Functional anatomical correlates of antidepressant drug treatment assessed using PET measures of regional glucose metabolism

https://doi.org/10.1016/S0924-977X(02)00102-5Get rights and content

Abstract

Neurophysiological studies of major depression performed using PET imaging have shown abnormalities of regional cerebral blood flow (CBF) and glucose metabolism in multiple prefrontal cortical and limbic structures that have been more generally implicated in emotional processing. The current study investigated the effects of antidepressant drug treatment in these regions using PET measures of glucose metabolism. Subjects with primary MDD (n=27) were imaged while unmedicated and depressed, and, of these, 20 were rescanned following chronic antidepressant drug treatment. Regional metabolism was compared between unmedicated depressives and controls and between the pre- and post-treatment conditions in regions-of-interest (ROI) where metabolism or flow had previously been shown to be abnormal in unmedicated depressives. At baseline, the mean metabolism was increased in the left and right lateral orbital cortex/ventrolateral prefrontal cortex (PFC), left amygdala, and posterior cingulate cortex, and decreased in the subgenual ACC and dorsal medial/dorsal anterolateral PFC in the unmedicated depressives relative to controls, consistent with the results of previous studies. Following treatment, metabolism significantly decreased in the left amygdala and left subgenual ACC, and corresponding changes in the orbital and posterior cingulate cortices approached significance. The metabolic reduction in the amygdala and right subgenual ACC appeared largely limited to those subjects who both responded to treatment and remained well at 6 months follow-up, in whom the reduction in amygdala metabolism tightly correlated with the reduction in HDRS scores. The magnitude of the treatment-associated, metabolic change in the amygdala also correlated positively with the change in the stressed plasma cortisol levels measured during scanning. These data converge with those from other PET studies to indicate that primary MDD is associated with abnormal metabolism in limbic and paralimbic structures of the mesiotemporal and prefrontal cortices. Chronic antidepressant drug treatment reduces metabolism in the amygdala and ventral ACC in subjects showing a persistent, positive treatment response. In contrast, the persistence of the abnormal metabolic deficits in the dorsomedial/dorsal anterolateral PFC in MDD during treatment may conceivably relate to the histopathological changes reported in these regions in post mortem studies of MDD.

Introduction

Neuroimaging studies of major depressive disorder (MDD) have provided invaluable information about the anatomical systems involved in depression. Positron emission tomography (PET) studies of cerebral blood flow (CBF) and glucose metabolism in depressed subjects demonstrate that neurophysiological activity is abnormal in several brain structures that have been shown by other types of evidence to participate in the modulation of emotional behavior. The relationships between metabolism and depression severity in these regions suggest that, in some structures, the abnormal neural activity is positively correlated with depressive symptoms, whereas, in others, activity may instead comprise a compensatory response that modules such symptoms (reviewed in Drevets, 2001).

The neural circuits implicated by PET studies of depression involve anatomical loops between the medial and orbital prefrontal cortex (PFC) and anatomically related areas of the mesiotemporal cortex, striatum, and thalamus (reviewed in Drevets, 2000, 2001; Drevets et al., 2002; Öngür and Price, 2000). In the PFC, most studies comparing unmedicated depressed subjects with primary MDD have reported elevated CBF and metabolism in the lateral orbital/ventrolateral PFC, the anterior cingulate cortex (ACC) anterior to the genu of the corpus callosum (i.e., pregenual ACC) and the anterior insula during the eyes-closed, at-rest condition (reviewed in Drevets, 2000; Brody et al., 2002). These areas share extensive, reciprocal anatomical connections to the amygdala, which has also been shown to have abnormally increased activity in MDD (reviewed in Drevets et al., 2002; Öngür and Price, 2000). Elevated physiological activity has also been reported in the posterior cingulate cortex, which sends extensive projections to the pregenual ACC, the anteroventral striatum (Wilson et al., 2002), which encompasses the accumbens area and would receive projections from the amygdala, the orbital and medial PFC (Drevets et al., 2001; Öngür and Price, 2000; Russchen et al., 1985), and the medial thalamus, within which the mediodorsal nucleus also shares reciprocal projections with the amygdala and PFC (Drevets et al., 1992, Drevets et al., 1995c; Price et al., 1996).

In addition to these areas of increased metabolic activity, areas of reduced CBF and metabolism were found in the ACC ventral to the genu of the corpus callosum (i.e., ‘subgenual’ ACC; Drevets et al., 1997) and the dorsomedial/dorsoanterolateral PFC in depressives relative to controls (Bell et al., 1999, and replicating Baxter et al., 1989 and Bench et al., 1992). The subgenual ACC and the dorsal anterolateral PFC were subsequently shown to contain abnormal reductions in cortex volume and/or histopathological changes in MDD and BD by in vivo, morphometric MRI and post mortem neuropathological studies. The reductions in physiological activity seen in PET images from MDD and BD samples may thus be accounted for by structural abnormalities of the corresponding cortex.

The current study assessed metabolic effects of antidepressant drug treatment in regions-of-interest (ROI) where abnormalities of basal metabolism were reported in MDD by Drevets et al., 1997, Drevets et al., 2002, Drevets et al., 1992. The ROI selected for comparing metabolism between pre- and post-treatment scans were defined in a manner that would avoid biasing results toward deviations from the mean in the pre-treatment group. Some of the differences reported between depressed and control samples in previous studies were detected using voxel-by-voxel approaches which optimized the localization of peak, intergroup differences in mean CBF or metabolism relative to the variance by computing statistical parametric images consisting of t- or Z-scores (Drevets et al., 1997, Drevets et al., 1992). These image data were thus sensitive to the noise distribution as well as to the mean intergroup differences in physiology (Drevets et al., 1992; Friston et al., 1991). The application of ROI defined based upon such peak differences between depressives versus controls could, therefore, introduce bias in the results of pre- versus post-treatment studies performed in the same subject sample (e.g., changes in the post-treatment scans obtained using such an approach could simply reflect regressions to the population mean).

The ROI of primary interest for brain structures studied herein were thus defined using either atlas-based, stereotaxic coordinates of peak differences between depressives and controls which had previously been identified in voxel-by-voxel analyses of image data from independent samples of depressives and controls (Drevets et al., 1992; Price et al., 1996; Talairach and Tournoux, 1988), or ROI defined on specific grey matter structures in MRI images that had been co-registered to the corresponding PET images (Drevets et al., 1997). The stereotaxic approach was used to assess treatment effects in the amygdala, while PET-MRI co-location was employed to investigate treatment-associated changes in the subgenual PFC (as described in Drevets et al., 1997) and lateral orbital cortex.

The lateral orbital ROI was positioned within the area of the left ventrolateral PFC where Drevets et al. (1992) demonstrated increased CBF in depressives versus controls (in samples independent from those studied herein). In this previous study a statistical parametric image (t-value image) was used to delimit an area where CBF inherently differed between one sample of depressives and controls, and the statistical significance of this difference between depressives and controls was then established by data obtained in the same ROI in a second, independent set of depressives and controls. This earlier study involved lower resolution PET images (PET VI measures of H215O uptake) than those available for the current study, however, so the vental PFC ROI delimited in Drevets et al. (1992) required modification in order to assess cortical glucose metabolism in the higher resolution images employed herein. This ventral PFC ROI where CBF had been abnormally elevated in depression in Drevets et al. (1992) included portions of the lateral orbital cortex, the ventrolateral PFC, the anterior insula, the pregenual ACC, and the frontal polar cortex (Drevets et al., 1992). An ROI defined to encompass this anatomical extent which also accommodated the higher spatial resolution FDG images acquired on the 953B scanner by narrowing its spatial extent was previously assessed in MDD by Drevets et al. (1995c). Metabolism was elevated in this area in depressives vs. controls, and decreased following antidepressant treatment (−5.4±10.0%, P<0.05) in a subsample of the subjects described herein. However, because of the marked anatomical variability of the orbital surface contour across humans (e.g., currently available spatial transformation algorithms have no ability to address the problem that the orbital cortex is convex in some subjects but concave in others), the coefficient of variance of these data was artifactually increased, as the stereotaxically positioned ROI extended outside the grey matter in some subjects.

In a subsequent analysis, Drevets et al. (1996) thus defined a more restricted PFC ROI in the posterior orbital cortex, stereotaxically defined using spherical volumes of 14 mm diameter (centered at x=±17, y=33, z=−16). The metabolism was also significantly elevated in this ROI in unmedicated depressives versus controls, and decreased during antidepressant drug treatment (−6.8±11%, P<0.02), with the reduction in metabolism and HDRS scores correlated at r=0.47 (P<0.05) in the MDD sample described herein. However, this ROI also extended beyond the brain edge in some subjects, leading to a high coefficient of variance across subjects.

Consequently, to more accurately and reliably measure cortical metabolism in a manner that addressed the variable contour of the orbital surface, the current study defined an orbital cortex ROI directly on each subject’s anatomical MRI image and extracted the corresponding metabolic data from coregistered PET images. This ROI sampled part of the lateral orbital area encompassed within the larger ROI from our original study (Drevets et al., 1992), and was more specifically defined to encompass the area where Rajkowska et al. (1999) reported abnormal reductions of grey matter volume and glial cells in a MDD sample studied post mortem. This histopathological pattern has also characterized the left subgenual PFC (Drevets et al., 1998; Öngür et al., 1998) and the amygdala (Bowley et al., 2002) in post mortem studies, and is hypothesized to be related to the hypermetabolic activity in these regions in PET studies (see Discussion).

An alternative approach was employed to address treatment effects in the dorsomedial/dorsal anterolateral PFC and the posterior cingulate cortex where abnormalities of regional CBF and metabolism had been reported in MDD, but where the specific localization of intergroup differences remained ambiguous. The spatial locations of reported differences between depressives and controls has varied so widely across studies in these areas that new studies performed in independent subject samples to assess treatment effects have had difficulty replicating the originally described abnormalities in ROI selected a priori. For example, using ROI predefined directly on PET images, Baxter et al. (1989) reported a reduction in the dorsal anterolateral PFC metabolism in depressives versus controls that reversed toward the normative baseline following antidepressant drug treatment. However, the same laboratory was unable to replicate the baseline abnormality in independent subject samples using ROI that were presumably positioned in the same area of the PFC in anatomical MRI images, and then transferred to the corresponding PET images to assess treatment effects (Brody et al., 2001; Saxena et al., 2002). Similarly, although several studies reported abnormalities of CBF or metabolism in the posterior cingulate cortex, the specific locations of these findings have differed widely across studies and have even differed in direction, being increased in most but reduced in some studies of depressives versus controls (reviewed in Drevets, 2000).

To assess treatment effects in such areas where the location of abnormalities in MDD remained unclear, a novel approach was applied in which the stereotaxic coordinates for areas implicated in interactions between emotional and cognitive processing were employed to guide ROI placement (Drevets and Raichle, 1998; Simpson et al., 2000). PET and fMRI studies of healthy humans have demonstrated several limbic and paralimbic cortical regions implicated in emotional behavior where hemodynamic activity consistently decreases as subjects perform attentionally demanding cognitive tasks (Shulman et al., 1997; Drevets and Raichle, 1998). In many of these regions, in contrast, hemodynamic activity increases during experimentally induced emotional states in healthy humans. The reciprocal patterns of the CBF changes in these limbic and paralimbic regions during attentionally demanding cognitive tasks were thus hypothesized to reflect interactions between cognitive and emotional processing (Drevets and Raichle, 1998).

Many of these regions closely correspond to areas where resting CBF and metabolism are abnormal during depression. Thus, the deactivation loci described by Shulman et al. (1997) in the amygdala, lateral orbital cortex, and subgenual PFC are encompassed within the ROI where we localized CBF and metabolism abnormalities in depression (Drevets et al., 1992, Drevets et al., 1995c, 1997, 2002). Moreover, the deactivation loci in the dorsomedial/dorsal anterolateral PFC and posterior cingulate areas from Shulman et al. (1997) were situated in the vicinity of CBF and metabolic abnormalities reported in studies of depression (reviewed in Drevets, 2001). We thus hypothesized that the stereotaxic coordinates for the deactivation loci of Shulman et al. (1997), which were carefully localized and replicated in large, independent samples of healthy humans, might guide ROI placement to specific areas where metabolism is abnormal in depression.

Section snippets

Subjects

Currently depressed subjects aged 18 to 59 who met DSM-IV criteria for recurrent MDD (APA, 1994) were recruited from the clinical services affiliated with Washington University School of Medicine. Subjects provided informed consent, as approved by the Washington University School of Medicine Institutional Review Board. Exclusion criteria included the presence of major medical and neurological disorders, history of other psychiatric disorders prior to MDD onset, history of mania, treatment with

Subjects

The mean age, gender composition, HDRS score and handedness [Edinburgh Handedness Inventory (Raczkowski et al., 1974)] of the subject samples appear in Table 1. Fourteen of the MDD subjects met the Winokur (1982) criteria for familial pure depressive disease (FPDD). None of the depressed subjects had a history of psychosis. One subject could not undergo MRI due to metal inside the head, so this subject’s data was used in the stereotaxic ROI assessments but not in the MRI-based ROI assessments.

Antidepressant treatment effects

Discussion

In the unmedicated phase of MDD, metabolism was abnormally increased in the left lateral orbital cortex, left amygdala, and posterior cingulate cortex, and decreased in the subgenual ACC and dorsal medial/anterolateral prefrontal cortex (PFC) in the unmedicated depressed relative to the control samples, consistent with the results of previous studies (see below). Following treatment, metabolism significantly decreased in the left amygdala and left subgenual ACC, and corresponding changes in the

Conclusion

These results converge with those of other PET studies, lesion analyses and post mortem studies to support a neural model in which the signs and symptoms of the major depressive syndrome involve dysfunction of modulatory systems in the prefrontal cortex, striatum and brainstem. Antidepressant therapies may compensate for dysfunction by attenuating this disinhibited, pathological limbic activity by augmenting monoamine neurotransmission and altering neuroreceptor sensitivity at various points in

Uncited references

Bechara et al., 1998; Casey et al., 2002; Chimowitz et al., 1992; Damasio et al., 1990; Davis, 1992; DiRocco et al., 1989; Drevets and Botteron, 1997; Drevets et al., 1995a; Drevets and Todd, 1997; Eastwood and Harrison, 2002; Fazekas, 1989; Folstein et al., 1991; Francis et al., 1989; Iversen and Mishkin, 1970; Nofzinger et al., 1999; Rubin et al., 1966; Teneback et al., 1999; Wooten and Collins, 1981; Young et al., 1993

Acknowledgements

The authors thank Joseph L. Price and Gordon Shulman for scientific discussions related to the methods and the interpretation of the results. Supported by NIH grants MH00928 and MH51137.

References (140)

  • W.C. Drevets et al.

    Amphetamine-induced dopamine release in human ventral striatum correlates with euphoria

    Biol. Psychiatry

    (2001)
  • W.C. Drevets et al.

    Glucose metabolism in the amygdala in depression: relationship to diagnostic subtype and stressed plasma cortisol levels

    Pharmacol. Biochem. Behav.

    (2002)
  • D. Ebert et al.

    Effects of sleep deprivation on the limbic system and the frontal lobes in affective disorders: a study with Tc-99m-HMPAO SPECT

    Psychiatry Res. Neuroimaging

    (1991)
  • S. Feldman et al.

    Differential effects of amygdaloid lesions on CRF-41, ACTH and corticosterone responses following neural stimuli

    Brain Res.

    (1994)
  • R.J. Frysztak et al.

    The effect of medial frontal cortex lesions on cardiovascular conditioned emotional responses in the rat

    Brain Res.

    (1994)
  • J.C. Gerber et al.

    The effect of antidepressant drugs on regional cerebral glucose utilization in the rat

    Brain Res.

    (1983)
  • J.P. Herman et al.

    Neurocircuitry of stress: central control of the hypothalamo–pituitary–adrenocortical axis

    Trends Neurosci.

    (1997)
  • T.A. Kimbrell et al.

    Regional cerebral glucose utilization in patients with a range of severities of unipolar depression

    Biol. Psychiatry

    (2002)
  • J.R. MacFall et al.

    Medial orbital frontal lesions in late onset depression

    Biol. Psychiatry

    (2001)
  • R.J. Maddock

    The retrosplenial cortex and emotion: new insights from functional neuroimaging of the human brain

    Trends Neurosci.

    (1999)
  • E.F. Nofzinger et al.

    Changes in forebrain function from waking to REM sleep in depression: preliminary analyses of [18F]FDG PET studies

    Psychiatry Res. Neuroimaging

    (1999)
  • G. Nowak et al.

    Alterations in the N-methyl-d-aspartate (NMDA) receptor complex in the frontal cortex of suicide victims

    Brain Res.

    (1995)
  • B. Baumann et al.

    Reduced volume of limbic system-affiliated basal ganglia in mood disorders: preliminary data from a post mortem study

    J. Neuropsychiatry Clin. Neurosci.

    (1999)
  • L.R. Baxter et al.

    Local cerebral glucose metabolic rates in obsessive-compulsive disorder—A comparison with rates in unipolar depression and in normal controls

    Arch. Gen. Psychiatry

    (1987)
  • L.R. Baxter et al.

    Cerebral metabolic rates for glucose in mood disorders

    Arch. Gen. Psychiatry

    (1985)
  • L.R. Baxter et al.

    Reduction of prefrontal cortex glucose metabolism common to three types of depression

    Arch. Gen. Psychiatry

    (1989)
  • A. Bechara et al.

    Dissociation of working memory from decision-making within the human prefrontal cortex

    J. Neurosci.

    (1998)
  • K.A. Bell et al.

    Decreased glucose metabolism in the dorsomedial prefrontal cortex in depression

    Biol. Psychiatry

    (1999)
  • C.J. Bench et al.

    Changes in regional cerebral blood flow on recovery from depression

    Psychol. Med.

    (1995)
  • C.J. Bench et al.

    The anatomy of melancholia—focal abnormalities of cerebral blood flow in major depression

    Psychol. Med.

    (1992)
  • C.J. Bench et al.

    Regional cerebral blood flow in depression measured by positron emission tomography: the relationship with clinical dimensions

    Psychol. Med.

    (1993)
  • K.J. Black et al.

    Baboon basal ganglia stereotaxy using internal MRI landmarks: validation and application to PET imaging

    J. Comput. Assist. Tomogr.

    (1997)
  • D.M. Bowen et al.

    Circumscribed changes of the cerebral cortex in neuropsychiatric disorders of later life

    Proc. Natl. Acad. Sci.

    (1989)
  • Bowley, M.P., Drevets, W.C., Öngür, D., Price, J.L., 2002. Glial changes in the amygdala and entorhinal cortex in mood...
  • J.D. Bremner et al.

    Positron emission tomography measurement of cerebral metabolic correlates of tryptophan depletion-induced depressive relapse

    Arch. Gen. Psychiatry

    (1997)
  • A.L. Brody et al.

    Regional brain metabolic changes in patients major depressive disorder from pre- to post-treatment with paroxetine

    Arch. Gen. Psychiatry

    (2001)
  • L. Brothers

    Neurophysiology of the perception of intentions by primates

  • S.T. Carmichael et al.

    Limbic connections of the orbital and medial prefrontal cortex in Macaque Monkeys

    J. Comp. Neurol.

    (1995)
  • Charney, D.S., Drevets, W.C., 2002. The neurobiological basis of anxiety disorders. In: Davis, K., Charney, D.S.,...
  • J. Corsellis et al.

    Neuropathological observations on yttrium implants and on undercutting in the orbito-frontal areas of the brain

  • Casey, B.J., Thomas, K.M., Eccard, C.H., Drevets, W.C., Dahl, R.E., Whalen, P.J., Perrett, D.J., Ryan, N.D., 2002....
  • M.I. Chimowitz et al.

    Further observations on the pathology of subcortical lesions identified on magnetic resonance imaging

    Arch. Neurol.

    (1992)
  • R.M. Cohen et al.

    Preliminary data on the metabolic brain pattern of patients with winter seasonal affective disorder

    Arch. Gen. Psychiatry

    (1992)
  • A.R. Damasio et al.

    Neural correlates of the experience of emotions

    Soc. Neurosci. Abstr.

    (1998)
  • M. Davis

    The role of the amygdala in conditioned fear

  • D. Dioro et al.

    The role of the medial prefrontal cortex (cingulate gyrus) in the regulation of hypothalamic–pituitary–adrenal responses to stress

    J. Neurosci.

    (1993)
  • W.C. Drevets

    PET and the functional anatomy of major depression

  • W.C. Drevets

    Prefrontal cortical-amygdalar metabolism in major depression

  • W.C. Drevets et al.

    Neuroimaging in psychiatry

  • W.C. Drevets et al.

    Blood flow changes in human somatosensory cortex during anticipated stimulation

    Nature

    (1995)
  • Cited by (498)

    View all citing articles on Scopus
    View full text