Introduction
The habenula is a paired diencephalic structure, positioned below the pineal gland in the dorsal diencephalon, adjacent to the posterior and dorsal part of the thalamus. It was commonly regarded as consisting of two parts, the medial and lateral habenula (Marburg, Reference Marburg1944). However, a detailed anatomical analysis of the rodent habenula revealed a much more complex subnuclear structure with up to 15 subdivisions (Andres et al. Reference Andres, von During and Veh1999; Geisler et al. Reference Geisler, Andres and Veh2003). Medial and lateral habenular nuclei are cytoarchitectonically distinct, with very different input and output connections. According to Nauta and others, the habenula is one of a few brain structures where limbic and striatal mechanisms intermix directly (Nauta, Reference Nauta1958; Herkenham & Nauta, Reference Herkenham and Nauta1977, Reference Herkenham and Nauta1979; Sutherland, Reference Sutherland1982; Haun et al. Reference Haun, Eckenrode and Murray1992; Andres et al. 1992, Reference Andres, von During and Veh1999). Because of its structural complexity and numerous connections with other brain areas, the habenular complex is implicated in many different neurotransmitter and modulator systems including biogenic amines (dopamine, serotonin and acetylcholine: Wang & Aghajanian, Reference Wang and Aghajanian1977; Herkenham & Nauta, Reference Herkenham and Nauta1979; Lisoprawski et al. Reference Lisoprawski, Herve, Blanc, Glowinski and Tassin1980; Wooten et al. Reference Wooten and Collins1981; Reisine et al. Reference Reisine, Soubrie, Artaud and Glowinski1982; Contestabile & Fonnum, Reference Contestabile and Fonnum1983; London et al. Reference London, Waller and Wamsley1985; Thornton et al. Reference Thornton, Evans and Harris1985; Christoph et al. Reference Christoph, Leonzio and Wilcox1986; Caldecott-Hazard et al. Reference Caldecott-Hazard, Mazziotta and Phelps1988; Mori et al. Reference Mori, Diehl, Chauhan, Ljunggren and Kristensson1999), amino acids [glutamate/aspartate: Kiss et al. Reference Kiss, Csaki, Bokor, Kocsis and Kocsis2002; gamma-aminobutyric acid (GABA): Gallager, Reference Gallager1976; Gottesfeld et al. Reference Gottesfeld, Massari, Muth and Jacobowitz1977; Contestabile & Fonnum, Reference Contestabile and Fonnum1983], neuropeptides (substance P, vasopressin: Ronnekleiv, Reference Ronnekleiv1988; Lucas et al. Reference Lucas, Hurley, Krause and Harlan1992) and purines [adenosine triphosphate (ATP): Pankratov et al. Reference Pankratov, Lalo, Verkhratsky and Nort2006]. Functionally, numerous reports have linked the habenular complex with avoidance learning, regulation of feeding behaviour, maternal behaviour and stress (Gonzalez-Lima & Scheich, Reference Gonzalez-Lima and Scheich1986; Wilcox et al. Reference Wilcox, Christoph, Double and Leonzio1986; Thornton & Bradbury, Reference Thornton and Bradbury1989; Chastrette et al. Reference Chastrette, Pfaff and Gibbs1991; Corodimas et al. Reference Corodimas, Rosenblatt, Canfield and Morrell1993; Matthews-Felton et al. Reference Matthews-Felton, Corodimas, Rosenblatt and Morrell1995; Vale-Martinez et al. Reference Vale-Martinez, Marti-Nicolovius, Guillazo-Blanch, Coll-Andreu and Morgado-Bernal1997; Amat et al. Reference Amat, Sparks, Matus-Amat, Griggs, Watkins and Maier2001). As the habenular complex seems to serve as a crucial modulatory relay between limbic forebrain structures and the midbrain, regulating levels of dopamine and serotonin utilization in the telencephalon and diencephalon, it is reasonable to consider the habenula a candidate brain area involved in the development of schizophrenic and affective disorders (Scheibel, Reference Scheibel1997). Indeed, there are both good theoretical arguments and experimental evidence in favour of a putative role of the habenula in processes with relevance for neuropsychiatric disorders (i.e. affective disorder and schizophrenia). Thus, the habenular complex is implicated in a variety of processes known to be disrupted in depression, such as disturbed motor and cognitive behaviour, nociception, anxiety, reward, behavioural inhibition and sleep (Ronnekleiv et al. Reference Ronnekleiv, Kelly and Wuttke1980; Sutherland & Nakajima, Reference Sutherland and Nakajima1981; Goldstein, Reference Goldstein1985; Mahieux & Benabid, Reference Mahieux and Benabid1987; Benabid & Jeaugey, Reference Benabid and Jeaugey1989; Haun et al. Reference Haun, Eckenrode and Murray1992; Murphy et al. Reference Murphy, DiCamillo, Haun and Murray1996; Valjakka et al. Reference Valjakka, Vartiainen, Tuomisto, Tuomisto, Olkkonen and Airaksinen1998; Carlson et al. Reference Carlson, Noguchi and Ellison2001; Ellison, Reference Ellison2002; Ullsperger & von Cramon, Reference Ullsperger and von Cramon2003; Hikosaka et al. Reference Hikosaka, Sesack, Lecourtier and Shepard2008). A variety of drugs have been demonstrated to induce degeneration within the lateral habenula (e.g. by amphetamine or cocaine) and the medial habenula (nicotine), and in the fasciculus retroflexus when given continuously (for an overview, see Carlson et al. Reference Carlson, Armstrong, Switzer and Ellison2000, Reference Carlson, Noguchi and Ellison2001; Ellison, Reference Ellison2002).
Neuronal activity in the lateral habenula is strongly inhibited by dopamine agonists and disinhibited by dopamine antagonists such as haloperidol, an antipsychotic drug (McCulloch et al. Reference McCulloch, Savaki and Sokoloff1980; Caldecott-Hazard et al. Reference Caldecott-Hazard, Mazziotta and Phelps1988; Ellison, Reference Ellison1994). Moreover, Shumake et al. (Reference Shumake, Edwards and Gonzalez-Lima2003, Reference Shumake, Conejo-Jimenez, Gonzalez-Pardo and Gonzalez-Lima2004) studied congenitally helpless rats (a widely used animal model of depression) and found an elevated metabolism within the habenula and the interpeduncular nucleus. These authors also revealed an almost complete decoupling of limbic forebrain regions from midbrain and diencephalic regions in newborn congenitally helpless rats. In three other rat models of depression an elevated metabolism in the lateral habenula was shown that was accompanied by reduced exploratory behaviour (Caldecott-Hazard et al. Reference Caldecott-Hazard, Mazziotta and Phelps1988). Lesions of the habenula have also been shown to eliminate the induction of learned helplessness (Amat et al. Reference Amat, Sparks, Matus-Amat, Griggs, Watkins and Maier2001; Klemm, Reference Klemm2004). In a neuroimaging study using positron emission tomography (PET), a strong correlation between habenula activity and depression severity was demonstrated in human subjects (Morris et al. Reference Morris, Smith, Cowen, Friston and Dolan1999).
Accumulating evidence from experiments with animals led Sartorius & Henn (Reference Sartorius and Henn2007) to suggest that an overactivation of the habenula might play a crucial role in human depression, and that functional inhibition of the lateral habenula by deep brain stimulation might have antidepressive effects.
Unlike depression, there is only little evidence in favour of an involvement of the habenular complex in schizophrenia. It has been shown that in rats bilateral lesions of the structure induces schizophrenia-like behaviour (Lecourtier et al. Reference Lecourtier, Neijt and Kelly2004). Human functional imaging studies have demonstrated that the habenula is activated following receipt of unexpected negative feedback or the absence of expected positive feedback. In patients with schizophrenia, a lack of appropriate modulation of habenula activity in response to feedback has been described (Kelly, 1998; Shepard et al. Reference Shepard, Holcomb and Gold2006). Hence, the habenula might play an important role in mediating the feedback-processing deficits in schizophrenia.
All these findings give good reason to investigate the habenula in human neuropsychiatric disorders. Surprisingly, the habenular complex has been largely overlooked by neuropsychiatric research, except for the above-mentioned study of Morris et al. (Reference Morris, Smith, Cowen, Friston and Dolan1999) and a computed tomography (CT) study by Sandyk (Reference Sandyk1992) that demonstrated an increased prevalence of habenular calcification in schizophrenia. One reason for that might be our limited knowledge on this brain structure in normal human brains with regard to its anatomy and physiology. To our knowledge, there is only one paper dealing with this topic (Marburg, Reference Marburg1944). The aim of the current study was therefore to investigate the morphology of the human habenular complex at a light microscopic level in patients with schizophrenia and depression and in normal control subjects.
Method
Subject characteristics
All brains were from the Magdeburg Brain Collection. Sampling of the human brain material and asservation was carried out in accordance with the Declaration of Helsinki (1964), German law and approval by the local ethics commission.
Brains were collected from 14 patients with mood disorder (eight women, six men). The age range was 26 to 69 years (mean 48.6±12.8 years). Of these 14, nine had died by suicide, including five women and four men; six patients displayed major depression and eight bipolar depression. We also investigated brains of 17 patients with schizophrenia (nine women and eight men) with an age range of 39 to 66 years (mean 52.6±7.5 years). Control brains were collected from five women and eight men with an age range of 38 to 66 years (mean 55.9±8.9 years). All control subjects died of natural causes. The demographic data for all subjects are summarized in Table 1.
Characteristics of subjects
F, Female; M, male; s.d., standard deviation.
All patients were matched for age, gender and post-mortem delay. The matching processes were carried out before the experimental procedures and quantitative analyses. Information for clinical diagnosis was obtained by the careful study of clinical records and by structured interviews with people who either lived with, or had frequent contact with, the subjects before death. Structured interviews were carried out to collect information to determine the presence or absence of psychiatric disorders. Data on lifetime and current mental illness were gathered with the Lifetime Version of the Schedule for Affective Disorders and Schizophrenia (SADS-L; Spitzer & Endicott, Reference Spitzer and Endicott1978). Final Axis I diagnoses were assigned in consensus meetings of two psychiatrists using all available information from informants and clinical charts. The final diagnosis was compatible with DSM-III-R (APA, 1987). In the same way, neuropsychiatric disorders were excluded in the control subjects. There was no current or lifetime psychoactive substance disorder (abuse or dependence according to DSM-III-R) in any of the subjects. All patients suffering from depressive syndrome had received long-term treatment with antidepressants. In addition, one of them had received neuroleptic drugs. All bipolar patients but one had received lithium. All schizophrenia patients had been treated with typical neuroleptics (haloperidol).
Qualitative neuropathological changes due to neurodegenerative disorders (such as Alzheimer's disease, Parkinson's disease, Pick's disease), tumours, inflammatory, vascular or traumatic processes were ruled out by an experienced neuropathologist (for demographical, histological and clinical data see Table 1).
Tissue processing and histology
All brains were obtained from pathologists or medical examination officers. Tissue preparation was performed as described previously (Bernstein et al. Reference Bernstein, Stanarius, Baumann, Henning, Krell, Danos, Falkai and Bogerts1998, Reference Bernstein, Baumann, Danos, Dieckmann, Bogerts, Gundelfinger and Braunewell1999). Brains were removed within 4 to 72 h after death and fixed in toto in 8% phosphate-buffered formaldehyde for at least 2 months. Frontal and occipital poles were separated by coronal cuts anterior to the genu and posterior to the splenium of the corpus callosum. After embedding all parts of the brains in paraffin, serial coronal sections of the middle block were cut on a microtome at 20 μm and mounted. Every 50th section was stained for anatomical orientation and morphometric investigations with a combined cell and fibre staining according to Nissl (Cresyl Violet) and Heidenhain–Woelcke procedures (Zech et al. Reference Zech, Roberts, Bogerts, Crow and Polak1986). Volume shrinkage was determined for each brain before and after dehydration and embedding of tissue. Volume shrinkage factors were calculated using the formula: VF=(A1/A2)3/2, where VF is the volume shrinkage factor, A1 the cross-sectional area before processing the tissue, and A2 the cross-sectional area after processing the tissue. Sections were taken at intervals of 400 μm (according to Cavalieri's principle; Mayhew, Reference Mayhew1992).
Morphometry and stereology
Seven sections per case were selected randomly for the analysis of the habenular complex. A computerized image analysis system (Digitrace®, Imatec, Germany) was used. It was attached to a high-resolution video camera on a Leica microscope, equipped with a motorized scanning stage. For the identification and delineation of the habenular, complex images of each side were taken using a fourfold magnification. These composed images were scanned by a digital camera and loaded by the analysis system. At a higher magnification the entire area of the habenular complex on the right or the left side was visible and allowed delineation of the medial and lateral habenular nuclei (Fig. 1). Volumes of both subnuclei were calculated from areas measured on the videoscreen, performing morphometrical operations described previously in detail (Bogerts et al. Reference Bogerts, Falkai, Haupts, Greve, Ernst, Tapernon-Franz and Heinzmann1990).
Overview of the medial (MHB) and lateral (LHB) habenula of the right and left side with adjacent structures. Aq, aqueduct, ant, anterior, pc, posterior commissure, hbc, habenular commissure. Combined cell and fibre with Nissl (Cresyl Violet) and Heidenhain–Woelcke procedures (magnification×1.25).
All further investigations (cell countings) were performed within the medial habenular nucleus only (see limitations of the study). The area of the medial habenular nucleus was divided into boxes of 250×250 μm using a counting grid. Every 20th box was selected randomly and all neurones within these boxes were counted manually. In each section, neurones of the medial habenular nucleus were counted by the optical dissector method using a ×40 objective in the dissector fields, which were selected using a systematic random sampling along the whole cross-sectional area of the medial habenular nucleus (Bernstein et al. Reference Bernstein, Stanarius, Baumann, Henning, Krell, Danos, Falkai and Bogerts1998). Application of the optical dissector made it necessary to measure movements in the z axis, which was performed by using a microcator as an integral part of the microscope. Only neurones in which a nucleolus was clearly visible were counted. Measurements of volumes and neurone counts were performed blind to the diagnosis.
Statistical analyses
All data are presented as mean±standard deviation (s.d.). A single-factor analysis of variance (ANOVA) was performed using diagnostic groups as a three-level independent variable (mood disorder patients versus schizophrenia patients versus non-psychiatric controls) and measured and calculated parameters (e.g. mean cell area) as dependent variables. Post-hoc Tukey HSD tests were performed to detect two-group differences. Pearson's correlation coefficients were calculated to determine the influence of demographic, histological and clinical variables such as age at time of death, brain weight, post-mortem delay, duration of illness or psychiatric medication, which might confound the results of the dependent variables. p values <0.05 were defined as statistically significant.
Results
Volumetry
Medial habenula
When estimating the volumes of habenular nuclei, we found in the medial habenula of the depressive patients a volume of 2.76±0.90 mm3 for the right side and 2.65±0.65 mm3 for the left side. There was a significant difference (p=0.025) for the right side in comparison to controls, which displayed a volume of 3.64±0.97 mm3 for the right side and 3.35±1.33 mm3 for the left side. There was a significant reduction of 24.1% on the right side (and 20.9% with a trend towards significance on the left, see Fig. 2 a).
Volumes of the (a) medial (mHb) and (b) lateral (lHb) habenular complex; (c) neuronal cell number and (d) neuronal cell areas of the medial habenula. Significant reductions of all parameters measured were found only in depressive patients (right side). □, Controls; 
, depressive patients; ▪, schizophrenia patients. * p⩽0.05, ** p⩽0.01, *** p⩽0.001.
Summary of results
Con, Control; Dep, depressive; Sz, Schizophrenia.
Values given as mean±standard deviation. Significant p values are in bold.
All interactions hemisphere×diagnosis were non-significant, therefore no single test per hemisphere was performed.
There was no difference between schizophrenia patients and controls with regard to habenular volumes (determined volumes in schizophrenia were 3.69±0.91 mm3 for the right side and 3.27±0.81 mm3 for the left side). However, the habenula as a whole had a different shape in schizophrenia compared to controls, in that it was plumper (shorter in the rostro-caudal dimension and larger in diameter). However, when comparing the volumes of the affective disorder group with those of schizophrenia patients, a significant reduction for both sides (p=0.01 and p=0.029) was revealed for the depressed patients. We failed to detect any differences of the left and right side within the groups (Table 2).
Lateral habenula
As shown in Fig. 2 b, we found a significant difference of the right side between depressive patients (right side 23.66±6.61 mm3 and left side 24.91±5.23 mm3) and controls (right side 29.59±4.83 mm3 and left side 27.57±5.05 mm3), with a reduction of 20.0% (p=0.014). There was also a significant reduction of the right side between schizophrenia and depressive patients of 26.7%, but no significant difference of schizophrenia patients in comparison to controls (right side 29.98±5.03 mm3, left side 28.56±5.70 mm3).
Cell number, cell area and cell density
Neuronal cell numbers
The estimated total cell numbers were 210731±54972 (right side) and 187142±88447 (left side) in controls, 137886±34729 (right side) and 129091±25945 (left side) in patients with affective disorder, and 206751±68261 (right side) and 168984±67978 (left side) in schizophrenia patients. Significant differences were found for the depressive patients in comparison to controls, with a reduction of 34.6% (p=0.000) and 31.0% (p=0.027) for the right and left side respectively, and between patients with affective disorder and schizophrenia, with a reduction of 49.9% (p=0.003) of the right side and 30.9% (p=0.047) of the left side. No differences were seen between schizophrenia patients and controls (Fig. 2 c).
Neuronal cell areas
We found an area of 0.236×106 μm2 on the right and 0.206×106 μm2 per 1 mm3 tissue on the left of controls, 0.142×106 μm2 on the right and 0.135×106 μm2 per 1 mm3 tissue on the left of the depressive patients, and 0.218×106 μm2 and 0.189×106 μm2 of the schizophrenia patients respectively. There was a significant reduction on both sides in depressive patients in comparison to controls of 39.6% (p=0.000) and 34.4% (p=0.016) for right and left respectively, and between patients with affective disorder and schizophrenia patients with 53.5% (p=0.002) and 39.8% (p=0.018) (see Fig. 2 d).
Neuronal cell densities
No significant differences were found between the groups. We could not detect any differences between patients with major depression and bipolar depression.
Confounding factors
Variables that could influence neurone numbers and volumes, such as brain weight, age at time of death, post-mortem delay or times of fixation, and duration of medication did not correlate with the parameters measured.
Methodical limitations
A limitation of this study is the lack of data about cell densities in the lateral habenula. Although we are aware that knowledge of cell numbers from the lateral habenula would be extremely interesting and helpful, we had to restrict ourselves on the medial part for methodical reasons. To properly determine volumes we had to delineate the habenula and its subdivisions (Herkenham & Nauta, Reference Herkenham and Nauta1977). This can be done best using a combined cell–fibre staining technique. Unfortunately, a consequence of this approach was a partial masking of stained neuronal cell bodies in the lateral habenula (but not in the medial part) by very dense fibre tracts, which hampered correct countings. Studies are in progress to overcome this problem.
A second important issue is the long-term treatment of the patients with antipsychotics (schizophrenia) or antidepressants (depression). Although our own calculations did not reveal significant correlations between the duration and/or dose of medication and volumes or cell numbers, others have found that haloperidol (Dorph-Petersen et al. Reference Dorph-Petersen, Pierri, Perel, Sun, Sampson and Lewis2005; Lieberman et al. Reference Lieberman, Tollefson, Charles, Zipursky, Sharma, Kahn, Keefe, Green, Gur, McEvoy, Perkins, Hamer, Gu and Tohen2005) and antidepressants (Young et al. Reference Young, Bonkale, Holcomb, Hicks and German2008) may affect brain volumes and should be taken into account as possible confounding factors in post-mortem studies.
Discussion
The limbic system has long been implicated in the pathogenesis of schizophrenia and affective disorders, and structural abnormalities have been found in many limbic areas (for reviews, see Falkai et al. Reference Falkai, Bogerts and Rozumek1988; Campbell & MacQueen, Reference Campbell and MacQueen2004; Dietrich et al. Reference Dietrich, Bonnenmann and Emrich2007; Geisler & Trimble, Reference Geisler and Trimble2008; White et al. Reference White, Cullen, Rohrer, Karatekin, Luciana, Schmidt, Hongwanishkul, Kumra, Schulz and Lim2008; Schmitt et al. Reference Schmitt, Steyskal, Bernstein, Schneider-Axmann, Parlapani, Schaeffer, Gattaz, Bogerts, Schmitz and Falkai2009). Very little is known about the limbic habenular nuclei in neuropsychiatric disorders, however. Although the need for studies on the habenula in the context of psychiatric diseases such as schizophrenia and depression has clearly been recognized (Sandyk, Reference Sandyk1992; Scheibel, Reference Scheibel1997; Deutsch et al. 2000; Klemm, Reference Klemm2004; Sartorius & Henn, Reference Sartorius and Henn2007; Hikosaka et al. Reference Hikosaka, Sesack, Lecourtier and Shepard2008), there are only two studies on this topic (Morris et al. Reference Morris, Smith, Cowen, Friston and Dolan1999; Shepard et al. Reference Shepard, Holcomb and Gold2006). The current study is, to our knowledge, the first morphometric study on the human habenular complex in schizophrenia and depression. Our data show reduced habenular volumes and cell numbers in patients with affective disorder, indicating distinct disease-related structural alterations of this brain region. This is in agreement with previous data of the human neuroimaging (PET) study that found a strong correlation between habenula activity and severity of depression (Morris et al. Reference Morris, Smith, Cowen, Friston and Dolan1999). Recent investigations show that depression is associated with an increase in the activation of the lateral habenular nucleus, which is accompanied by down-regulation of monaminergic systems. Deep brain stimulation is being applied increasingly to cure treatment-resistant depressive patients by reducing this overactivation (reviewed in Hauptman et al. Reference Hauptman, DeSalles, Espinoza, Sedrak and Ishida2008). This increased habenular activity in patients with affective disorder is difficult to explain on grounds of reduced volumes and cell numbers only. Rather, changes in the chemical composition of the habenula would be expected to contribute significantly to this altered activity pattern in depression. Unfortunately, very little is known about the regional distribution and cellular localization of neurotransmitters and modulators in the human habenula. Our own unpublished data on altered cellular expression of choline acetyltransferase and glutamate decarboxylase in habenular neurones of depressive patients are in favour of this assumption.
Almost all available data for the habenular complex originate from animal studies. In animal models of depression tested so far, an abnormally elevated habenular metabolism was found (Caldecott-Hazard et al. Reference Caldecott-Hazard, Mazziotta and Phelps1988; Shumake et al. Reference Shumake, Edwards and Gonzalez-Lima2003).
Stress experiments demonstrate a role of the habenular complex in mediating the neurochemical and behavioural responses to chronic stress. Moreover, the habenula seems to be necessary for the induction of learned helplessness/behavioural depression (Amat et al. Reference Amat, Sparks, Matus-Amat, Griggs, Watkins and Maier2001).
The raphe nuclei are implicated in the pathophysiology of depression (Bielau et al. Reference Bielau, Mawrin, Krell, Agelink, Trübner, Davis, Gos, Bogerts, Bernstein and Baumann2005). Morris et al. (Reference Morris, Smith, Cowen, Friston and Dolan1999) and Hikosaka et al. (Reference Hikosaka, Sesack, Lecourtier and Shepard2008) proposed that the habenula represents a point of convergence in a feedback loop that controls raphe activity. A recently published lesion experiment suggests that the lateral habenula might be a necessary structure for the induction of behavioural depression in the rat, and that activities of the lateral habenula are involved in depressive disorders (Yang et al. Reference Yang, Hu, Xia, Zhang and Zhao2008). Lateral habenula lesions improved behavioural responses of depressed rats most probably by increasing serotonin levels in the dorsal raphe nucleus (Hikosaka et al. Reference Hikosaka, Sesack, Lecourtier and Shepard2008; Yang et al. Reference Yang, Hu, Xia, Zhang and Zhao2008).
Mainly based on genetic findings, bipolar depression is currently discussed as a more distinct disease that is possibly more closely related to schizophrenia than to major depressive disorder (for a review on this topic, see Möller, Reference Möller2003). However, in the case of the habenula, bipolar patients show structural alterations that are similar to those in unipolar depression, but not to schizophrenia. Most of our results were obtained from the medial habenula, which plays an important role in the septo-habenulo-interpeduncular pathway. Efferences of the medial habenula are travelling primarily to the interpeduncular nucleus (Ramon y Cajal, Reference Ramon y Cajal1911; Nauta, Reference Nauta1958; Akagi & Powell, Reference Akagi and Powell1968; Herkenham & Nauta, Reference Herkenham and Nauta1977), which is reciprocally connected to many midbrain areas. The medial habenula itself has some minor connections with the median raphe, the ventral tegmental area (VTA) and some other areas. In congenitally helpless rats the medial and lateral habenular nuclei were metabolically hyperactive and the VTA activity was diminished, which suggests that the habenula might play a role in depression and reflect a defective dopamine innervation (Shumake et al. Reference Shumake, Edwards and Gonzalez-Lima2003).
In general, most interest is directed to the lateral habenula (reviewed in Geisler & Trimble, Reference Geisler and Trimble2008). We found a strongly reduced volume of the lateral habenula, but additional investigations at the cellular level are necessary to learn more about the structural impact of this part of the habenular complex in neuropsychiatric disorders, especially because the lateral habenula has a very complex subnuclear structure with abundant input and output patterns (Andres et al. Reference Andres, von During and Veh1999; Geisler et al. Reference Geisler, Andres and Veh2003).
We would have expected more differences in schizophrenia patients compared to controls according to the often mentioned changes in the dopamine system. In a motion prediction test (a human MRI study) with positive and negative feedback, blood flow to the habenula decreased following successful trials and increased following failures. This again indicates that the habenula might inhibit dopamine activity in response to failure (Ullsperger & von Cramon, Reference Ullsperger and von Cramon2003).
The habenular complex seems to include a wide range of neurotransmitters in its trajectory, from its septal–basal forebrain origins to the habenula, interpeduncular nucleus and midbrain tegmentum. These neurotransmitters include glutamate, aspartate, acetylcholine, GABA, and opioid peptides. However, on the morphological basis we could only detect subtle but not significant differences in schizophrenia patients in comparison to controls. More interesting, we found robust and statistical significant difference between schizophrenia and depressive patients. We assume there might be disturbances on other levels, for example within subnuclei or certain neuronal groups. Further immunohistochemical investigations are necessary to relate different neurotransmitter to those alterations or search for targets in schizophrenia.
The basic picture that emerges is that of a robust feedback control system regulating levels of dopamine and serotonin utilization in the telencephalon and diencephalon, unique in combining and directing striatal and limbic output caudally on major midbrain sources of monaminergic and cholinergic innervation of forebrain structures (Nauta, Reference Nauta1974; Scheibel, Reference Scheibel1997). These neurotransmitter systems are crucial in the modulation of affect and behaviour and have been implicated in schizophrenia and affective psychoses (Scheibel, Reference Scheibel1997).
In summary, we have shown a reduced volume of the medial and lateral habenular complex and a reduced cell number and area of the medial habenula in patients with affective disorder in comparison to controls and schizophrenia patients.
Acknowledgements
We thank S. Funke and R. Stauch for excellent technical assistance. This study was supported by The Stanley Foundation and Graduiertenkolleg der Deutschen Forschungsgemeinschaft ‘Biologische Grundlagen von Erkrankungen des Nervensystems’.
Declaration of Interest
None.
