Midline brain structures in patients with current and remitted major depression

Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

Contents lists available at ScienceDirect

Progress in Neuro-Psychopharmacology & Biological Psychiatry j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p n p

Midline brain structures in patients with current and remitted major depression Tsutomu Takahashi a,b,c,⁎, Murat Yücel a,d, Valentina Lorenzetti a, Kazue Nakamura b, Sarah Whittle a,d, Mark Walterfang a, Michio Suzuki b,c, Christos Pantelis a, Nicholas B. Allen d,e a

Melbourne Neuropsychiatry Centre, Department of Psychiatry, University of Melbourne and Melbourne Health, Victoria, Australia Department of Neuropsychiatry, University of Toyama, Toyama, Japan Core Research for Evolutional Science and Technology, Japan Science and Technology Corporation, Tokyo, Japan d ORYGEN Youth Health Research Centre, Centre for Youth Mental Health, University of Melbourne, Victoria, Australia e Department of Psychology, University of Melbourne, Victoria, Australia b c

a r t i c l e

i n f o

Article history: Received 23 April 2009 Received in revised form 25 May 2009 Accepted 26 May 2009 Available online 6 June 2009 Keywords: Adhesio interthalamica Cavum septum pellucidum Depression State factors Trait factors

a b s t r a c t Brain morphologic changes of limbic-cortical regions have been reported in major depressive disorder (MDD). However, it remains largely unknown whether MDD is associated with abnormalities in midline brain structures, which play a critical role in limbic-cortical connectivity, and whether such changes reflect state or trait markers of the disorder. We used magnetic resonance imaging to investigate the length of the adhesio interthalamica (AI) and cavum septum pellucidum (CSP) in 29 currently depressed patients, 27 remitted depressed patients, and 33 age- and gender-matched healthy control subjects. The currently depressed patients had a significantly shorter AI compared with controls, but there was no difference in the AI length between the remitted patients and controls. The AI length in the overall patient group was negatively correlated with the severity of symptoms of “loss of interest” at the time of scanning. Furthermore, the patients with co-morbid anxiety disorders tended to have a shorter AI compared with those without. The CSP length and prevalence of a large CSP (≥ 6 mm) did not differ between the groups. Although a comprehensive investigation of medication effects was not possible due to incomplete medication data, these findings suggest that a shorter length of the AI may be associated with state-related brain changes in major depression rather than a stable marker of illness vulnerability. Whether the AI length exhibits ongoing changes across the course of the illness remains to be determined in longitudinal studies. © 2009 Elsevier Inc. All rights reserved.

1. Introduction Magnetic resonance imaging (MRI) studies of major depressive disorder (MDD) have demonstrated neuroanatomical alterations predominantly in the amygdala, hippocampus, basal ganglia, orbitofrontal and anterior cingulate cortices (Konarski et al., 2008; Lorenzetti et al., in press), consistent with a limbic-cortical dysregulation model of unipolar depression (Mayberg, 2003). The precise nature of these morphologic changes is unclear, but illness stage (e.g., acutely depressed versus remitted) may affect these findings in a regionally specific manner (Caetano et al., 2004, 2006; Lorenzetti et al., in press). While some brain abnormalities (e.g., volume reduction of subgenual cingulate region)

Abbreviations: AUDIT, Alcohol Use Disorders Identification Test; AI, adhesio interthalamica; ANCOVA, analysis of covariance; ANOVA, analysis of variance; BDI, Beck Depression Inventory; CSP, cavum septum pellucidum; ICV, intracranial volume; MASQ, Mood and Anxiety Symptom Questionnaire; MDD, major depressive disorder; MRI, magnetic resonance imaging; PANAS, Positive Affect and Negative Affect Scale; SCID-IV, Structured Clinical Interview for DSM-IV. ⁎ Corresponding author. Melbourne Neuropsychiatry Centre, c/o National Neuroscience Facility, 161 Barry St, Carlton South, Victoria 3053, Australia. Tel.: +61 3 8344 1800; fax: +61 3 9348 0469. E-mail address: [email protected] (T. Takahashi). 0278-5846/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.pnpbp.2009.05.020

appear to be a trait marker of MDD (Lorenzetti et al., in press), regional expansion of the corpus callosum has been observed in currently depressed patients, but not in remitted patients (Walterfang et al., 2009), implicating state-related impairments of interhemispheric connectivity in the neurobiology of MDD. However, it remains unknown whether other midline brain abnormalities are present in MDD, and whether they also show specific state-related changes. The adhesio interthalamica (AI), a midline structure connecting the medial surfaces of the thalami, is variable in size among individuals and absent in about 20% of human brains (Kretschmann and Weinrich,1992; Percheron, 2004). While a smaller AI in schizophrenia spectrum disorders (Shimizu et al., 2008; Takahashi et al., 2008a,b) has been implicated in early neurodevelopmental insult (O'Rahilly and Muller, 1990; Rosales et al., 1968), the fact that acceleration of the age-related AI atrophy has been observed in schizophrenia suggests state-dependent dynamic changes of the AI (Rosales et al., 1968; Takahashi et al., 2008c). Given that AI abnormalities have also been reported in affective psychosis and borderline personality disorder (Takahashi et al., 2008c, 2009), it is important to examine whether other neuropsychiatric disorders also exhibit abnormalities of the AI. Although the functional significance of the AI in humans is unclear, the midline nuclei of the thalamus, including the AI, has efferent connections with the amygdala

T. Takahashi et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

and anterior cingulate cortex (Graff-Radford, 1997; Percheron, 2004), suggestive of a role for the AI in functional disconnectivity of the limbic– thalamic-cortical circuitry that mediates depressive symptoms (state effects) in MDD (Anand et al., 2005a,b). To our knowledge, however, no MRI studies have examined the size of the AI in MDD. The cavum septum pellucidum (CSP), which is formed by the incomplete fusion of the septum pellucidi, is thought to be a normal anatomical variant, but an unusually large CSP might reflect abnormalities in the fetal neurodevelopment of midline and limbic system structures (Rakic and Yakovlev, 1968). While several MRI studies have reported increased prevalence of a large CSP in schizophrenia and affective psychosis (e.g., de Souza Crippa et al., 2006; Kasai et al., 2004; Kwon et al., 1998; Nopoulos et al., 1997), further studies in larger samples did not replicate these findings (Hagino et al., 2001; Rajarethinam et al., 2008; Takahashi et al., 2007, 2008d). On the other hand, the prevalence of a large CSP in non-psychotic affective disorders has not been adequately investigated; negative CSP findings in small numbers of affective disorder patients (Brisch et al., 2007; Kwon et al., 1998) or in MDD patients (Shioiri et al., 1996) warrant further replication as they have important implications for potential neurodevelopmental models of these disorders (Sanches et al., 2008). This MRI study investigated the size of the AI and CSP in currently depressed patients (cMDD), individuals with a history of major depression but who were currently in remission (rMDD), and healthy comparison subjects. This approach potentially enabled us to examine whether abnormalities in these midline structures in MDD, if present, reflect state or trait influences. On the basis of putative state-related limbic-cortical abnormalities in MDD (Anand et al., 2005a,b) and AI abnormalities in various neuropsychiatric disorders (Takahashi et al., 2008a,b,c), we predicted that only cMDD patients would have shorter length of the AI compared with controls, and that the AI length in the MDD patients would be associated with their symptoms. We also predicted negative CSP findings based on previous studies in mood disorders. 2. Methods 2.1. Participants Eighty-nine subjects were recruited in the study, of which 29 received a current diagnosis of major depressive disorder (cMDD), 27

1059

were currently medically and psychiatrically well individuals with a previous history of major depressive disorder (rMDD), and 33 were healthy control subjects (Table 1). Seven rMDD patients had a total Beck Depression Inventory (BDI; Beck and Steer, 1987) score N 18 at the study, but they did not fulfill the criteria of MDD by the Structured Clinical Interview for DSM-IV (SCID-IV-TR) (First et al., 2001). Demographic and clinical characteristics of the same MDD subjects, recruited through advertisement in the local media from the general community and via outpatient mental health clinics, have been described previously (Lorenzetti et al., 2009). The relevant Human Research Ethics Review committees approved the study protocol, and the participants gave written informed consent after a complete description of the study. Participant inclusion criteria were: age 18–50 years, English as a preferred language, and current IQ N 70. Exclusion criteria were: a history of significant head injury, seizures, impaired thyroid function and steroid use, neurological diseases, electroconvulsive therapy within the past 6 months, and brain abnormalities detected on a T2 scan (e.g., periventricular hyper-intensities). All the depressed subjects with another current Axis I psychiatric disorder (other than anxiety disorders) were excluded, as well as any healthy controls who had a personal history of psychiatric illness, drug or alcohol dependence. All participants underwent a clinical and neuropsychological assessment by experienced research psychologists at Orygen Youth Health, Melbourne. Participants were assessed with the SCID-IVTR, the BDI, the Mood and Anxiety Symptom Questionnaire (MASQ) (Watson et al., 1995), Positive Affect and Negative Affect Scale (PANAS) (Watson et al., 1988) and the Alcohol Use Disorders Identification Test (AUDIT) (Babor et al., 1992). Measures of premorbid and current intelligence were obtained using the Wechsler Test of Adult Reading (Wechsler, 2001) and the Wechsler Abbreviated Scale of Intelligence (Wechsler, 1999). Medication use in the preceding 6 months was also assessed. Thirty-three (21 cMDD and 12 rMDD) patients were on a stable medication regimen for at least 6 months preceding the scan. Of these, seventeen patients were on selective serotonin reuptake inhibitors, four on venlafaxine, three on mirtazapine, two each tricyclics and monoamine-oxidase inhibitors, and one each on lithium and reboxetine. Three patients were receiving combination therapy (paroxetine and benzodiazepine; escitolopram and mirtazapine; and lithium and

Table 1 Demographic and clinical characteristics of the participants.

Age (years) Male/female Age of onset Number of episodes First episode/recurrent Melancholic/atypical Medication past 6 months: yes/no Current anxiety disorder: yes/no Current IQ Premorbid IQ Beck Depression Inventory MASQ General distress General depression General anxiety Anxious arousal High positive affect Loss of interest PANAS Positive affect Negative affect AUDIT

Controls (N = 33)

cMDD (N = 29)

rMDD (N = 27)

Group comparisonsa

34.0 ± 9.9 12/21 – – – – – – 111.1 ± 10.9 111.6 ± 12.3 3.6 ± 4.1 (range, 0–16)

32.5 ± 8.3 7/22 21.1 ± 8.0 3.7 ± 3.4 7/22 10/3 21/6 18/10 104.9 ± 8.7 107.5 ± 11.4 36.8 ± 8.9 (range, 20–56)

35.1 ± 10.0 9/18 26.0 ± 9.4 3.1 ± 2.6 – – 12/13 4/23 111.4 ± 9.9 111.7 ± 8.9 13.0 ± 11.7 (range, 0–40)

F(2, 86) = 0.52, p = 0.595 Chi-square = 1.13, p = 0.568 F(1, 54) = 4.56, p = 0.037; cMDD b rMDD F(1, 38) = 0.37, p = 0.547 – – Chi-square = 4.96, p = 0.026 Chi-square = 14,02 p b 0.001 F(2, 85) = 4.03, p = 0.021; not significant (Scheffé test) F(2, 86) = 1.41, p = 0.250 F(2, 86) = 120.57, p b 0.001; cMDD N rMDD N controls

27.9 ± 8.3 19.5 ± 7.2 16.4 ± 6.4 22.0 ± 4.4 81.1 ± 14.3 14.7 ± 5.0

50.5 ± 7.8 47.3 ± 9.2 32.2 ± 8.7 42.0 ± 12.2 43.6 ± 13.5 31.6 ± 6.4

40.4 ± 10.3 35.0 ± 11.7 24.7 ± 7.7 28.9 ± 7.7 65.0 ± 12.4 23.5 ± 6.8

F(2, 81) = 49.21, p b 0.001; cMDD N rMDD N controls F(2, 82) = 66.85, p b 0.001; cMDD N rMDD N controls F(2, 82) = 32.31, p b 0.001; cMDD N rMDD N controls F(2, 79) = 40.47, p b 0.001; cMDD N rMDD N controls F(2, 80) = 57.19, p b 0.001; cMDD b rMDD b controls F(2, 82) = 58.68, p b 0.001; cMDD N rMDD N controls

32.9 ± 7.3 11.2 ± 1.6 4.6 ± 3.0

21.6 ± 6.5 21.2 ± 8.5 5.4 ± 6.2

28.7 ± 8.0 14.2 ± 4.7 5.7 ± 4.8

F(2, 82) = 18.57, p b 0.001; cMDD b rMDD, controls F(2, 83) = 24.98, p b 0.001; cMDD N rMDD, controls F(2, 83) = 0.42, p = 0.662

The values represent mean ± SD, except where noted. AUDIT, Alcohol Use Disorders Identification Test; cMDD, currently depressed patients; MASQ, Mood and Anxiety Symptom Questionnaire; PANAS, Positive and Negative Affect Schedule; rMDD, remitted depressed patients. a Difference between the degree of freedom across measures is due to missing data.

1060

T. Takahashi et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

Table 2 Measures of midline brain structures. Brain region

Controls (N = 33)

cMDD (N = 29)

rMDD (N = 27)

Group comparison

Intracranial volume (cm3) CSP (≥1 slice) present; N (%) Large CSP (≥ 6 slices) present; N (%) CSP length (log) AI absent; N (%) AI length (mm)

1493 ± 143 28 (84.9) 5 (15.2) 0.31 ± 0.4 1 (3.0) 12.3 ± 2.7 (range, 5–16)

1477 ± 138 26 (89.7) 1 (3.5) 0.32 ± 0.3 4 (13.8) 10.8 ± 4.2b (range, 2–17)

1470 ± 150 22 (81.5) 2 (7.4) 0.33 ± 0.5 0 (0) 12.4 ± 2.8 (range, 6–17)

F(2, 85) = 0.20, p = 0.816a Fisher's exact test, p = 0.683 Fisher's exact test, p = 0.259 F(2, 81) = 0.10, p = 0.905 Fisher's exact test: p = 0.059 F(2, 81) = 3.84, p = 0.025

The values represent mean ± SD, except where noted. cMDD, currently depressed patients; rMDD, remitted depressed patients. a ANCOVA with age as a covariate and group as a between-subject factor was used. b Significantly shorter than controls (p = 0.019) and tended to be shorter than rMDD (p = 0.064).

dothiepin), while 3 cMDD and 6 rMDD patients were medicationnaïve.

kappa = 0.79) and 100% (30/30 agreement, kappa = 1.00)/100% (30/ 30 agreement, kappa = 1.00), respectively.

2.2. Magnetic resonance imaging procedures

2.3. Statistical analysis

MR scans were acquired with a 1.5-T Magnetom Avanto scanner (Siemens Medical System, Inc., Erlangen, Germany) at the St. Vincent's Hospital Melbourne, Victoria. A structural T1-weighted image analysis was performed, with time to echo = 2.3 ms, time repetition = 2.1 ms, flip angle = 15°, matrix size = 256 × 256, voxel dimension = 1 × 1 × 1 mm. Additionally, MRI abnormalities were assessed using a high-resolution T2-weighted scan. The intracranial volume (ICV) was measured on T1-weighted images to correct for differences in head size as previously described (Eritaia et al., 2000); the groups did not significantly differ in their ICVs (Table 2). For the assessment of the AI and CSP (Fig. 1), the T1-weighted images were processed using Dr. View software (AJS, Tokyo, Japan) as described elsewhere (Takahashi et al., 2007, 2008a). Briefly, brain images were realigned in three dimensions and then reconstructed into entire contiguous coronal images with a 1-mm thickness, perpendicular to the AC-PC line. One rater (TT), who was blind to the subjects' identity, counted the number of coronal slices where each midline structure was clearly seen. Since the images were 1-mm thick with no gap, the rating was a reflection of the actual anterior– posterior length of the AI and CSP (in mm). On the basis of previous studies (Takahashi et al., 2008a,b), we considered the AI as present when it could be identified on three or more 1-mm slices in both coronal and axial views. The CSP was defined as present when it could be identified on at least one coronal slice. A CSP equal to or greater than 6 slices was defined as large (Takahashi et al., 2007). Intra- and inter-rater (TT and KN) intraclass correlation coefficients for the AI and CSP lengths in 30 randomly selected brains were over 0.97. Intra-/ inter-rater reliabilities for the prevalence of the AI and large CSP were 100% (30/30 agreement, kappa = 1.00)/97% (29/30 agreement,

Clinical and demographic differences between groups were examined with one-way analysis of variance (ANOVA) or chi-square test. The length of each midline structure was analyzed using the analysis of covariance (ANCOVA), with ICV and age as covariates and with diagnosis and gender as between-subject factors. For this analysis, the subjects without CSP (3 cMDD, 5 rMDD, and 5 control subjects) were regarded as having a CSP of 0.5 mm and then logtransformed because of their skewed distribution (Takahashi et al., 2007). Post hoc Scheffé tests were used to follow-up the significant main effects or interactions yielded by these analyses. Chi-square tests, of Fisher's exact tests when expected cell sizes were less than five, were used for assessing the frequency of the AI and CSP. The relationships between the length of the midline structures and clinical variables were examined using Spearman's rank correlation coefficients because of a skewed distribution of these variables. The association of premorbid or current IQ (adequately normally distributed according to the Kolmogorov–Smirnov test) to the length of the AI and CSP (log) was examined using Pearson's partial correlation coefficients controlling for age and ICV. Statistical significance was defined as p b 0.05. 3. Results 3.1. Demographic and clinical data Comparison of the groups revealed no significant difference in age, gender, and intelligence but, as expected, measures of depressive and anxiety symptoms were significantly different between groups (Table 1). The cMDD patients, as compared to rMDD patients, had an earlier age of onset, a higher proportion of participants on medication

Fig. 1. Sample coronal slices showing the cavum septum pellucidum (A) and adhesio interthalamica (B). Arrows indicate the position of each midline brain structure.

T. Takahashi et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

1061

3.3. Correlational analysis Age and IQ (current, premorbid) did not correlate with the AI or CSP length in controls, the depression group as a whole, or the cMDD and rMDD subgroups. There was a negative correlation between the AI length and MASQ scale reflecting loss of interest in pleasurable activities in the depression group as a whole (rho= −0.404, p = 0.003). When the cMDD and rMDD subgroups were examined separately, this negative correlation was significant only in the rMDD group (Table 3). Age of onset was negatively correlated with AI length only in the rMDD group (rho = −0.518, p = 0.006). No significant correlations were found between the length of midline structures and other clinical measures (i.e., number of episodes, total BDI score, PANAS subscale scores, and AUDIT score) when the depression group was examined as a whole, or in the cMDD and rMDD subgroups separately. 4. Discussion Fig. 2. Length of the cavum septi pellucisi (CSP) in patients with current (cMDD) and remitted (rMDD) depression and healthy control participants.

in the previous 6 months, and a higher rate of co-morbid anxiety disorder. 3.2. AI and CSP measurements There was no significant group difference in the prevalence of an absent AI (Table 2). ANCOVA of the AI length revealed significant main effects for diagnosis [F(2, 81) = 3.84, p = 0.025] and gender [F(1, 81) = 4.78, p = 0.032], but there was no gender-by-group interaction [F(2, 81) = 0.87, p = 0.422]. Post hoc tests indicated that the cMDD patients had a significantly shorter AI than controls (p = 0.019) and showed a non-significant trend to have a shorter AI than rMDD patients (p = 0.064). Males had a significantly shorter AI than females (p b 0.001). The MDD patients with co-morbid anxiety disorder (N = 22) tended to have a shorter AI compared with those without (N = 33) [F(1, 51) = 3.68, p = 0.061], but there were no differences in the AI length when melancholic (N = 10) and non-melancholic (N = 18) subgroups of cMDD patients [F(1, 24) = 0.00, p = 0.959] or depressed patients who were (N = 33) and were not (N = 19) taking medication in the preceding 6 months [F(1, 48) = 0.81, p = 0.371] were compared. Overall frequencies of the CSP and large CSP in the present sample were 85.4% and 9.0% respectively, showing no group differences (Table 2, Fig. 2). ANCOVA of the CSP (log) revealed a significant effect for gender [F(1, 81) = 4.54, p = 0.036], but post hoc analysis showed no gender difference (p = 0.904). There was no effect for diagnosis [F(2, 81) = 0.10, p = 0.905] or gender-by-group interaction [F(2, 81) = 1.26, p = 0.289]. Co-morbid anxiety disorder [F(1, 51)= 2.25, p = 0.140], melancholic/non-melancholic subgroups [F(1, 24) = 0.07, p = 0.805], and medication [F(1, 48) = 0.44, p = 0.509] did not significantly affect the CSP length in the MDD patients.

To our knowledge, this is the first MRI study to report the size of both AI and CSP in patients with current and remitted MDD. As predicted, the currently depressed patients had a significantly shorter AI compared with controls, but AI length in remitted patients was not different from healthy controls. The AI length in the MDD patients was negatively correlated with the degree of loss of interest at the time of scanning. On the other hand, we did not identify any differences in the prevalence or length of the CSP between the groups. These findings suggest that different biological processes are responsible for the development of these midline brain structures, and that a shorter length of the AI may be associated with state-related brain changes in major depression rather than a stable marker of illness vulnerability. The prevalence and size of the AI are considered to reflect early neurodevelopmental factors because it develops early during the gestation period, but the AI also exhibits age-related morphologic changes, especially after the third decade (Rosales et al., 1968). As discussed elsewhere (Takahashi et al., 2008a), previous studies have reported a wide variance from 2.3% (Takahashi et al., 2008c) to 22.3% (Nopoulos et al., 2001) in the prevalence of an absent AI in healthy subjects, possibly reflecting methodological and sample differences across studies (e.g., imaging techniques, age, and gender). The present finding of low prevalence of absent AI [3.0% (1/33)] in relatively young control subjects is consistent with recent studies using highresolution MRI [missing in 3/51 controls (Shimizu et al., 2008) and 2/ 87 controls (Takahashi et al., 2008c)] and with the notion of agerelated atrophy of the AI. We also replicated a large interindividual variability (Table 2) as well as a sexual dimorphism of the size of the AI reported in healthy subjects, with females having a larger AI than males (Allen and Gorski, 1991). Our finding of the shorter AI in currently depressed patients is consistent with the limbic-cortical dysregulation model of major depression that is primarily based on observations from functional neuroimaging (Anand et al., 2005a,b; Mayberg, 2003), since the midline nuclei of the thalamus, as well as the AI, are likely to be connected with

Table 3 Correlations between the length of the adhesio interthalamica (AI) and MASQ scores.

MASQ General distress General depression General anxiety Anxious arousal High positive affect Loss of interest

Controls

Whole MDD

cMDD

rMDD

− 0.070 (0.710) 0.026 (0.887) 0.020 (0.914) − 0.317 (0.082) − 0.201 (0.279) − 0.016 (0.931)

− 0.266 (0.054) − 0.256 (0.064) 0.002 (0.987) 0.006 (0.966) 0.178 (0.208) − 0.404 (0.003)⁎

0.073 (0.711) 0.025 (0.900) 0.376 (0.049) 0.443 (0.018) 0.077 (0.698) − 0.279 (0.151)

− 0.542 (0.005)⁎ − 0.499 (0.011) − 0.256 (0.217) − 0.233 (0.284) 0.191 (0.370) − 0.520 (0.008)⁎

The values represent Spearman's rho (p). cMDD, currently depressed patients; MASQ, Mood and Anxiety Symptom Questionnaire; MDD, major depressive disorder; rMDD, remitted depressed patients. ⁎ Significant even after correction for multiple comparisons [6 subscales; p b 0.0083(0.05/6)].

1062

T. Takahashi et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

limbic-cortical regions implicated in the neurobiology of MDD (e.g., amygdala, frontal and anterior cingulate cortices; Graff-Radford, 1997; Percheron, 2004). On the other hand, remitted depressed patients, who were psychiatrically well but remained vulnerable to illness relapse, had normal AI length, suggesting that the AI abnormalities are not related to an underlying disease diathesis that is observable between episodes. Interestingly, serotonergic dysfunction or imbalances in activity and connectivity of limbic–thalamic-cortical regions, which may contribute to negative mood states in major depression (Anand et al., 2005a; Holthoff et al., 2004; Reimold et al., 2008), appear to be normalized following the antidepressant treatment and/or recovery from depression (Anand et al., 2005b; Bhagwagar et al., 2007; Holthoff et al., 2004). Our finding of a negative correlation between the AI length and severity of some depressive symptoms (i.e., the MASQ loss of interest subscale score) also supports the notion that AI alterations in MDD may represent a state marker related to symptom severity rather than a trait marker of illness vulnerability. The precise nature or anatomical pattern of brain morphologic changes in the course of major depression remains unclear, but previous cross-sectional MRI studies have demonstrated that some of brain morphologic changes, such as reduced hippocampal volume (Caetano et al., 2004) or corpus callosum expansion (Walterfang et al., 2009), are also apparent only in currently depressed patients. While MRI studies cannot address the underlying mechanism of these potentially state-related brain morphologic changes, stress-toxicity (Duman, 2002), neurohumoral changes (e.g., hypothalamic–pituitary–adrenal axis dysregulation) (Lorenzetti et al., in press), and/or neuroplastic effects of medication (Frodl et al., 2008; Moore et al., 2002) might be relevant. Given that the AI contains thalamic neurons and glial cells (Percheron, 2004), our findings might be in line with putative glial pathology of depression (Páv et al., 2008; Rajkowska and Miguel-Hidalgo, 2007), as well as with experimental evidence that stress-induced reductions in glial metabolism and glial fibrillary associated protein (GFAP) mRNA expression were reversed by chronic treatment with the glutamate-modulating drug riluzole (Banasr et al., in press). However, a follow-up MRI study in MDD did not find longitudinal volume changes of the hippocampus during 3 years after an acute depressive episode (Frodl et al., 2008) and no MRI studies have examined the longitudinal morphologic changes of midline brain structures during depression and after remission in the same patients. While antidepressants (Frodl et al., 2008) and mood stabilizers (Nakamura et al., 2007; Moore et al., 2002) may increase gray matter volume, it is unknown whether the medication also affects the size of the AI. In addition, our finding of a shorter AI only in the cMDD patients could be interpreted as indicative of severe neurodevelopmental abnormalities in refractory depression. The possibility also exists that our AI findings may be associated with or even be consequent to other brain structural changes. Thus, further longitudinal study is required to examine the nature of morphologic changes of the AI and related brain regions as well as the potential influence of key clinical features (e.g., medication, clinical course) on hypothesized state-related brain changes in MDD. In contrast to the AI findings, we found no difference in the prevalence of a large CSP, a marker of fetal neurodevelopmental abnormalities (Rakic and Yakovlev, 1968), between MDD patients and healthy controls. In addition, there was no relationship between the size of the CSP and clinical measures including depressive symptoms as well as subtypes (i.e., currently depressed versus remitted patients, co-morbid anxiety disorders, melancholic versus non-melancholic) of the MDD patients. Taken together with negative CSP findings in pervious postmortem (Brisch et al., 2007) and MRI (Kwon et al., 1998; Shioiri et al., 1996) studies of MDD, the present findings suggest that CSP is unlikely to be related to vulnerability or clinical presentation of major depression. A few potentially confounding factors in this study should be taken into account. First, complete lifetime medication data (e.g., dose, duration) of the patients were not available, representing a clear

limitation of the study. No difference in the CSP or AI length was found between patients who were and were not on antidepressant medication, but medication could affect self-reported answers concerning feelings, thus possibly decoupling the structural changes of depression from the feeling of being depressed. In addition, it remains unclear whether antidepressant-like substances (e.g. nicotine or caffeine) affect brain structures in healthy subjects (Gallinat et al., 2006). Thus, the potential effects of medication or other factors on the midline brain structures in MDD remain to be determined in future studies. Second, in addition to a relatively small sample size, potential heterogeneity between the cMDD and rMDD groups might have biased our results because assessment of psychotic symptoms or detailed subtype of rMDD patients was not comprehensively undertaken. Lower current IQ in the cMDD patients compared with other groups might have also biased our results. However, there was no significant effect of premorbid or current IQ on the length of midline structures in any group. The rMDD patients had a later onset compared with cMDD patients, but this difference alone could not explain the shorter AI in the cMDD group because the AI length was negatively correlated with onset age. Finally, high prevalence of comorbid anxiety disorders of the cMDD group (18 of 28 patients) raises the possibility that the AI abnormalities in this group were driven predominantly by symptomatic anxiety. Although we found the association of the AI length to depressive symptoms (i.e., loss of interest in pleasurable activities) rather than anxiety symptoms, the patients with co-morbid anxiety disorder tended to have a shorter AI compared with those without. We have also demonstrated in this cohort that comorbidly anxious patients exhibit callosal expansions similar to those seen in cMDD patients compared to depressed patients without anxiety (Walterfang et al., 2009). Given that the AI abnormalities have been also reported in a range of psychotic (Shimizu et al., 2008; Takahashi et al., 2008a,c) and personality (Takahashi et al., 2008b, 2009) disorders, it is possible that reduced AI length is a marker of general psychopathology, rather than a specific illness marker. Thus, further work should investigate the specificity of AI findings in various neuropsychiatric disorders including anxiety disorders without depression. 5. Conclusion In conclusion, our findings of a shorter AI in currently depressed but not in remitted MDD patients suggest that the AI abnormalities are state-dependent. A significant negative correlation between the AI length and severity of depressive symptoms at scanning further supports this hypothesis. Further studies are needed to confirm these putative state-related AI changes during the course of illness using longitudinal designs. Acknowledgements This research was supported by a grant from the Australian Research Council (I.D. DP0557663) awarded to A/Prof. Allen and A/ Prof. Yücel. Neuroimaging analysis was facilitated by the Neuropsychiatry Imaging Laboratory managed by Ms. Bridget Soulsby at the Melbourne Neuropsychiatry Centre and supported by Neurosciences Victoria. The authors thank Ms. Orli Schwartz and Ms. Diana Maud for recruitment and assessment of the participants. A/Prof. Yücel is supported by an NHMRC Clinical Career Development Award (I.D. 509345). Ms. Lorenzetti is supported by a scholarship of the Faculty of Psychology, The University of Bologna, Italy. Dr. Walterfang was supported by a Pfizer Neuroscience Research Grant and a Stanley Research Centre Grant. Dr. Whittle was supported by an Australian Research Council Postdoctoral Fellowship. Dr. Takahashi was supported to undertake this work by a Grant-in-Aid for Scientific Research (No. 19591346) from the Japanese Society for the Promotion of Science; and a Research Grant (17-2,18-6) for Nervous and Mental

T. Takahashi et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 33 (2009) 1058–1063

Disorders from the Ministry of Health and Welfare (Dr. Suzuki), Japan. A/Prof. Allen is supported by a grant from the Colonial Foundation. References Allen LS, Gorski RA. Sexual dimorphism of the anterior commissure and massa intermedia of the human brain. J Comp Neurol 1991;312:97-104. Anand A, Li Y, Wang Y, Wu J, Gao S, Bukhari L, et al. Activity and connectivity of brain mood regulating circuit in depression: a functional magnetic resonance study. Biol Psychiatry 2005a;57:1079–88. Anand A, Li Y, Wang Y, Wu J, Gao S, Bukhari L, et al. Antidepressant effect on connectivity of the mood-regulating circuit: an FMRI study. Neuropsychopharmacology 2005b;30:1334–44. Babor TF, De La Fuente JR, Saunders J, Grant M. The alcohol use disorders identification test: guidelines for use in primary health care. Geneva: World Health Organization; 1992. Banasr M, Chowdhury GM, Terwilliger R, Newton SS, Duman RS, Behar KL, et al. Glial pathology in an animal model of depression: reversal of stress-induced cellular, metabolic and behavioral deficits by the glutamate-modulating drug riluzole. Mol Psychiatry in press. doi:10.1038/mp.2008.106. Beck AT, Steer RT. Beck Depression Inventory Manual. San Antonio: HBJ; 1987. Bhagwagar Z, Murthy N, Selvaraj S, Hinz R, Taylor M, Fancy S, et al. 5-HTT binding in recovered depressed patients and healthy volunteers: a positron emission tomography study with [11C]DASB. Am J Psychiatry 2007;164:1858–65. Brisch R, Bernstein HG, Krell D, Stauch R, Trübner K, Dobrowolny H, et al. Volumetric analysis of septal region in schizophrenia and affective disorder. Eur Arch Psychiatry Clin Neurosci 2007;257:140–8. Caetano SC, Hatch JP, Brambilla P, Sassi RB, Nicoletti M, Mallinger AG, et al. Anatomical MRI study of hippocampus and amygdala in patients with current and remitted major depression. Psychiatry Res 2004;132:141–7. Caetano SC, Kaur S, Brambilla P, Nicoletti M, Hatch JP, Sassi RB, et al. Smaller cingulate volumes in unipolar depressed patients. Biol Psychiatry 2006;59:702–6. de Souza Crippa JA, Zuardi AW, Busatto GF, Sanches RF, Santos AC, Araújo D, et al. Cavum septum pellucidum and adhesio interthalamica in schizophrenia: an MRI study. Eur Psychiatry 2006;21:291–9. Duman RS. Pathophysiology of depression: the concept of synaptic plasticity. Eur Psychiatry 2002;17(Suppl 3):306–10. Eritaia J, Wood SJ, Stuart GW, Bridle N, Dudgeon P, Maruff P, et al. An optimized method for estimating intracranial volume from magnetic resonance images. Magn Reson Med 2000;44:973–7. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for Axis 1 DSM-IV Disorders. New York: New York State Psychiatric Institute; 2001. Frodl T, Jäger M, Smajstrlova I, Born C, Bottlender R, Palladino T, et al. Effect of hippocampal and amygdala volumes on clinical outcomes in major depression: a 3-year prospective magnetic resonance imaging study. J Psychiatry Neurosci 2008;33:423–30. Gallinat J, Meisenzahl E, Jacobsen LK, Kalus P, Bierbrauer J, Kienast T, et al. Smoking and structural brain deficits: a volumetric MR investigation. Eur J Neurosci 2006;24:1744–50. Graff-Radford NR. Syndromes due to acquired thalamic damage. In: Feinberg TE, Farah MJ, editors. Behavioral and neurology and neuropsychology. New York: McGrawHill; 1997. p. 433–43. Hagino H, Suzuki M, Kurokawa K, Mori K, Nohara S, Takahashi T, et al. Magnetic resonance imaging study of the cavum septi pellucidi in patients with schizophrenia. Am J Psychiatry 2001;158:1717–9. Holthoff VA, Beuthien-Baumann B, Zündorf G, Triemer A, Lüdecke S, Winiecki P, et al. Changes in brain metabolism associated with remission in unipolar major depression. Acta Psychiatr Scand 2004;110:184–94. Kasai K, McCarley RW, Salisbury DF, Onitsuka T, Demeo S, Yurgelun-Todd D, et al. Cavum septi pellucidi in first-episode schizophrenia and first-episode affective psychosis: an MRI study. Schizophr Res 2004;71:65–76. Kwon JS, Shenton ME, Hirayasu Y, Salisbury DF, Fischer IA, Dickey CC, et al. MRI study of cavum septi pellucidi in schizophrenia, affective disorder, and schizotypal personality disorder. Am J Psychiatry 1998;155:509–15. Konarski JZ, McIntyre RS, Kennedy SH, Rafi-Tari S, Soczynska JK, Ketter T. Volumetric neuroimaging investigations in mood disorders: bipolar disorder versus major depressive disorder. Bipolar Disord 2008;10:1-37. Kretschmann HJ, Weinrich W. Clinical neuroimaging and clinical neuroanatomy. 2nd ed. New York: Thieme Medical Publishers; 1992. Lorenzetti V, Allen NB, Fornito A, Pantelis C, De Plato G, Ang A, et al. Pituitary gland volume in currently depressed and remitted depressed patients. Psychiatry Res 2009;172:55–60. Lorenzetti V, Allen NB, Fornito A, Yücel M. Structural brain abnormalities in major depressive disorder: a selective review of recent MRI studies. J Affect Disord, in press. doi:10.1016/j.jad.2008.11.021.

1063

Mayberg HS. Modulating dysfunctional limbic-cortical circuits in depression: towards development of brain-based algorithms for diagnosis and optimised treatment. Br Med Bull 2003;65:193–207. Moore GJ, Bebchuk JM, Wilds IB, Chen G, Manji HK. Lithium-induced increase in human brain grey matter. Lancet 2002;356:1241–2. Nakamura M, Salisbury DF, Hirayasu Y, Bouix S, Pohl KM, Yoshida T, et al. Neocortical gray matter volume in first-episode schizophrenia and first-episode affective psychosis: a cross-sectional and longitudinal MRI study. Biol Psychiatry 2007;62:773–83. Nopoulos P, Swayze V, Flaum M, Ehrhardt JC, Yuh WT, Andreasen NC. Cavum septi pellucidi in normals and patients with schizophrenia as detected by magnetic resonance imaging. Biol Psychiatry 1997;41:1102–8. Nopoulos PC, Rideout D, Crespo-Facorro B, Andreasen NC. Sex differences in the absence of massa intermedia in patients with schizophrenia versus healthy controls. Schizophr Res 2001;48:177–85. O'Rahilly R, Muller F. Ventricular system and choroid plexuses of the human brain during the embryonic period proper. Am J Anat 1990;189:285–302. Páv M, Kováru H, Fiserová A, Havrdová E, Lisá V. Neurobiological aspects of depressive disorder and antidepressant treatment: role of glia. Physiol Res 2008;57:151–64. Percheron G. Thalamus. In: Paxinos G, Mai JK, editors. The human nervous system. 2nd ed. California: Elsevier Academic Press; 2004. p. 592–675. Rajarethinam R, Sohi J, Arfken C, Keshavan MS. No difference in the prevalence of cavum septum pellucidum (CSP) between first-episode schizophrenia patients, offspring of schizophrenia patients and healthy controls. Schizophr Res 2008;103:22–5. Rajkowska G, Miguel-Hidalgo JJ. Gliogenesis and glial pathology in depression. CNS Neurol Disord Drug Targets 2007;6:219–33. Rakic P, Yakovlev PI. Development of the corpus callosum and cavum septi in man. J Comp Neurol 1968;132:45–72. Reimold M, Batra A, Knobel A, Smolka MN, Zimmer A, Mann K, et al. Anxiety is associated with reduced central serotonin transporter availability in unmedicated patients with unipolar major depression: a [11C]DASB PET study. Mol Psychiatry 2008;13:606–13 [557]. Sanches M, Keshavan MS, Brambilla P, Soares JC. Neurodevelopmental basis of bipolar disorder: a critical appraisal. Prog Neuropsychopharmacol Biol Psychiatry 2008;32:1617–27. Rosales RK, Lemay MJ, Yakovlev PI. The development and involution of massa intermedia with regard to age and sex. J Neuropathol Exp Neurol 1968;27:166. Shimizu M, Fujiwara H, Hiraol K, Namiki C, Fukuyama H, Hayashi T, et al. Structural abnormalities of the adhesio interthalamica and mediodorsal nuclei of the thalamus in schizophrenia. Schizophr Res 2008;101:331–8. Shioiri T, Oshitani Y, Kato T, Murashita J, Hamakawa H, Inubushi T, et al. Prevalence of cavum septum pellucidum detected by MRI in patients with bipolar disorder, major depression and schizophrenia. Psychol Med 1996;26:431–4. Takahashi T, Suzuki M, Hagino H, Niu L, Zhou SY, Nakamura K, et al. Prevalence of large cavum septi pellucidi and its relation to the medial temporal lobe structures in schizophrenia spectrum. Prog Neuropsychopharmacol Biol Psychiatry 2007;31:1235–41. Takahashi T, Suzuki M, Nakamura K, Tanino R, Zhou SY, Hagino H, et al. Association between absence of the adhesio interthalamica and amygdala volume in schizophrenia. Psychiatry Res 2008a;162:101–11. Takahashi T, Suzuki M, Zhou SY, Nakamura K, Tanino R, Kawasaki Y, et al. Prevalence and length of the adhesio interthalamica in schizophrenia spectrum disorders. Psychiatry Res 2008b;164:90–4. Takahashi T, Yücel M, Yung AR, Wood SJ, Phillips LJ, Berger GE, et al. Adhesio interthalamica in individuals at high-risk for developing psychosis and patients with psychotic disorders. Prog Neuropsychopharmacol Biol Psychiatry 2008c;32:1708–14. Takahashi T, Yung AR, Yücel M, Wood SJ, Phillips LJ, Harding IH, et al. Prevalence of large cavum septi pellucidi in ultra high-risk individuals and patients with psychotic disorders. Schizophr Res 2008d;105:236–44. Takahashi T, Chanen AM, Wood SJ, Walterfang M, Harding IH, Yücel M, et al. Midline brain structures in teenagers with first-presentation borderline personality disorder. Prog Neuropsychopharmacol Biol Psychiatry 2009;33:842–6. Walterfang M, Yücel M, Barton S, Reutens DC, Wood AG, Chen J, et al. Corpus callosum size and shape in individuals with current and past depression. J Affect Disord 2009;115(3):411–20. Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol 1988;54:1063–70. Watson D, Clark L, Weber K, Assenheimer J, Strauss M, McCormick R. Testing a tripartite model: I. Evaluating the convergent and discriminant validity of anxiety and depression symptom scales. J Abnorm Psychology 1995;104:3-14. Wechsler D. Manual for the Wechsler Abbreviated Scale of Intelligence. San Antonio, Texas: The Psychological Corporation; 1999. Wechsler D. Manual for the Wechsler Test of Adult Reading (WTAR). San Antonio, Texas: The Psychological Corporation; 2001.