- Email: [email protected]
Patterns of brain activity in patients with epilepsy and depression
Seizure 1999; 8: 390–397
Article No. seiz.1999.0320, available online at http://www.idealibrary.com on
Patterns of brain activity in patients with epilepsy and depression † §¶
‡k
§
‡
‡
H. A. RING , P. D. ACTON , D. SCULL , D. C. COSTA , S. GACINOVIK & M. R. GACINOVIK
§
†
Department of Psychological Medicine, St Bartholomew’s and the Royal London School of Medicine, London E1, UK; ‡ Institute of Nuclear Medicine, University College London Medical School, London W1, UK; § Institute of Neurology, University College London Medical School, London WC1, UK Correspondence to: Dr H. A. Ring, Academic Department of Psychological Medicine, 3rd Floor, Alexandra Wing, Royal London Hospital, Whitechapel Road, London E1 1BB, UK
Depression is a recognized feature of epilepsy. This study tested the hypothesis that depression arising in patients with epilepsy would be associated with decreased activity in brain regions previously demonstrated to be hypoperfused both in primary depression and in depression secondary to movement disorders. Two groups of patients with temporal lobe epilepsy were studied, one of which also met DSM IV criteria for a major depressive episode. All underwent a SPECT scan using the blood flow marker, 99m Tc-HMPAO. An automated voxel-based analysis demonstrated no regions of relatively decreased activity in the depressed compared with the non-depressed patients. Sites of relative hyperactivity in the depressed group were concentrated in the left hemisphere, particularly in dorsolateral prefrontal cortex, striatum, thalamus and temporo-parietal regions. Comparison of these data with normal population data revealed that in the depressed epilepsy group regional activities were within the normal range whilst corresponding results from the non-depressed group were below it. Depressed patients with epilepsy have cerebral regions with greater perfusion than non-depressed people with epilepsy, although they are not hyperperfused compared with normals. Our results suggest that depression in people with epilepsy may arise from a mechanism which differs from that underlying the development of depression in patients with movement disorders. c 1999 BEA Trading Ltd
Key words: epilepsy; depression; SPECT; cerebral blood flow.
INTRODUCTION Depression is a clinically important concomitant of epilepsy and is seen relatively frequently in this population1, 2 . A study of the prevalence of psychiatric morbidity in people with epilepsy in the community recorded neurotic depression in 22%3 . The importance of understanding depression in epilepsy is highlighted when the frequency of suicide in this population is considered. Barraclough (1981)4 , reviewing 11 previous studies, reported a suicide rate five times that in the general population. Although depression may occur in reaction to any adverse life event or stress, including epilepsy, there is evidence that in the case of epilepsy it is likely that there is a pathophysiological relationship between § E-mail: [email protected] ¶ Present address: Department of Radiology, University of Pennsyl-
vania, Philadelphia, PA 19104, USA. 1059–1311/99/070390 + 08
$12.00/0
the epilepsy and the development of depression in at least a proportion of the patients in which these conditions coexist. Depression is more common in patients with epilepsy than in non-epileptic patients with comparable degress of disability5 , and within patients suffering from epilepsy there is evidence that depression is reported more frequently in those with temporal lobe foci than in those with generalized epilepsy or extra-temporal foci5, 6 . Indeed, in a study of more than 600 patients with temporal lobe epilepsy, 19% were found to have clinically apparent anxiety whilst 11% were depressed7 . Another strand of evidence supporting a biological basis to depression in epilepsy is provided by accounts of the development of episodes of depression in association with periods of decreased seizure frequency and relative or absolute reduction in epileptiform abnormalities in the electroencephalogram; so-called ‘forced normalization’. It has been suggested that the development c 1999 BEA Trading Ltd
Patterns of brain activity in patients with epilepsy and depression
of psychopathology in these circumstances relates to ongoing electrical epileptiform discharges spreading through brain regions different from those associated with the development of seizures8 . There is evidence that the manifestation of depression secondary to some other neurological disorders may also be related to the pathophysiology of those disorders. In particular, patients with idiopathic Parkinson’s disease (IPD) have an increased rate of depression9 . In this group of patients it has been noted that development of depression may predate the onset of obvious motor symptoms10 and that severity of depression does not correlate with specific measures of motor impairment11 . It has been suggested that depression occurring in patients with basal ganglia disorders (IPD, Huntington’s disease, ischaemic lesions of the striatum) may arise out of a common mechanism involving frontal subcortical and paralimbic systems12 . Disturbances in similar regions have been identified as being associated with the manifestation of states of primary depression, with functional neuroimaging studies of such patients demonstrating focal disturbances concentrated in frontal lobe regions13, 14 . In general, regions of dorsolateral and dorsomedial prefrontal cortex tend to be hypoperfused whilst more ventral and posterior regions of the brain have been reported as being more active in depressed subjects15, 16 . A direct comparison of regional cerebral blood flow patterns in patients with depression occurring in people with IPD to patients with primary depression observed bilateral decreases in rCBF in anteromedial regions of the medial frontal cortex and the cingulate cortex (Brodmann’s areas 9 and 32) in the depressed IPD group compared with those with IPD alone and compared with normal controls17 . This regional disturbance overlapped that observed in patients with primary depression. These observations suggest that medial prefrontal cortex is a common area of neural dysfunction in the manifestation of both primary depression and depression in IPD. There is already some evidence that depression in patients with temporal lobe epilepsy (TLE) is associated with disturbed frontal lobe function. It has been reported18, 19 that in patients with left TLE there is a significant correlation between the presence of a self-reported dysphoric mood state and the degree of frontal lobe dysfunction, measured using a frontal lobe cognitive task (the Wisconsin Card Sorting Test). We have previously reported that in non-depressed patients with left TLE, higher scores on the Beck Depression Inventory were associated with relatively reduced activity in bilateral frontal and right temporal regions20 . The aim of the current study was to further explore the proposal that different patient groups who share
391
a psychopathological state (i.e. depression), but who otherwise have different primary diagnoses (i.e. Major Affective Disorder, IPD, TLE), will demonstrate similar functional disturbances relating to their depression. Specifically, we wished to test the hypothesis that depression meeting formal diagnostic criteria arising in patients with TLE would be associated with evidence of decreased activity in prefrontal cortical regions that have previously been demonstrated to be hypoperfused both in patients with primary depression and those with motor disorders. MATERIALS AND METHODS Study design Two groups of patients were studied. Both groups had previously been diagnosed as having TLE. The experimental group were ascertained on the basis of meeting additional criteria for the diagnosis of depression. The control group were selected from patients with TLE but no evidence of depression. Subjects were initially screened for suitability for the study and informed consent was obtained. On the day of their study each subject received a semistructured clinical interview lasting about 45 minutes which included an assessment of depressive state. Subjects then underwent a resting state SPECT scan lasting 30 minutes. The study was approved by the local research ethics committee and the Administration of Radioactive Substances Advisory Committee (ARSAC) of the UK Department of Health.
Subject selection and assessment Ten subjects were recruited into each group from specialist epilepsy outpatient clinics. All subjects had clear evidence of TLE, based on EEG and structural imaging investigations already performed as part of their clinical management. The same diagnostic criteria were employed with respect to depression as were used by Ring et al. (1994)17 . Hence as well as TLE, the experimental group had depression meeting DSM IIIR (and DSM IV) diagnostic criteria for a major depressive episode21 . The severity of mood disturbance was also measured using the 17-item Hamilton Depression Scale. Only subjects scoring 19 or more were included in the depressed group. Non-depressed subjects had no significant depressive symptoms elicited on clinical assessment. No subjects with any evidence of a progressive neurodegenerative condition were included.
392
Scanning protocol Each subject underwent a single SPECT scan. They were studied in an awake resting state with eyes closed. The blood flow marker employed was 99m TcHMPAO, available commercially from Amersham International as Ceretec. Subjects received a dose of approximately 500 MBq, administered through a vein in the antecubital fossa. Scanning was performed using a triple-headed IGE Neurocam SPECT scanner (GE Medical Systems, Milwaukee, WI, USA), with general purpose collimators. In air the reconstructed image resolution of this system is 10.7 mm full width at half maximum22 . Scanning was for 40 s per projection for 128 projections, acquired into a 64 × 64 matrix, giving a total duration of scanning of approximately 30 min.
Image analysis Images were reconstructed using filtered backprojection, with a Hanning filter with a cut-off frequency of 1.0 cycles/cm and corrected for attenuation (Chang type = 0.12 cm−1 ). Reconstructed images consisted of a 64 × 64 matrix, with a slice thickness of 4 mm, and an in-plane pixel size of 4 mm. Two image analysis techniques were employed; a manual region of interest (ROI) approach and an automated voxel-based comparison. To perform the ROI analysis, after reorientation into transverse (parallel to the inferior frontal–superior occipital plane), sagittal and coronal slices, square ROIs of 4×4 pixels were placed in the cortical and subcortical grey matter structures to obtain average counts per pixel for frontal, parietal and temporal cortices, caudate, thalamus and cerebellum in each hemisphere. At least three contiguous slices were sampled using as many ROIs as were needed to cover each structure. Average counts per pixel in each region were divided by the counts in the cerebellar hemisphere with higher average counts in order to obtain relative perfusion ratios for these brain regions. In the first stage of the automated voxel-based comparison, images were coregistered to each other, using a standard HMPAO template as the baseline image (Nuclear Diagnostics, Sweden). This template consisted of the mean image from 10 coregistered HMPAO scans taken previously of a group of normal, healthy volunteers. All images were registered using linear transformations (i.e. translation, rotation and scaling, with no image warping), minimizing the absolute difference between images23 . After co-registration with each other, the significance of differences between the two groups of sub-
H. A. Ring et al.
jects was assessed. In the absence of absolute quantification, each image was normalized to the mean counts per pixel in the cerebellum, defined by placing a standard region-of-interest template over the coregistered scans. Unpaired t-tests were performed at every voxel, and corrected for non-independent multiple comparisons, using the theory of random Gaussian fields [Statistical Parametric Mapping (SPM)]24 . Parametric images of statistically significant differences between groups were generated, thresholded at P < 0.05 (cluster-level), and overlaid on the standard template image.
RESULTS Clinical data A total of 19 patients were included in the final analysis of the data. Variables describing the subjects and their epilepsy are listed in Table 1. One of the depressed patients was excluded from the final analysis because examination of the SPECT image demonstrated extensive rCBF anomalies. No patients experienced a seizure during scanning and none had reported any seizures in the 24 hours preceding scanning. Table 1: Clinical variables describing the depressed and non-depressed subject groups.
Age (SD) Sex (male : female) Hamilton score (SD) Laterality of focusa Seizure frequency/month mean, (range) Duration of epilepsy in years (SD) Number of anticonvulsants mean, (range)
Depressed
Non-depressed
33 (7.1) 3:6 21 (2) 6 = left, 3 = right
31 (6.8) 6:4 4 (2) 7 = left, 3 = right
6, (1–12) 20 (8.6)
3, (1–5) 22 (6.2)
2, (1–3)
2, (1–3)
a If bilateral foci were present then the laterality of the greatest
disturbance was recorded.
All patients were diagnosed to be suffering from complex partial seizures with a temporal lobe focus. Three of the depressed subjects and one of the nondepressed subjects demonstrated bilateral epileptiform abnormalities on interictal scalp EEG recording. A total of seven subjects had in the past had temporal lobe resections as attempted treatment for their epilepsy. Two of the depressed group and two of the controls had undergone left temporal lobectomy whilst two of the depressed group and one of the controls had undergone right temporal lobectomy. All the patients were taking between one and three antiepileptic drugs. The mean number of agents prescribed to both groups was two and the most frequently prescribed drug was carbamazepine. In addi-
Patterns of brain activity in patients with epilepsy and depression
393
tion, subjects were taking a range of first and second line drugs with no systematic difference in regimens between the groups. Four of the depressed group were taking antidepressant medication at the time of the study. Considering the phenomenology of the depression occurring in the patients with epilepsy and depression, all these subjects met DSM IV criteria for a major depressive episode. The duration of their depression was of at least three weeks and had persisted beyond the immediate post-ictal period in all these subjects. In addition, four of the depressed group had a past history of depressive episodes whilst one of the non-depressed group had previously suffered with panic attacks.
Imaging data Examination of the results of the SPM analysis reveal that although there were several regions at which there was greater activity in the depressed than in the non-depressed subjects, there were no regions of relatively decreased activity in the depressed group. The sites of relative hyperactivity in the depressed group are displayed in Fig. 1. The threshold for display is P < 0.05 and the image is scaled as a P-map with the scale running from P < 0.05 to P < 0.001. It may be seen that the sites of significantly increased activity are concentrated in the left hemisphere, particularly in the left dorsolateral prefrontal cortex in the region of the inferior frontal gyrus including BA 44 and 45 and extending as far anteriorly as the more lateral area of BA 46, the left insula, the left striatum and thalamus, the left superior temporal gyrus (BA 22), the left angular gyrus (BA 39) and the left inferior parietal lobule (BA 40). More rostrally there is also some bilateral relative hyperactivity in the depressed group in the rostral anterior cingulate (BA 24), middle frontal and precentral gyrii (BA 8, 9, 6) and supplementary motor area (BA 6). In addition, there is relative hyperactivity in the region of the left hippocampus in the depressed group. Although it is not possible to obtain absolute quantification of blood flow from the SPECT methodology used in this study, it is possible to gain a semiquantitative estimate of how regional activities in the two groups compare with activities in normal subjects (Table 2). This was achieved by comparing activity in regions of interest (ROIs) containing the areas identified in the SPM analysis as being significantly different between the two groups with activities previously measured from a normal control data set of similar ROIs in a group of 40 normal subjects (age range; 21– 59, sex ratio; 26 males to 14 females). It is noted that, as also observed in the SPM analysis, there is relatively greater activity in frontal, parietal and thalamic
Fig. 1: Sites of relative hyperactivity in the depressed group. The threshold for display is P < 0.05 and the image is scaled as a P-map with the scale running from P < 0.05 (red) to P < 0.001 (yellow). In the image the left side of the brain is presented on the right.
regions in the depressed than the non-depressed subjects. However, it is also observed that nevertheless the mean levels of activity in the depressed group are not, over each region as a whole, above the upper limit of the normal range observed in a normal control population. It is also seen that in the non-depressed epilepsy group regional activities in left frontal and parietal regions are below the lower levels of the normal range. Hence it appears that the depressed patients with epilepsy are less hypoperfused than the non-depressed subjects with respect to a normal population (but are not hyperperfused compared with the normals). Table 2: Region of interest analysis contrasting activitya in areas identified as significantly different in the SPM analysis in the depressed and non-depressed subjects with epilepsy to ROI activity in a group of 40 healthy control subjects. Region
Epilepsydepressed
Epilepsynondepressed
Pb
Normal rangec
left frontal left parietal left thalamus
0.77 0.81 0.91
0.70 0.75 0.84
0.005 0.02 <0.001
0.71–0.81 0.77–0.93 0.79–1.03
a Activity is calculated as a ratio in each subject of counts in
ROI/counts in cerebellum. b P calculated using one-way ANOVA. c Normal range defined as mean +/− two standard deviations for ratios calculated as above in a group of 40 normal controls. SPM = Statistical Parametric Mapping, ROI = region of interest.
DISCUSSION The aim of this study was to explore the hypothesis that depression occurring in association with epilepsy
394
would be associated with prefrontal cortical hypoperfusion similar to that which has been reported in patients with primary depression and in patients with depression associated with Parkinson’s disease. Although our patients with epilepsy and depression did not display hyperperfusion compared with a normal sample, nevertheless there were no sites of hypoactivity relative to the non-depressed epilepsy group. However, widespread regions of relative hyperactivity in the depressed compared with the non-depressed subjects were found across left frontal, temporal, parietal and subcortical structures. Hence contrasting the results of the two epilepsy groups, our results do not support the hypothesis that all secondary depression is associated with prefrontal cortical hypoperfusion in depressed subjects relative to a non-depressed comparison group matched for neurological disease. Two functional imaging studies have previously reported on patterns of cerebral activity in patients with epilepsy and varying degrees of depression20, 25 . Neither of these studies investigated patients with a specific diagnosis of depression meeting diagnostic criteria. Rather, both examined patients unselected with respect to mood, and then commented on the correlations between patterns of cerebral activity and scores on a self-rated depression questionnaire. Both studies found that patients with higher depression scores had evidence of decreased frontal activity. Bromfield et al. (1992)25 noted bilateral inferior frontal hypometabolism in five patients with left temporal lobe epilepsy compared with five non-depressed controls with epilepsy, whilst Schmitz et al. (1997)20 noted a significant relationship between increased scores on the Beck Depression Inventory and decreased activity in bilateral frontopolar and dorsolateral prefrontal regions. Although the regions reported by Bromfield et al. (1992)25 are inferior to those identified by Schmitz and colleagues, nevertheless, all these areas have previously been reported as being underactive in patients with primary depressive illnesses. However, it may be argued that the depression reported in the patients in the present study was qualitatively different from that observed in the former two studies. In those studies patients reported symptoms of depression but did not meet diagnostic criteria for depression. It may be that they were manifesting symptoms of mild to moderate depression in reaction to their life circumstances, whilst the depressed subjects in the current study were suffering from significant depressive states arising in relation to the pathophysiology of their epilepsy. It has been proposed that depression, both primary and that arising secondary to a movement disorder, involves disturbed function in a network of cortical and subcortical sites including dorsolateral prefrontal, anterior cingulate and inferior parietal cortices and the basal ganglia as well as the ventral insula and the
H. A. Ring et al.
hypothalamic-pituitary axis15 . In the majority of studies of patients with depression, both primary and that arising secondarily to movement disorders, activity in these regions, particularly the more dorsal regions, has been found to be reduced13, 14, 17, 26 . Hence the results of the current study, in which many of these same areas were more perfused in the depressed than in the non-depressed epilepsy subjects, suggests that similar functional circuits may be involved in the manifestation of depression in epilepsy but that they are dysfunctional by virtue of being more active, rather than less active. Although the majority of imaging studies of depressed subjects in a resting state have reported evidence of decreased activity in prefrontal regions, several studies exploring the functional correlates of induced sadness in normal volunteers, often females, have found a different pattern of results27–30 . In those studies, by using various mood induction methods and PET measures of rCBF, patterns of cerebral activity associated with induced states of brief but significant sadness have been identified in otherwise euthymic subjects. Although an induced state of sadness in normal volunteers is self-evidently very different from depression occurring in people with epilepsy, there are some aspects of depression as it can develop in those with epilepsy which suggest that consideration of states of induced sadness may be relevant. In both post-ictal depressions and the depression that may develop in the context of forced normalization, the symptoms of depression arise over a short time, are of relatively short duration (in the order of hours of weeks), and tend to be associated with an active sense of sadness rather than the emptiness and flattening of affect that are more likely to be a feature of primary depressive states. Patients suffering from post-ictal depressions have reported that even though their mood was not particularly low prior to a seizure, and even though they know that they are very likely to have more seizures and are not too distressed by this knowledge, that following a seizure their mood changes rapidly and profoundly and that they feel that this is happening beyond their control, as depressive thoughts and feelings overwhelm them31 . These episodes of post-ictal depression may develop repeatedly in people who then make a good recovery from their depressive state. As well as these similarities to artifically induced mood states, there are phenomenological differences between interictal depressive states and primary depression. The most prominent mood disorders of epilepsy have been described as rapid shifting moods with predominant depression and episodic anger32 . It has also been reported that patients with interictal depression have high state anxiety scores33 and that they have more psychotic traits than a compara-
Patterns of brain activity in patients with epilepsy and depression
bly depressed group of patients with non-neurological chronic illness5 . Betts (1981)34 has noted that interictal depressions in people with epilepsy tend to be of sudden onset and departure with fluctuations in severity whilst present. Although the patients with epilepsy and depression in the current study were a relatively heterogeneous group, whose depression was not limited to the post-ictal period or to a process of forced normalization, nevertheless in the majority subjective clinical observations suggested that their depressive states were modulated to an extent by their seizures. Hence overall, there are a number of similarities between states of depression arising in people with epilepsy and states of transient sadness induced in normal volunteers. In studies of negative mood induction in normals it is observed that the negative mood states have been predominantly associated with areas of cerebral hyperactivity and with fewer decreases in activity27–30 . The regions reported as demonstrating increased activity in those studies overlap with the regions identified in the current study as being relatively more active in the patients with depression and epilepsy than in those with epilepsy alone. Apart from the work by Lane et al. (1997)30 who report bilateral changes, the reports note a preponderance of left-sided activation. George et al. (1995)28 report, amongst other findings, activation in left thalamus and striatum and Lane et al. (1997)30 also found increased activity in the striatum, more so on the left, in states of induced sadness, regardless of whether the emotion was precipitated by watching a film or recalled from memory. Pardo et al. (1993)27 , George et al. (1995)28 and Baker et al. (1997)29 describe increased activation in regions of left inferior dorsolateral prefrontal cortex close to areas activated in the current study. In addition, Baker et al. (1997)29 also reported increased rCBF associated with a state of induced sadness in left middle frontal gyrus (BA 9), right lateral premotor (BA 6) and supplementary motor (BA 6) areas that were also relatively more active in the current study in the depressed compared with the non-depressed patients. Lane et al. (1997)30 report that sadness, but not happiness or disgust induced by recall, was associated with bilateral increases in insula activity. Those authors conclude that the anterior insula cortex appears to be preferentially involved in aspects of negative emotions and it is noted that Pardo et al. (1993)27 and George et al. (1995)28 also found activation of the left insula during a state of induced sadness. It is noted that the insula has been described as having a role as limbic integration cortex35 , with connections to amygdala, as well as to lateral prefrontal cortex, superior temporal and retroinsular parietal regions. Sites within these latter regions were relatively hyperperfused in the depressed patients with epilepsy
395
in the current study. Other brain regions more perfused in these patients and also generally reported to be more active in states of induced sadness include; left striatum, thalamus, dorsolateral prefrontal cortex and rostral anterior cingulate gyrus. These areas largely correspond to the three components of a ‘limbic-cortical’ model of depression proposed by Mayberg (1997)15 , which itself builds on earlier models of limbic regulation of mood (Papez 1937)36 . Mayberg’s model suggests that attentional and cognitive features of depression such as apathy and psychomotor slowing may relate to disturbances of dorsolateral prefrontal, inferior parietal and striatal regions, whilst the biological features including changes in sleep, appetite and libido involve hypothalamic regions (not implicated in our current study) and insula. Finally, she suggests an integratory role for rostral cingulate, in part on the basis of its reciprocal connections with the above regions. It is also noted that Devinsky et al. (1995)37 consider the rostral cingulate gyrus (Brodmann’s area 24), a region that was relatively hyperperfused in the depressed subjects in the current study, to be part of the affective cingulate. Mayberg’s model includes a pattern of decreases and increases in activity across these regions. Nevertheless, the similarity in patterns of regional involvement suggests that in our patients a state of relative hyperactivity across an extended limbic system may be associated with the manifestation of depression in epilepsy, whilst in primary depressive states these same areas may be hypoperfused. Some previous reports of increased blood flow in orbitofrontal and anterior temporal regions associated with induced mood states have subsequently been considered to be confounded by increases in flow in temporalis muscle38 . In the study by Reiman et al. (1997)39 those authors again noted increased blood flow bilaterally in these regions and concluded that in these areas of the brain the increased activity was at least partially attributable to increased temporalis muscle activity. In our study it is noted that there were no increases in flow in what appeared to be similar orbitofrontal or anterior temporal regions. In addition, where there was increased flow in inferior dorsolateral and more rostral temporal regions, this was unilateral. Hence we believe that it is unlikely that our findings can be explained by such muscle flow artefacts. It is noted that the ratio of females to males is greater in the depressed than in the non-depressed group; respectively 2 to 1 and 1 to 1.5. Previous studies using a variety of imaging techniques have failed to provide a consensus on gender differences in rCBF. However, a study employing 99m Tc-HMPAO, as was also used in our own study, failed to observe any sex differences40 . Whilst we consider it unlikely that differences in sex distributions between our two study groups could explain our results, the issue of gender
396
differences in flow indicates that future studies of these patient groups should attempt better matching for gender. A total of seven of the patients studied had previously undergone temporal lobe surgery but these patients were distributed approximately equally across the two groups. In addition, the surgical procedure employed was similar in all these subjects. It is therefore unlikely that a history of temporal lobe surgery in some of the subjects would have accounted for the differences between the depressed and the non-depressed groups. The semi-quantitative ROI analysis performed revealed that whilst both the voxel-based analysis and the ROI analysis demonstrate relative hyperperfusion in the depressed compared with the non-depressed patients with epilepsy, that nevertheless the relative hyperperfusion noted in the depressed patients was within normal limits as defined by the range previously obtained from 40 normal subjects. This observation together with the finding that in the non-depressed patients left frontal and parietal regions were hypoperfused compared with normal values may be explained by the established finding that interictal SPECT scans in patients with focal seizures frequently demonstrate relatively extensive hypometabolism41, 42 . CONCLUSIONS Depression develops relatively frequently in people with epilepsy, as it does in people with IPD. It has been suggested in both of these conditions that elements of the pathophysiology underlying the neurological states may also contribute directly to the development of the depression. Both IPD and primary depression are associated with decreased activity in prefrontal cortical areas and associated paralimbic areas. However, it does not appear to be the case that all types of secondary depression arise from this mechanism. In this study, patients with depression and epilepsy demonstrated relatively greater cerebral activity in a range of cerebral sites than those with epilepsy alone. The pattern of differences between these two groups was similar to the changes in cerebral activity that have been observed in normal volunteers in whom a transient state of depression has been induced. These results suggest that the mechanism underlying the development of epilepsy-related depression may be distinct from that occurring in primary or movement disorderrelated depression and may be associated with relatively greater activity in emotion-related cerebral regions.
H. A. Ring et al.
ACKNOWLEDGEMENTS This work was supported by a project grant from the Medical Research Council (UK). The authors are grateful to Dr MS George for his helpful comments on the manuscript and to the Raymond Way Fund for financial support.
REFERENCES 1. Ring, H. A. and Trimble, M. R. Depression in epilepsy. In: Depression in Neurological Disease (Eds S. E. Starkstein and R. G. Robinson). Baltimore, Johns Hopkins Press, 1993: pp. 63– 83. 2. Devinsky, O. and Bear, D. M. Varieties of depression in epilepsy. Neuropsychiatry, Neuropsychology, and Behavioural Neurology 1991; 4: 49–61. 3. Edeh, J. and Toone, B. K. Antiepileptic therapy, folate deficiency, and psychiatric morbidity: a general practice survey. Epilepsia 1985; 26: 434–440. 4. Barraclough, B. Suicide and epilepsy. In: Epilepsy and Psychiatry (Eds E. H. Reynolds and M. R. Trimble). Edinburgh, Churchill Livingstone, 1981: pp. 72–76. 5. Mendez, M. F., Cummings, J. L. and Benson, F. Depression in epilepsy: significance and phenomenology. Archives of Neurology 1986; 43: 766–770. 6. Perini, G. and Mendius, R. Depression and anxiety in complex partial seizures. Journal of Nervous and Mental Diseases 1984; 172: 287–290. 7. Currie, S., Heathfield, K. W. G., Henson, R. A. and Scott, D. F. Clinical course and prognosis of temporal lobe epilepsy: a survey of 666 patients. Brain 1971; 94: 173–190. 8. Wolf, P. Acute behavioral symptomatology at disappearance of epileptiform EEG abnormality. In: Neurobehavioral Problems in Epilepsy. Advances in Neurology, Vol. 55 (Eds D. B. Smith, D. M. Treiman and M. R. Trimble). New York, Raven Press, 1991: pp. 127–142. 9. Cummings, J. L. Depression and Parkinson’s disease: a review. American Journal of Psychiatry 1992; 149: 443–454. 10. Fukunishi, I., Hosokawa, K. and Ozaki, S. Depression antedating the onset of Parkinson’s disease. The Japanese Journal of Psychiatry and Neurology 1991; 45: 7–11. 11. Starkstein, S. E., Preziosi, T. J., Bolduc, P. L. and Robinson, R. G. Depression in Parkinson’s disease. Journal of Nervous and Mental Disease 1990; 178: 27–31. 12. Mayberg, H. S. Frontal lobe dysfunction in secondary depression. Journal of Neuropsychiatry and Clinical Neurosciences 1994; 6: 428–442. 13. Baxter, L. R., Scwartz, J. M., Phelps, M. E., Mazziota, J. C., Guze, B. H., Selin, C. E., Gerner, R. H. and Sumida, R. Reduction of prefrontal cortex glucose metabolism common to three types of depression. Archives of General Psychiatry 1989; 46: 243–250. 14. Bench, C. J., Friston, K. J., Brown, R. G., Scott, L. C., Frackowiak, R. S. J. and Dolan, R. J. The anatomy of melancholia— focal abnormalities of cerebral blood flow in major depression and the cognitive impairment of depression. Psychological Medicine 1992; 22: 607–615. 15. Mayberg, H. S. Limbic-cortical dysregulation: a proposed model of depression. Journal of Neuropsychiatry and Clinical Neurosciences 1997; 9: 471–481. 16. Drevets, W. C. Functional neuroimaging studies of depression. The anatomy of melancholia. Annual Review of Medicine 1998; 49: 341–361. 17. Ring, H. A., Bench, C. J., Trimble, M. R., Brooks, D. J., Frackowiak, R. S. J. and Dolan, R. J. Depression in Parkinson’s dis-
Patterns of brain activity in patients with epilepsy and depression
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
ease: a Positron Emission Tomography study. British Journal of Psychiatry 1994; 165: 333–339. Hermann, B. P., Seidenberg, M., Haltiner, A. and Wyler, A. R. Mood state in unilateral temporal lobe epilepsy. Biological Psychiatry 1991; 30: 1205–1218. Seidenberg, M., Hermann, B., Noe, A. and Wyler, A. R. Depression in temporal lobe epilepsy: interaction between laterality of lesion and Wisconsin Card Sort performance. Neuropsychiatry, Neuropsychology and Behavioral Neurology 1995; 8: 81–87. Schmitz, E. B., Moriarty, J., Costa, D. C., Ring, H. A., Ell, P. J. and Trimble, M. R. Psychiatric profiles and patterns of cerebral blood flow in focal epilepsy: interactions between depression, obsessionality and perfusion related to the laterality of the epilepsy. Journal of Neurology, Neurosurgery and Psychiatry 1997; 62: 458–463. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders. 3rd and 4th Editions, Washington DC, American Psychiatric Association, 1987, 1994. Kouris, K., Jarritt, P. H., Costa, D. C. and Ell, P. J. Physical assessment of the GE/CGR Neurocam and comparison with a single rotating gamma-camera. European Journal of Nuclear Medicine 1992; 19: 236–242. Hoh, C. K., Dahlbom, M., Harris, G., Choi, Y., Hawkins, R. A., Phelps, M. E. and Maddahi, J. Automated iterative threedimensional registration of positron emission tomography images. Journal of Nuclear Medicine 1993; 34: 2009–2018. Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. P., Frith, C. D. and Frackowiack, R. S. J. Statistical parametric maps in functional imaging: a general linear approach. Human Brain Mapping 1995; 2: 189–210. Bromfield, E. B., Altshuler, L. L., Leiderman, D. B., Balst, M., Ketter, T. A., Devinsky, O., Post, R. M. and Thoodore, W. H. Cerebral metabolism and depression in patients with complex partial seizures. Archives of Neurology 1992; 49: 976. Mayberg, H. S., Starkstein, S. E., Sadzot, B., Preziosi, T. J., Andrezejewski, P. L., Dannals, R. F., Wagner, H. N. and Robinson, R. G. Selective hypometabolism in the inferior frontal lobe in depressed patients with Parkinson’s disease. Annals of Neurology 1990; 28: 57–64. Pardo, J. V., Pardo, P. J. and Raichle, M. E. Neural correlates of self-induced dysphoria. American Journal of Psychiatry 1993; 150: 713–719. George, M. S., Ketter, T. A., Parekh, P. I., Horwitz, B., Herscovitch, P. and Post, R. M. Brain activity during transient sadness and happiness in healthy women. American Journal of Psychiatry 1995; 152: 341–351.
397
29. Baker, S. C., Frith, C. D. and Dolan, R. J. The interaction between mood and cognitive function studied with PET. Psychological Medicine 1997; 27: 565–578. 30. Lane, R. D., Reiman, E. M., Ahern, G. L., Schwartz, G. E. and Davidson, R. J. Neuroanatomical correlates of happiness, sadness and disgust. American Journal of Psychiatry 1997; 154: 926–933. 31. Blumer, D. Postictal depression: significance for the treatment of the neurobehavioral disorder of epilepsy. Journal of Epilepsy 1992; 5: 214–219. 32. Blumer, D. Epilepsy and disorders of mood. In: Advances in Neurology, Vol. 55 (Eds D. Smith, D. Treiman and M. Trimble). New York, Raven Press Ltd, 1991: pp. 185–195. 33. Robertson, M. M., Trimble, M. R. and Townsend, H. R. A. Phenomenology of depression in epilepsy. Epilepsia 1987; 28: 364–372. 34. Betts, T. A. Depression, anxiety and epilepsy. In: Epilepsy and Psychiatry (Eds E. H. Reynolds and M. R. Trimble). Edinburgh, Churchill Livingstone, 1981: pp. 60–71. 35. Augustine, J. R. Circuitry and functional aspects of the insular lobe in primates including humans. Brain Research Reviews 1996; 22: 229–244. 36. Papez, J. W. A proposed mechanism of emotion. Archives of Neurology and Psychiatry 1937; 79: 217–224. 37. Devinsky, O., Morrell, M. J. and Vogt, B. A. Contributions of anterior cingulate cortex to behaviour. Brain 1995; 118: 279– 306. 38. Drevets, W. C., Videen, T., Macleod, A., Haller, J. and Raiche, M. PET images of blood flow changes during anxiety: correction. Science 1989; 256: 1696. 39. Reiman, E. M., Lane, R. D., Ahern, G. L., Schwartz, G. E., Davidson, R. J., Friston, K. J., Yun, L. S. and Chen, K. Neuroanatomical correlates of externally and internally generated human emotion. American Journal of Psychiatry 1997; 154: 918–925. 40. Catafau, A. M., Lomena, F. J., Pavia, J., Paredella, E., Bernardo, M., Setoain, J. and Tolosa, E. Regional cerebral blood flow pattern in normal young and aged volunteers: a 99mTc-HMPAO SPET study. European Journal of Nuclear Medicine 1996; 23: 1329–1337. 41. Andersen, A. D., A’Rovi Hansen, B., Hogenhaven, H. H. M., Paulson, O. B. and Dam, M. Interictal SPECT of rCBF is of clinical utility in the preoperative evaluation of patients with partial epilepsy. Epilepsy Research 1996; 25: 41–50. 42. Newton, M. R., Berkovic, S. F., Austin, M. C., Rowe, C. C., McKay, W. J. and Bladin, P. F. Ictal, postictal and interictal SPECT in the lateralisation of temporal lobe epilepsy. European Journal of Nuclear Medicine 1994; 21: 1067–1071.
Article No. seiz.1999.0320, available online at http://www.idealibrary.com on
Patterns of brain activity in patients with epilepsy and depression † §¶
‡k
§
‡
‡
H. A. RING , P. D. ACTON , D. SCULL , D. C. COSTA , S. GACINOVIK & M. R. GACINOVIK
§
†
Department of Psychological Medicine, St Bartholomew’s and the Royal London School of Medicine, London E1, UK; ‡ Institute of Nuclear Medicine, University College London Medical School, London W1, UK; § Institute of Neurology, University College London Medical School, London WC1, UK Correspondence to: Dr H. A. Ring, Academic Department of Psychological Medicine, 3rd Floor, Alexandra Wing, Royal London Hospital, Whitechapel Road, London E1 1BB, UK
Depression is a recognized feature of epilepsy. This study tested the hypothesis that depression arising in patients with epilepsy would be associated with decreased activity in brain regions previously demonstrated to be hypoperfused both in primary depression and in depression secondary to movement disorders. Two groups of patients with temporal lobe epilepsy were studied, one of which also met DSM IV criteria for a major depressive episode. All underwent a SPECT scan using the blood flow marker, 99m Tc-HMPAO. An automated voxel-based analysis demonstrated no regions of relatively decreased activity in the depressed compared with the non-depressed patients. Sites of relative hyperactivity in the depressed group were concentrated in the left hemisphere, particularly in dorsolateral prefrontal cortex, striatum, thalamus and temporo-parietal regions. Comparison of these data with normal population data revealed that in the depressed epilepsy group regional activities were within the normal range whilst corresponding results from the non-depressed group were below it. Depressed patients with epilepsy have cerebral regions with greater perfusion than non-depressed people with epilepsy, although they are not hyperperfused compared with normals. Our results suggest that depression in people with epilepsy may arise from a mechanism which differs from that underlying the development of depression in patients with movement disorders. c 1999 BEA Trading Ltd
Key words: epilepsy; depression; SPECT; cerebral blood flow.
INTRODUCTION Depression is a clinically important concomitant of epilepsy and is seen relatively frequently in this population1, 2 . A study of the prevalence of psychiatric morbidity in people with epilepsy in the community recorded neurotic depression in 22%3 . The importance of understanding depression in epilepsy is highlighted when the frequency of suicide in this population is considered. Barraclough (1981)4 , reviewing 11 previous studies, reported a suicide rate five times that in the general population. Although depression may occur in reaction to any adverse life event or stress, including epilepsy, there is evidence that in the case of epilepsy it is likely that there is a pathophysiological relationship between § E-mail: [email protected] ¶ Present address: Department of Radiology, University of Pennsyl-
vania, Philadelphia, PA 19104, USA. 1059–1311/99/070390 + 08
$12.00/0
the epilepsy and the development of depression in at least a proportion of the patients in which these conditions coexist. Depression is more common in patients with epilepsy than in non-epileptic patients with comparable degress of disability5 , and within patients suffering from epilepsy there is evidence that depression is reported more frequently in those with temporal lobe foci than in those with generalized epilepsy or extra-temporal foci5, 6 . Indeed, in a study of more than 600 patients with temporal lobe epilepsy, 19% were found to have clinically apparent anxiety whilst 11% were depressed7 . Another strand of evidence supporting a biological basis to depression in epilepsy is provided by accounts of the development of episodes of depression in association with periods of decreased seizure frequency and relative or absolute reduction in epileptiform abnormalities in the electroencephalogram; so-called ‘forced normalization’. It has been suggested that the development c 1999 BEA Trading Ltd
Patterns of brain activity in patients with epilepsy and depression
of psychopathology in these circumstances relates to ongoing electrical epileptiform discharges spreading through brain regions different from those associated with the development of seizures8 . There is evidence that the manifestation of depression secondary to some other neurological disorders may also be related to the pathophysiology of those disorders. In particular, patients with idiopathic Parkinson’s disease (IPD) have an increased rate of depression9 . In this group of patients it has been noted that development of depression may predate the onset of obvious motor symptoms10 and that severity of depression does not correlate with specific measures of motor impairment11 . It has been suggested that depression occurring in patients with basal ganglia disorders (IPD, Huntington’s disease, ischaemic lesions of the striatum) may arise out of a common mechanism involving frontal subcortical and paralimbic systems12 . Disturbances in similar regions have been identified as being associated with the manifestation of states of primary depression, with functional neuroimaging studies of such patients demonstrating focal disturbances concentrated in frontal lobe regions13, 14 . In general, regions of dorsolateral and dorsomedial prefrontal cortex tend to be hypoperfused whilst more ventral and posterior regions of the brain have been reported as being more active in depressed subjects15, 16 . A direct comparison of regional cerebral blood flow patterns in patients with depression occurring in people with IPD to patients with primary depression observed bilateral decreases in rCBF in anteromedial regions of the medial frontal cortex and the cingulate cortex (Brodmann’s areas 9 and 32) in the depressed IPD group compared with those with IPD alone and compared with normal controls17 . This regional disturbance overlapped that observed in patients with primary depression. These observations suggest that medial prefrontal cortex is a common area of neural dysfunction in the manifestation of both primary depression and depression in IPD. There is already some evidence that depression in patients with temporal lobe epilepsy (TLE) is associated with disturbed frontal lobe function. It has been reported18, 19 that in patients with left TLE there is a significant correlation between the presence of a self-reported dysphoric mood state and the degree of frontal lobe dysfunction, measured using a frontal lobe cognitive task (the Wisconsin Card Sorting Test). We have previously reported that in non-depressed patients with left TLE, higher scores on the Beck Depression Inventory were associated with relatively reduced activity in bilateral frontal and right temporal regions20 . The aim of the current study was to further explore the proposal that different patient groups who share
391
a psychopathological state (i.e. depression), but who otherwise have different primary diagnoses (i.e. Major Affective Disorder, IPD, TLE), will demonstrate similar functional disturbances relating to their depression. Specifically, we wished to test the hypothesis that depression meeting formal diagnostic criteria arising in patients with TLE would be associated with evidence of decreased activity in prefrontal cortical regions that have previously been demonstrated to be hypoperfused both in patients with primary depression and those with motor disorders. MATERIALS AND METHODS Study design Two groups of patients were studied. Both groups had previously been diagnosed as having TLE. The experimental group were ascertained on the basis of meeting additional criteria for the diagnosis of depression. The control group were selected from patients with TLE but no evidence of depression. Subjects were initially screened for suitability for the study and informed consent was obtained. On the day of their study each subject received a semistructured clinical interview lasting about 45 minutes which included an assessment of depressive state. Subjects then underwent a resting state SPECT scan lasting 30 minutes. The study was approved by the local research ethics committee and the Administration of Radioactive Substances Advisory Committee (ARSAC) of the UK Department of Health.
Subject selection and assessment Ten subjects were recruited into each group from specialist epilepsy outpatient clinics. All subjects had clear evidence of TLE, based on EEG and structural imaging investigations already performed as part of their clinical management. The same diagnostic criteria were employed with respect to depression as were used by Ring et al. (1994)17 . Hence as well as TLE, the experimental group had depression meeting DSM IIIR (and DSM IV) diagnostic criteria for a major depressive episode21 . The severity of mood disturbance was also measured using the 17-item Hamilton Depression Scale. Only subjects scoring 19 or more were included in the depressed group. Non-depressed subjects had no significant depressive symptoms elicited on clinical assessment. No subjects with any evidence of a progressive neurodegenerative condition were included.
392
Scanning protocol Each subject underwent a single SPECT scan. They were studied in an awake resting state with eyes closed. The blood flow marker employed was 99m TcHMPAO, available commercially from Amersham International as Ceretec. Subjects received a dose of approximately 500 MBq, administered through a vein in the antecubital fossa. Scanning was performed using a triple-headed IGE Neurocam SPECT scanner (GE Medical Systems, Milwaukee, WI, USA), with general purpose collimators. In air the reconstructed image resolution of this system is 10.7 mm full width at half maximum22 . Scanning was for 40 s per projection for 128 projections, acquired into a 64 × 64 matrix, giving a total duration of scanning of approximately 30 min.
Image analysis Images were reconstructed using filtered backprojection, with a Hanning filter with a cut-off frequency of 1.0 cycles/cm and corrected for attenuation (Chang type = 0.12 cm−1 ). Reconstructed images consisted of a 64 × 64 matrix, with a slice thickness of 4 mm, and an in-plane pixel size of 4 mm. Two image analysis techniques were employed; a manual region of interest (ROI) approach and an automated voxel-based comparison. To perform the ROI analysis, after reorientation into transverse (parallel to the inferior frontal–superior occipital plane), sagittal and coronal slices, square ROIs of 4×4 pixels were placed in the cortical and subcortical grey matter structures to obtain average counts per pixel for frontal, parietal and temporal cortices, caudate, thalamus and cerebellum in each hemisphere. At least three contiguous slices were sampled using as many ROIs as were needed to cover each structure. Average counts per pixel in each region were divided by the counts in the cerebellar hemisphere with higher average counts in order to obtain relative perfusion ratios for these brain regions. In the first stage of the automated voxel-based comparison, images were coregistered to each other, using a standard HMPAO template as the baseline image (Nuclear Diagnostics, Sweden). This template consisted of the mean image from 10 coregistered HMPAO scans taken previously of a group of normal, healthy volunteers. All images were registered using linear transformations (i.e. translation, rotation and scaling, with no image warping), minimizing the absolute difference between images23 . After co-registration with each other, the significance of differences between the two groups of sub-
H. A. Ring et al.
jects was assessed. In the absence of absolute quantification, each image was normalized to the mean counts per pixel in the cerebellum, defined by placing a standard region-of-interest template over the coregistered scans. Unpaired t-tests were performed at every voxel, and corrected for non-independent multiple comparisons, using the theory of random Gaussian fields [Statistical Parametric Mapping (SPM)]24 . Parametric images of statistically significant differences between groups were generated, thresholded at P < 0.05 (cluster-level), and overlaid on the standard template image.
RESULTS Clinical data A total of 19 patients were included in the final analysis of the data. Variables describing the subjects and their epilepsy are listed in Table 1. One of the depressed patients was excluded from the final analysis because examination of the SPECT image demonstrated extensive rCBF anomalies. No patients experienced a seizure during scanning and none had reported any seizures in the 24 hours preceding scanning. Table 1: Clinical variables describing the depressed and non-depressed subject groups.
Age (SD) Sex (male : female) Hamilton score (SD) Laterality of focusa Seizure frequency/month mean, (range) Duration of epilepsy in years (SD) Number of anticonvulsants mean, (range)
Depressed
Non-depressed
33 (7.1) 3:6 21 (2) 6 = left, 3 = right
31 (6.8) 6:4 4 (2) 7 = left, 3 = right
6, (1–12) 20 (8.6)
3, (1–5) 22 (6.2)
2, (1–3)
2, (1–3)
a If bilateral foci were present then the laterality of the greatest
disturbance was recorded.
All patients were diagnosed to be suffering from complex partial seizures with a temporal lobe focus. Three of the depressed subjects and one of the nondepressed subjects demonstrated bilateral epileptiform abnormalities on interictal scalp EEG recording. A total of seven subjects had in the past had temporal lobe resections as attempted treatment for their epilepsy. Two of the depressed group and two of the controls had undergone left temporal lobectomy whilst two of the depressed group and one of the controls had undergone right temporal lobectomy. All the patients were taking between one and three antiepileptic drugs. The mean number of agents prescribed to both groups was two and the most frequently prescribed drug was carbamazepine. In addi-
Patterns of brain activity in patients with epilepsy and depression
393
tion, subjects were taking a range of first and second line drugs with no systematic difference in regimens between the groups. Four of the depressed group were taking antidepressant medication at the time of the study. Considering the phenomenology of the depression occurring in the patients with epilepsy and depression, all these subjects met DSM IV criteria for a major depressive episode. The duration of their depression was of at least three weeks and had persisted beyond the immediate post-ictal period in all these subjects. In addition, four of the depressed group had a past history of depressive episodes whilst one of the non-depressed group had previously suffered with panic attacks.
Imaging data Examination of the results of the SPM analysis reveal that although there were several regions at which there was greater activity in the depressed than in the non-depressed subjects, there were no regions of relatively decreased activity in the depressed group. The sites of relative hyperactivity in the depressed group are displayed in Fig. 1. The threshold for display is P < 0.05 and the image is scaled as a P-map with the scale running from P < 0.05 to P < 0.001. It may be seen that the sites of significantly increased activity are concentrated in the left hemisphere, particularly in the left dorsolateral prefrontal cortex in the region of the inferior frontal gyrus including BA 44 and 45 and extending as far anteriorly as the more lateral area of BA 46, the left insula, the left striatum and thalamus, the left superior temporal gyrus (BA 22), the left angular gyrus (BA 39) and the left inferior parietal lobule (BA 40). More rostrally there is also some bilateral relative hyperactivity in the depressed group in the rostral anterior cingulate (BA 24), middle frontal and precentral gyrii (BA 8, 9, 6) and supplementary motor area (BA 6). In addition, there is relative hyperactivity in the region of the left hippocampus in the depressed group. Although it is not possible to obtain absolute quantification of blood flow from the SPECT methodology used in this study, it is possible to gain a semiquantitative estimate of how regional activities in the two groups compare with activities in normal subjects (Table 2). This was achieved by comparing activity in regions of interest (ROIs) containing the areas identified in the SPM analysis as being significantly different between the two groups with activities previously measured from a normal control data set of similar ROIs in a group of 40 normal subjects (age range; 21– 59, sex ratio; 26 males to 14 females). It is noted that, as also observed in the SPM analysis, there is relatively greater activity in frontal, parietal and thalamic
Fig. 1: Sites of relative hyperactivity in the depressed group. The threshold for display is P < 0.05 and the image is scaled as a P-map with the scale running from P < 0.05 (red) to P < 0.001 (yellow). In the image the left side of the brain is presented on the right.
regions in the depressed than the non-depressed subjects. However, it is also observed that nevertheless the mean levels of activity in the depressed group are not, over each region as a whole, above the upper limit of the normal range observed in a normal control population. It is also seen that in the non-depressed epilepsy group regional activities in left frontal and parietal regions are below the lower levels of the normal range. Hence it appears that the depressed patients with epilepsy are less hypoperfused than the non-depressed subjects with respect to a normal population (but are not hyperperfused compared with the normals). Table 2: Region of interest analysis contrasting activitya in areas identified as significantly different in the SPM analysis in the depressed and non-depressed subjects with epilepsy to ROI activity in a group of 40 healthy control subjects. Region
Epilepsydepressed
Epilepsynondepressed
Pb
Normal rangec
left frontal left parietal left thalamus
0.77 0.81 0.91
0.70 0.75 0.84
0.005 0.02 <0.001
0.71–0.81 0.77–0.93 0.79–1.03
a Activity is calculated as a ratio in each subject of counts in
ROI/counts in cerebellum. b P calculated using one-way ANOVA. c Normal range defined as mean +/− two standard deviations for ratios calculated as above in a group of 40 normal controls. SPM = Statistical Parametric Mapping, ROI = region of interest.
DISCUSSION The aim of this study was to explore the hypothesis that depression occurring in association with epilepsy
394
would be associated with prefrontal cortical hypoperfusion similar to that which has been reported in patients with primary depression and in patients with depression associated with Parkinson’s disease. Although our patients with epilepsy and depression did not display hyperperfusion compared with a normal sample, nevertheless there were no sites of hypoactivity relative to the non-depressed epilepsy group. However, widespread regions of relative hyperactivity in the depressed compared with the non-depressed subjects were found across left frontal, temporal, parietal and subcortical structures. Hence contrasting the results of the two epilepsy groups, our results do not support the hypothesis that all secondary depression is associated with prefrontal cortical hypoperfusion in depressed subjects relative to a non-depressed comparison group matched for neurological disease. Two functional imaging studies have previously reported on patterns of cerebral activity in patients with epilepsy and varying degrees of depression20, 25 . Neither of these studies investigated patients with a specific diagnosis of depression meeting diagnostic criteria. Rather, both examined patients unselected with respect to mood, and then commented on the correlations between patterns of cerebral activity and scores on a self-rated depression questionnaire. Both studies found that patients with higher depression scores had evidence of decreased frontal activity. Bromfield et al. (1992)25 noted bilateral inferior frontal hypometabolism in five patients with left temporal lobe epilepsy compared with five non-depressed controls with epilepsy, whilst Schmitz et al. (1997)20 noted a significant relationship between increased scores on the Beck Depression Inventory and decreased activity in bilateral frontopolar and dorsolateral prefrontal regions. Although the regions reported by Bromfield et al. (1992)25 are inferior to those identified by Schmitz and colleagues, nevertheless, all these areas have previously been reported as being underactive in patients with primary depressive illnesses. However, it may be argued that the depression reported in the patients in the present study was qualitatively different from that observed in the former two studies. In those studies patients reported symptoms of depression but did not meet diagnostic criteria for depression. It may be that they were manifesting symptoms of mild to moderate depression in reaction to their life circumstances, whilst the depressed subjects in the current study were suffering from significant depressive states arising in relation to the pathophysiology of their epilepsy. It has been proposed that depression, both primary and that arising secondary to a movement disorder, involves disturbed function in a network of cortical and subcortical sites including dorsolateral prefrontal, anterior cingulate and inferior parietal cortices and the basal ganglia as well as the ventral insula and the
H. A. Ring et al.
hypothalamic-pituitary axis15 . In the majority of studies of patients with depression, both primary and that arising secondarily to movement disorders, activity in these regions, particularly the more dorsal regions, has been found to be reduced13, 14, 17, 26 . Hence the results of the current study, in which many of these same areas were more perfused in the depressed than in the non-depressed epilepsy subjects, suggests that similar functional circuits may be involved in the manifestation of depression in epilepsy but that they are dysfunctional by virtue of being more active, rather than less active. Although the majority of imaging studies of depressed subjects in a resting state have reported evidence of decreased activity in prefrontal regions, several studies exploring the functional correlates of induced sadness in normal volunteers, often females, have found a different pattern of results27–30 . In those studies, by using various mood induction methods and PET measures of rCBF, patterns of cerebral activity associated with induced states of brief but significant sadness have been identified in otherwise euthymic subjects. Although an induced state of sadness in normal volunteers is self-evidently very different from depression occurring in people with epilepsy, there are some aspects of depression as it can develop in those with epilepsy which suggest that consideration of states of induced sadness may be relevant. In both post-ictal depressions and the depression that may develop in the context of forced normalization, the symptoms of depression arise over a short time, are of relatively short duration (in the order of hours of weeks), and tend to be associated with an active sense of sadness rather than the emptiness and flattening of affect that are more likely to be a feature of primary depressive states. Patients suffering from post-ictal depressions have reported that even though their mood was not particularly low prior to a seizure, and even though they know that they are very likely to have more seizures and are not too distressed by this knowledge, that following a seizure their mood changes rapidly and profoundly and that they feel that this is happening beyond their control, as depressive thoughts and feelings overwhelm them31 . These episodes of post-ictal depression may develop repeatedly in people who then make a good recovery from their depressive state. As well as these similarities to artifically induced mood states, there are phenomenological differences between interictal depressive states and primary depression. The most prominent mood disorders of epilepsy have been described as rapid shifting moods with predominant depression and episodic anger32 . It has also been reported that patients with interictal depression have high state anxiety scores33 and that they have more psychotic traits than a compara-
Patterns of brain activity in patients with epilepsy and depression
bly depressed group of patients with non-neurological chronic illness5 . Betts (1981)34 has noted that interictal depressions in people with epilepsy tend to be of sudden onset and departure with fluctuations in severity whilst present. Although the patients with epilepsy and depression in the current study were a relatively heterogeneous group, whose depression was not limited to the post-ictal period or to a process of forced normalization, nevertheless in the majority subjective clinical observations suggested that their depressive states were modulated to an extent by their seizures. Hence overall, there are a number of similarities between states of depression arising in people with epilepsy and states of transient sadness induced in normal volunteers. In studies of negative mood induction in normals it is observed that the negative mood states have been predominantly associated with areas of cerebral hyperactivity and with fewer decreases in activity27–30 . The regions reported as demonstrating increased activity in those studies overlap with the regions identified in the current study as being relatively more active in the patients with depression and epilepsy than in those with epilepsy alone. Apart from the work by Lane et al. (1997)30 who report bilateral changes, the reports note a preponderance of left-sided activation. George et al. (1995)28 report, amongst other findings, activation in left thalamus and striatum and Lane et al. (1997)30 also found increased activity in the striatum, more so on the left, in states of induced sadness, regardless of whether the emotion was precipitated by watching a film or recalled from memory. Pardo et al. (1993)27 , George et al. (1995)28 and Baker et al. (1997)29 describe increased activation in regions of left inferior dorsolateral prefrontal cortex close to areas activated in the current study. In addition, Baker et al. (1997)29 also reported increased rCBF associated with a state of induced sadness in left middle frontal gyrus (BA 9), right lateral premotor (BA 6) and supplementary motor (BA 6) areas that were also relatively more active in the current study in the depressed compared with the non-depressed patients. Lane et al. (1997)30 report that sadness, but not happiness or disgust induced by recall, was associated with bilateral increases in insula activity. Those authors conclude that the anterior insula cortex appears to be preferentially involved in aspects of negative emotions and it is noted that Pardo et al. (1993)27 and George et al. (1995)28 also found activation of the left insula during a state of induced sadness. It is noted that the insula has been described as having a role as limbic integration cortex35 , with connections to amygdala, as well as to lateral prefrontal cortex, superior temporal and retroinsular parietal regions. Sites within these latter regions were relatively hyperperfused in the depressed patients with epilepsy
395
in the current study. Other brain regions more perfused in these patients and also generally reported to be more active in states of induced sadness include; left striatum, thalamus, dorsolateral prefrontal cortex and rostral anterior cingulate gyrus. These areas largely correspond to the three components of a ‘limbic-cortical’ model of depression proposed by Mayberg (1997)15 , which itself builds on earlier models of limbic regulation of mood (Papez 1937)36 . Mayberg’s model suggests that attentional and cognitive features of depression such as apathy and psychomotor slowing may relate to disturbances of dorsolateral prefrontal, inferior parietal and striatal regions, whilst the biological features including changes in sleep, appetite and libido involve hypothalamic regions (not implicated in our current study) and insula. Finally, she suggests an integratory role for rostral cingulate, in part on the basis of its reciprocal connections with the above regions. It is also noted that Devinsky et al. (1995)37 consider the rostral cingulate gyrus (Brodmann’s area 24), a region that was relatively hyperperfused in the depressed subjects in the current study, to be part of the affective cingulate. Mayberg’s model includes a pattern of decreases and increases in activity across these regions. Nevertheless, the similarity in patterns of regional involvement suggests that in our patients a state of relative hyperactivity across an extended limbic system may be associated with the manifestation of depression in epilepsy, whilst in primary depressive states these same areas may be hypoperfused. Some previous reports of increased blood flow in orbitofrontal and anterior temporal regions associated with induced mood states have subsequently been considered to be confounded by increases in flow in temporalis muscle38 . In the study by Reiman et al. (1997)39 those authors again noted increased blood flow bilaterally in these regions and concluded that in these areas of the brain the increased activity was at least partially attributable to increased temporalis muscle activity. In our study it is noted that there were no increases in flow in what appeared to be similar orbitofrontal or anterior temporal regions. In addition, where there was increased flow in inferior dorsolateral and more rostral temporal regions, this was unilateral. Hence we believe that it is unlikely that our findings can be explained by such muscle flow artefacts. It is noted that the ratio of females to males is greater in the depressed than in the non-depressed group; respectively 2 to 1 and 1 to 1.5. Previous studies using a variety of imaging techniques have failed to provide a consensus on gender differences in rCBF. However, a study employing 99m Tc-HMPAO, as was also used in our own study, failed to observe any sex differences40 . Whilst we consider it unlikely that differences in sex distributions between our two study groups could explain our results, the issue of gender
396
differences in flow indicates that future studies of these patient groups should attempt better matching for gender. A total of seven of the patients studied had previously undergone temporal lobe surgery but these patients were distributed approximately equally across the two groups. In addition, the surgical procedure employed was similar in all these subjects. It is therefore unlikely that a history of temporal lobe surgery in some of the subjects would have accounted for the differences between the depressed and the non-depressed groups. The semi-quantitative ROI analysis performed revealed that whilst both the voxel-based analysis and the ROI analysis demonstrate relative hyperperfusion in the depressed compared with the non-depressed patients with epilepsy, that nevertheless the relative hyperperfusion noted in the depressed patients was within normal limits as defined by the range previously obtained from 40 normal subjects. This observation together with the finding that in the non-depressed patients left frontal and parietal regions were hypoperfused compared with normal values may be explained by the established finding that interictal SPECT scans in patients with focal seizures frequently demonstrate relatively extensive hypometabolism41, 42 . CONCLUSIONS Depression develops relatively frequently in people with epilepsy, as it does in people with IPD. It has been suggested in both of these conditions that elements of the pathophysiology underlying the neurological states may also contribute directly to the development of the depression. Both IPD and primary depression are associated with decreased activity in prefrontal cortical areas and associated paralimbic areas. However, it does not appear to be the case that all types of secondary depression arise from this mechanism. In this study, patients with depression and epilepsy demonstrated relatively greater cerebral activity in a range of cerebral sites than those with epilepsy alone. The pattern of differences between these two groups was similar to the changes in cerebral activity that have been observed in normal volunteers in whom a transient state of depression has been induced. These results suggest that the mechanism underlying the development of epilepsy-related depression may be distinct from that occurring in primary or movement disorderrelated depression and may be associated with relatively greater activity in emotion-related cerebral regions.
H. A. Ring et al.
ACKNOWLEDGEMENTS This work was supported by a project grant from the Medical Research Council (UK). The authors are grateful to Dr MS George for his helpful comments on the manuscript and to the Raymond Way Fund for financial support.
REFERENCES 1. Ring, H. A. and Trimble, M. R. Depression in epilepsy. In: Depression in Neurological Disease (Eds S. E. Starkstein and R. G. Robinson). Baltimore, Johns Hopkins Press, 1993: pp. 63– 83. 2. Devinsky, O. and Bear, D. M. Varieties of depression in epilepsy. Neuropsychiatry, Neuropsychology, and Behavioural Neurology 1991; 4: 49–61. 3. Edeh, J. and Toone, B. K. Antiepileptic therapy, folate deficiency, and psychiatric morbidity: a general practice survey. Epilepsia 1985; 26: 434–440. 4. Barraclough, B. Suicide and epilepsy. In: Epilepsy and Psychiatry (Eds E. H. Reynolds and M. R. Trimble). Edinburgh, Churchill Livingstone, 1981: pp. 72–76. 5. Mendez, M. F., Cummings, J. L. and Benson, F. Depression in epilepsy: significance and phenomenology. Archives of Neurology 1986; 43: 766–770. 6. Perini, G. and Mendius, R. Depression and anxiety in complex partial seizures. Journal of Nervous and Mental Diseases 1984; 172: 287–290. 7. Currie, S., Heathfield, K. W. G., Henson, R. A. and Scott, D. F. Clinical course and prognosis of temporal lobe epilepsy: a survey of 666 patients. Brain 1971; 94: 173–190. 8. Wolf, P. Acute behavioral symptomatology at disappearance of epileptiform EEG abnormality. In: Neurobehavioral Problems in Epilepsy. Advances in Neurology, Vol. 55 (Eds D. B. Smith, D. M. Treiman and M. R. Trimble). New York, Raven Press, 1991: pp. 127–142. 9. Cummings, J. L. Depression and Parkinson’s disease: a review. American Journal of Psychiatry 1992; 149: 443–454. 10. Fukunishi, I., Hosokawa, K. and Ozaki, S. Depression antedating the onset of Parkinson’s disease. The Japanese Journal of Psychiatry and Neurology 1991; 45: 7–11. 11. Starkstein, S. E., Preziosi, T. J., Bolduc, P. L. and Robinson, R. G. Depression in Parkinson’s disease. Journal of Nervous and Mental Disease 1990; 178: 27–31. 12. Mayberg, H. S. Frontal lobe dysfunction in secondary depression. Journal of Neuropsychiatry and Clinical Neurosciences 1994; 6: 428–442. 13. Baxter, L. R., Scwartz, J. M., Phelps, M. E., Mazziota, J. C., Guze, B. H., Selin, C. E., Gerner, R. H. and Sumida, R. Reduction of prefrontal cortex glucose metabolism common to three types of depression. Archives of General Psychiatry 1989; 46: 243–250. 14. Bench, C. J., Friston, K. J., Brown, R. G., Scott, L. C., Frackowiak, R. S. J. and Dolan, R. J. The anatomy of melancholia— focal abnormalities of cerebral blood flow in major depression and the cognitive impairment of depression. Psychological Medicine 1992; 22: 607–615. 15. Mayberg, H. S. Limbic-cortical dysregulation: a proposed model of depression. Journal of Neuropsychiatry and Clinical Neurosciences 1997; 9: 471–481. 16. Drevets, W. C. Functional neuroimaging studies of depression. The anatomy of melancholia. Annual Review of Medicine 1998; 49: 341–361. 17. Ring, H. A., Bench, C. J., Trimble, M. R., Brooks, D. J., Frackowiak, R. S. J. and Dolan, R. J. Depression in Parkinson’s dis-
Patterns of brain activity in patients with epilepsy and depression
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
ease: a Positron Emission Tomography study. British Journal of Psychiatry 1994; 165: 333–339. Hermann, B. P., Seidenberg, M., Haltiner, A. and Wyler, A. R. Mood state in unilateral temporal lobe epilepsy. Biological Psychiatry 1991; 30: 1205–1218. Seidenberg, M., Hermann, B., Noe, A. and Wyler, A. R. Depression in temporal lobe epilepsy: interaction between laterality of lesion and Wisconsin Card Sort performance. Neuropsychiatry, Neuropsychology and Behavioral Neurology 1995; 8: 81–87. Schmitz, E. B., Moriarty, J., Costa, D. C., Ring, H. A., Ell, P. J. and Trimble, M. R. Psychiatric profiles and patterns of cerebral blood flow in focal epilepsy: interactions between depression, obsessionality and perfusion related to the laterality of the epilepsy. Journal of Neurology, Neurosurgery and Psychiatry 1997; 62: 458–463. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders. 3rd and 4th Editions, Washington DC, American Psychiatric Association, 1987, 1994. Kouris, K., Jarritt, P. H., Costa, D. C. and Ell, P. J. Physical assessment of the GE/CGR Neurocam and comparison with a single rotating gamma-camera. European Journal of Nuclear Medicine 1992; 19: 236–242. Hoh, C. K., Dahlbom, M., Harris, G., Choi, Y., Hawkins, R. A., Phelps, M. E. and Maddahi, J. Automated iterative threedimensional registration of positron emission tomography images. Journal of Nuclear Medicine 1993; 34: 2009–2018. Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. P., Frith, C. D. and Frackowiack, R. S. J. Statistical parametric maps in functional imaging: a general linear approach. Human Brain Mapping 1995; 2: 189–210. Bromfield, E. B., Altshuler, L. L., Leiderman, D. B., Balst, M., Ketter, T. A., Devinsky, O., Post, R. M. and Thoodore, W. H. Cerebral metabolism and depression in patients with complex partial seizures. Archives of Neurology 1992; 49: 976. Mayberg, H. S., Starkstein, S. E., Sadzot, B., Preziosi, T. J., Andrezejewski, P. L., Dannals, R. F., Wagner, H. N. and Robinson, R. G. Selective hypometabolism in the inferior frontal lobe in depressed patients with Parkinson’s disease. Annals of Neurology 1990; 28: 57–64. Pardo, J. V., Pardo, P. J. and Raichle, M. E. Neural correlates of self-induced dysphoria. American Journal of Psychiatry 1993; 150: 713–719. George, M. S., Ketter, T. A., Parekh, P. I., Horwitz, B., Herscovitch, P. and Post, R. M. Brain activity during transient sadness and happiness in healthy women. American Journal of Psychiatry 1995; 152: 341–351.
397
29. Baker, S. C., Frith, C. D. and Dolan, R. J. The interaction between mood and cognitive function studied with PET. Psychological Medicine 1997; 27: 565–578. 30. Lane, R. D., Reiman, E. M., Ahern, G. L., Schwartz, G. E. and Davidson, R. J. Neuroanatomical correlates of happiness, sadness and disgust. American Journal of Psychiatry 1997; 154: 926–933. 31. Blumer, D. Postictal depression: significance for the treatment of the neurobehavioral disorder of epilepsy. Journal of Epilepsy 1992; 5: 214–219. 32. Blumer, D. Epilepsy and disorders of mood. In: Advances in Neurology, Vol. 55 (Eds D. Smith, D. Treiman and M. Trimble). New York, Raven Press Ltd, 1991: pp. 185–195. 33. Robertson, M. M., Trimble, M. R. and Townsend, H. R. A. Phenomenology of depression in epilepsy. Epilepsia 1987; 28: 364–372. 34. Betts, T. A. Depression, anxiety and epilepsy. In: Epilepsy and Psychiatry (Eds E. H. Reynolds and M. R. Trimble). Edinburgh, Churchill Livingstone, 1981: pp. 60–71. 35. Augustine, J. R. Circuitry and functional aspects of the insular lobe in primates including humans. Brain Research Reviews 1996; 22: 229–244. 36. Papez, J. W. A proposed mechanism of emotion. Archives of Neurology and Psychiatry 1937; 79: 217–224. 37. Devinsky, O., Morrell, M. J. and Vogt, B. A. Contributions of anterior cingulate cortex to behaviour. Brain 1995; 118: 279– 306. 38. Drevets, W. C., Videen, T., Macleod, A., Haller, J. and Raiche, M. PET images of blood flow changes during anxiety: correction. Science 1989; 256: 1696. 39. Reiman, E. M., Lane, R. D., Ahern, G. L., Schwartz, G. E., Davidson, R. J., Friston, K. J., Yun, L. S. and Chen, K. Neuroanatomical correlates of externally and internally generated human emotion. American Journal of Psychiatry 1997; 154: 918–925. 40. Catafau, A. M., Lomena, F. J., Pavia, J., Paredella, E., Bernardo, M., Setoain, J. and Tolosa, E. Regional cerebral blood flow pattern in normal young and aged volunteers: a 99mTc-HMPAO SPET study. European Journal of Nuclear Medicine 1996; 23: 1329–1337. 41. Andersen, A. D., A’Rovi Hansen, B., Hogenhaven, H. H. M., Paulson, O. B. and Dam, M. Interictal SPECT of rCBF is of clinical utility in the preoperative evaluation of patients with partial epilepsy. Epilepsy Research 1996; 25: 41–50. 42. Newton, M. R., Berkovic, S. F., Austin, M. C., Rowe, C. C., McKay, W. J. and Bladin, P. F. Ictal, postictal and interictal SPECT in the lateralisation of temporal lobe epilepsy. European Journal of Nuclear Medicine 1994; 21: 1067–1071.