patients with epilepsy.
Imaging was performed at either the NYU Center for Brain Imaging on a 3T Siemens Allegra scanner or at the UCSD Radiology Imaging Laboratory on a GE 1.5T EXCITE scanner with an 8-channel phased-array head coil. Acquisition parameters were optimized for gray/white contrast. Acquisition included two T1-weighted images (NYU: TE=3.25ms, TR=2530ms, TI=1.1ms, flip angle=7°, FOV=256mm, voxel size=1×1×1.33mm; UCSD: TE=3.8ms, TR=10.7ms, TI=1000 ms, flip angle=8°, FOV=240mm, voxel size=1.25×1.25×1.2mm) which were averaged to improve signal-to-noise.
Cortical thickness was measured using FreeSurfer 5.0 (http://surfer.nmr.mgh.harvard.edu) which employs the following procedures, described in detail elsewhere: white matter segmentation; tessellation of the gray/white boundary; inflation of the folded surface tessellation; correction of topological defects; measuring the distance between each point on the white and pial surfaces; smoothing (15mm FWHM); and averaging across participants using a spherical technique which matches homologous regions while minimizing distortions. For each hemisphere, a general linear model estimated the effects of BDI-II score on cortical thickness at each vertex along the cortical surface. Subject age and scanning site were included as covariates of no interest. The correlation between BDI-II score and cortical thickness was assessed within and between subject groups. Results were mapped onto the average brain and considered significant at p<.05 corrected for multiple comparisons using cluster-based thresholding .
BDI-II scores ranged from 0–34 in controls and 0–36 in TLE patients, indicating a spectrum of depressive symptoms in both TLE patients and controls. TLE patients were more depressed than controls (TLE mean=11.7; control mean=5.6; t(65)=3.1, p<.005). There was no difference in severity of depression between patients with right and left seizure foci (right TLE mean=9.8; left TLE mean=10.4; t(32)=.87, p>.5)
In controls, BDI-II scores correlated positively with cortical thickness in right dorsolateral prefrontal cortex and negatively with left lateral OFC and right precuneus, as shown in Figure 1. In TLE patients, BDI-II scores correlated positively with cortical thickness in left lateral OFC and right fusiform gyrus, and negatively with a small region of right superior parietal cortex.
In a direct comparison between TLE patients and controls, the BDI-thickness correlation differed significantly only in bilateral lateral OFC, as shown in Figure 2. The correlation was positive in TLE and negative in controls.
We show for the first time that the neuroanatomical correlates of depressive symptoms in patients with epilepsy differ from those in controls without epilepsy, and include opposite findings in OFC, a limbic component of prefrontal cortex involved in the subjective experience of emotional and social stimuli .
Within-group comparison showed that in controls, higher levels of self-reported depressive symptoms were associated with thinning of left lateral OFC. These results are consistent with prior studies demonstrating OFC abnormalities in major depression [16, 17], and extend the association between OFC abnormality and depression to subjects with subclinical depressive symptoms (since no enrolled subjects had a current or past diagnosis of depression). In contrast, patients with TLE had thickening of left lateral OFC in association with greater self-reported depressive symptoms. See supplemental materials for discussion of brain regions other than OFC which showed significant BDI-thickness correlations in either TLE patients or controls.
Between-group comparison confirmed that BDI-thickness correlations in left OFC differed significantly between TLE patients and controls. In addition, this analysis revealed similar between-group differences in right OFC, suggesting bilateral OFC involvement in TLE-related depressive symptoms. However, conclusions about laterality effects are limited by the underrepresentation of patients with right TLE in our sample (12/35), precluding separate analyses of right and left TLE patients .
OFC has bidirectional connections with sensory association cortices and temporal lobe regions . In particular, lateral OFC has robust, direct connections with the amygdala , suggesting a mechanism by which mesial temporal epileptic activity could affect lateral OFC structure. In depressed patients with TLE, FDG PET studies found OFC hypometabolism [27, 28]. Together with our demonstration of OFC thickening in association with depression, this suggests that the cellular components contributing to OFC thickening are unlikely to be normally-functioning neurons or glia. Future studies, perhaps using MR spectroscopy, may better define the nature of this thickening.
While a potential limitation of this study is the use of two different MRI scanners, prior multi-site studies of cortical thickness have shown negligible impact of initial scan acquisition . In addition, we scanned roughly equal numbers of patients and controls on each scanner, and included site as a covariate in analyses, making it unlikely that findings relate to local hardware or other factors.
It is a limitation of this study that we did not perform structured diagnostic interviews to diagnose major depression. Our findings can be considered relevant to understanding the neural basis only of depressive symptoms, not major depression. However, it is important to note that DSM-IV  was not designed for the diagnosis of psychiatric disorders in neurologic patients, and we believe use of a symptom measure such as the BDI-II is appropriate in patients with epilepsy who often do not meet DSM criteria for major depression [1, 6–8]. Use of the BDI-II, an instrument validated in people both with and without epilepsy , facilitates comparison between these two group. Our use of a symptom measure rather than an all-or-none diagnosis is also in accord with the idea that depression may be best investigated using a dimensional rather than a categorical approach . Still, it will be important for future studies to utilize a measure of symptom severity like the BDI-II as well as a structured clinical interview to fully assess depression, and to include subjects both with and without a formal diagnosis of major depression and with and without epilepsy. It will also be important to investigate different components of depression (e.g. affective, somatic and cognitive) to learn if they have specific neural correlates in subjects with and without epilepsy; while the BDI-II has been used for this purpose, it has not been independently validated.
Our findings support the idea that depression associated with temporal lobe epilepsy is not only phenomenologically distinct from depression unassociated with epilepsy, but is also neurobiologically distinct, thus highlighting the need for specifically-designed diagnostic instruments , and perhaps, specific antidepressant treatments different from those used in patients without epilepsy.
Supported by FACES (Finding a Cure for Epilepsy and Seizures) and NIH grants NS18741 (E.H.), NS44623 (E.H.) and NS056091 (C.M.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.