Alan R. Prossin, MD; and Melvin G. McInnis, MD, are with the Department of Psychiatry, University of Michigan Medical School, Ann Arbor, Michigan. Amit Anand, MD, is with the Department of Psychiatry and the Department of Radiology, Indiana University School of Medicine, Indianapolis. Mary M. Heitzeg, MD, is with the Department of Psychiatry, and the Molecular and Behavioral Neuroscience Institute, University of Michigan Medical School. Jon-Kar Zubieta, MD, is with the Department of Psychiatry, Molecular and Behavioral Neuroscience Institute, and Department of Radiology University of Michigan Medical School.
Funding from The Phil F. Jenkins Foundation and The Prechter Bipolar Research Fund at the University of Michigan supported this work.
Dr. Prossin; Dr. McInnis; Dr. Anand; Dr. Heitzeg; and Dr. Zubieta have disclosed no relevant financial relationships.
Address correspondence to: Alan R. Prossin, MD, Comprehensive Depression Center, Department of Psychiatry, University of Michigan Medical School, 4250 Plymouth Road, Suite# 2409, Ann Arbor, MI 48105; fax 734-936-2690; or e-mail email@example.com.
The description of schizophrenia and bipolar disorder as disparate clinical entities arose from a Kraepelinian distinction in the time period from the late 19th to early 20th century, which soon became known as the Kraepelinian dichotomy. Over the years, many studies were conducted to gain further insight as to whether or not this distinction continued to hold true. The increase in interest in the study and effect of the phenomenology associated with psychiatric illness in the 1980s led to the Kraepelinian dichotomy being operationalized in the third revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III).1 The Kraepelinian dichotomy continues to be ingrained in psychiatric practice. However, the age-old debate as to whether or not the dichotomy “holds true” persists. This article reviews the existent neuroimaging literature on this topic in an effort to illustrate the neurobiological evidence that either supports or refutes this dichotomy. The article concludes with related recommendations as to suggested revisions to the current categorical nosology of the DSM-IV-TR.
Despite Kraepelin’s initial conceptualization, he later acknowledged that schizophrenia and bipolar disorder may, indeed, share some overlapping clinical features. Kraepelin began to appreciate that it was not individual symptoms that would distinguish schizophrenia and bipolar disorder but rather the pattern of their presentation that would provide these clinical delineations. Although Kraepelin believed schizophrenia to be associated with a gradual and progressive cognitive and functional decline, he believed manic depression to be more representative of a course of intermittent decline in functioning interwoven with periods of return to baseline and full recovery of illness. Kraepelin found, through his studies of families, that indeed what he termed dementia praecox (Bleuler later renamed this as schizophrenia) and manic depression (later differentiated into various mood disorders, including bipolar illness and depression)2,3 appeared to be inherited traits based on the similarities in course and outcome among generations within families.
Some clinicians of the era debated the idea of a mutually exclusive or dichotomous characterization of these debilitating illnesses. In a more integrative fashion, Kasanin introduced the term schizoaffective disorder with its inherent overlap between mood and psychotic symptoms. Some clinicians and researchers continue to argue that given the significant clinical overlap, a certain degree of shared biological (and likely associated genetic) predisposition is likely present. Indeed, Kraepelin’s discussions on the ability of these disorders to “breed true” emphasized the inherited nature of these syndromes, providing evidence of his foresight of the need to identify biological features of these illnesses to gain a mechanistic understanding of their similarities and differentiation.
Despite some early questions and debate as to the degree of overlap between these clinical illnesses, the increase in genetic and family studies over the past 2 to 3 decades has resulted in the reignition of this age-old debate. The extent of clinical overlap between schizophrenia and bipolar disorder that were identified in research studies support the overlap that was readily apparent from clinical practice. Some of the factors involved in this overlap include epidemiological factors (ie, age of onset, sex distributions, and prevalence) and clinical presentations (ie, schizophrenia with superimposed mood symptoms and psychotic presentations of bipolar disorder). Until recently, however, there has been limited research involving neurobiological phenotypes of this overlap.
We set out to review the literature that explores this issue by reviewing studies that specifically compared and contrasted the various imaging findings between schizophrenia and bipolar disorder. Although the bulk of the neuroimaging literature comparing and contrasting the findings between these two clinical paradigms relates to volume comparisons derived from anatomical studies (ie, postmortem anatomy, structural magnetic resonance imaging [MRI]), the evolution of functional neuroimaging research paradigms has resulted in a slow, but gradual, increase in the numbers and quality of studies that compare and contrast the brain function between these two debilitating illnesses.
An extensive literature review of original research articles comparing neuroimaging findings in schizophrenia and bipolar disorder was conducted using PubMed. In all articles reviewed here, diagnoses were to be made based on the DSM version that was current at the time of article submission. A postmortem anatomical study was also included to represent earlier work and as a reference to compare more recent neuroimaging findings against. The literature search was further narrowed by including only those original research articles whose paradigms were based on the following neuroimaging techniques: structural MRI, diffusion tensor imaging (DTI), magnetic resonance spectroscopy (MRS), positron emission tomography (PET), and functional MRI (fMRI). Of the approximately 80 original research articles obtained from our intensive search (“Brain Imaging Bipolar Schizophrenia;” MRI Bipolar Schizophrenia;” PET Bipolar Schizophrenia;” “fMRI Bipolar Schizophrenia;” “Functional Bipolar Schizophrenia”), the results were narrowed down to approximately 20 articles based on methodological soundness and quality of research. The results were further organized categorically based on imaging modality and then further subdivided based on regional findings. These findings are outlined in the Table (see page 156) under the following categories: postmortem, MRI, MRS, DTI, PET, and fMRI. This was not intended to be a thorough review of neuroimaging for each individual disorder but rather an outline of findings from studies that specifically compared schizophrenia and bipolar disorder and which possibly had related these findings to those of healthy controls.
Table. A Review of Neuroimaging Studies Comparing Schizophrenia with Bipolar Disorder
Additionally, to further investigate whether the Kraeplinian dichotomy and the associated neuroimagong findings hold true, some studies focusing on genetic studies have been included. Where possible, comparisons of neuronal circuitry from a systems-based approach to neuroimaging has been highlighted. However, despite these efforts, the bulk of the studies comparing these two illnesses involved anatomical comparisons from structural MRI.
Most articles identified involved comparing and contrasting the brain structure in individuals with schizophrenia and bipolar disorder with those of healthy controls (see Table, page 156). Although there has been some work done in PET imaging (see Table, page 156) investigating the differences in the various receptor systems, specifically across these diagnostic boundaries, this volume of this research is not great, perhaps because of the ethical concerns of withholding treatment from this clinically unstable group of individuals.
A postmortem study4 investigating diagnostic comparisons of neuronal somal size within the layers of the insular cortex has determined that schizophrenic subjects have a reduction in neuronal somal size within a particular layer of the insula as compared with bipolar subjects.
The structural MRI studies reviewed (see Table, page 156) show consistent evidence that in schizophrenia, there is a reduction in hippocampal volumes and an increase in ventricular size compared with patients with bipolar disorder. However, there is less consistency regarding frontal and subcortical changes in terms of replication of findings, presence or absence of lateralization, and clinical association.
In an MRS investigation5 (see Table, page 156), although no difference in metabolites between diagnostic categories was detected, bipolar subjects on lithium treatment evidenced elevated NAA/PCr-Cr in the basal ganglia region as compared with healthy controls, an effect that could be confounded by the effects of medication.
Of the fMRI studies reviewed (see Table, page 156), a common theme emerged indicating activation differences in the prefrontal cortex between subjects with bipolar disorder and subjects with schizophrenia. These differences do not appear to be limited to psychotic presentations of mood episodes and include subjects with bipolar disorder type I, in general. Of particular interest is that these differences occurred across a variety of tasks, including emotional prosody, sentence completion, and verbal fluency. It should be noted that one study identified a link between the effect of genotype for neuroreglin-1 and activation under a verbal fluency task, with effect of genotype being associated with greater deactivation in the left precuneus across diagnostic boundaries (schizophrenia, bipolar disorder, and healthy controls). This study additionally identified an association between effect of genotype and activation that appeared to separate out with clinical categories (inferior frontal gyrus in subjects with high risk for schizophrenia; posterior orbital gyrus in subjects with high risk for bipolar disorder).
We identified and selected two PET studies that specifically compared the neurobiology of schizophrenia and bipolar disorder I (see Table, page 156). One of these studies directly measured particular receptor or transporter binding potentials, and the second study used PET to measure cerebral blood flow. Zubieta et al found differences in the concentration of vesicular monoamine transporters (VMAT2), as measured with the radiotracer [11C]DTBZ, reflecting aminergic synaptic density, between schizophrenic and bipolar subjects in the thalamus, as well as between bipolar disorder and healthy subjects.6,7 In another study that investigated the diagnostic differences in blood flow in the superior temporal cortex, no effects of diagnosis were found.8
There has been little work done to date examining the differences in measures of white matter tracks and neuronal connectivity between schizophrenia and bipolar disorder. However, one recent study investigating connectivity with diffusion tensor imaging found that although patient groups and controls had differences in fractional anisotropy compared with healthy controls, these differences did not hold for comparisons across the diagnostic categories of schizophrenia and bipolar disorder9 (see Table, page 156).
The neurobiological/neuropsychological domains affected in bipolar disorder have been identified as being more widespread and more closely related to schizophrenia than originally expected. However, the degree of severity of these changes is disparate, with schizophrenia resulting in a typically more severe neuropsychological deterioration than that of bipolar disorder. The convergence that is seen within endophenotypes (ie, neuroimaging data) associated with certain phenotypic domains (ie, psychosis) suggests the presence of a dimensional pattern of illness. However, the existence of overlap with schizophrenia appears more consistent in psychotic forms of bipolar disorder as compared with non-psychotic bipolar disorder. Although on comparing respective measures of connectivity, fractional anisotropy was, indeed, found to be similar across diagnostic boundaries in certain brain regions (ie, anterior limb of internal capsule, anterior thalamic radiation, and uncinate fasciculus) and with FA differences noted between either patient group and healthy controls.
Attempts to separate these psychiatric pathologies by spectroscopic measures have not proven beneficial. Specifically, evidence of changes in the concentrations of various cellular markers (ie, choline, inositol, and NAA/PCr-Cr) in the basal ganglia has been observed only in bipolar patients on lithium treatment, whereas concentrations in both the basal ganglia and occipital cortex were found to be similar between schizophrenia and bipolar disorder. In addition, these similarities overlapped with those of healthy controls, suggesting that this measure may not be helpful in differentiating these psychopathologies from controls.
Additionally, other associations (ie, left amygdala volume associated with verbal memory) highlight the differences between these illnesses and, in so doing, provide more support to the categorical nosology present in the current version of the DSM. Indeed, just as many nonoverlapping find-ings are evident from our review of neuroimaging in schizophrenia and bipolar disorder. Most notable are the volumetric differences in the ventricles, hippocampus, and amygdala between schizophrenia and bipolar disorder that have been widely reproduced. However, if these changes were to be represented on a continuum, we would not expect such disparate association with neuropsychological functioning. This has been pointed out by Kilgore et al, who were unable to find differences in amygdala volume but did find that left amygdala volume in bipolars was positively correlated to verbal memory. However, in schizophrenia, it was negatively correlated with memory.10
The PET findings illustrate the idea of partial overlap and coexisting heterogeneity of these two psychiatric illnesses, with diagnostic effects being associated with vesicular monoamine transporter (VMAT2) binding in both the thalamus (bipolar disorder having greater binding than schizophrenia or healthy controls) and the ventral brainstem (similar binding in bipolar disorder and schizophrenia, both of which are greater than that in healthy controls). Other research showed no diagnostic difference in regional cerebral blood flow measurements.
The results of the fMRI studies yield similar results, some supporting the dichotomy and others supporting a dimensional explanation. There is evidence to show that bipolar and schizophrenic subjects differentially activated portions of their insula and prefrontal cortex during sentence completion and differential activation of the prefrontal cortex during verbal fluency tasks. Additionally, tasks of emotional prosody appeared to highlight both differential activation and similarities in activation between schizophrenia and bipolar disorder. Despite the fact that recent studies have suggested neuroregulin-1 as a susceptibility gene for schizophrenia and bipolar disorder, the results outlined have shown regions of deactivation associated with the neuroregulin-1 gene that appear to be common to bipolar disorder, schizophrenia, and healthy controls (left precuneus). Additionally, an interaction between genotype and regional activation in the prefrontal cortex is present in both bipolar disorder and schizophrenia. However, these regions of activation are disparate, outlining the possibility that despite a possible genetic similarity across these diagnostic paradigms, the functional implication of this genetic susceptibility appears to be somewhat disparate.
In summary, the available neuroimaging findings that have been reviewed support the presence of both an overlap and a distinction between the circuitry that appears to be involved in the pathophysiology of schizophrenia and bipolar disorder. Currently, most available data are limited to neuroanatomical measures of both gray matter and white matter, with a recent trend toward an increase in functional studies. The findings of anatomical and functional overlap would be helpful in informing genetic association studies in this area. Additionally, further longitudinal neuroimaging incorporated into family studies would likely provide more evidence to support these hypotheses.
The above findings neither refute nor prove the Kraepelinian dichotomy but certainly provide evidence for an overlap between neurobiological factors underlying both disorders. The results of this review validate the need for the further investigation of biomarkers that will assist in the subclassification of those psychiatric illnesses whose symptom clusters lie along the spectrum of mood and psychotic dimensions. These discoveries will likely highlight the complexity of disordered neurobiology, yielding a greater prominence to biologically driven diagnostic and treatment approaches in chronic neuropsychiatric illnesses. The key question that emerges is not surrounding the validity of the Kraepelinian dichotomy but rather is how we move forward in our transition from descriptive clinical classifications toward a diagnostic system that is aligned with the complex neurobiological underpinnings of these disorders — and more akin to clusters of biological disparities and similarities.
- Winokur G. Psychosis in bipolar and unipolar affective illness with special reference to schizo-affective disorder. Br J Psychiatry. 2984;145:236–242.
- Lake CR, Hurwitz N. The schizophrenias, the neuroses and the covered wagon; a critical review. Neuropsychiatr Dis Treat. 2007;3(1):133–143. doi:10.2147/nedt.2007.3.1.133 [CrossRef]
- Lake CR, Hurwitz N. Schizoaffective disorder merges schizophrenia and bipolar disorders as one disease — there is no schizoaffective disorder. Curr Opin Psychiatry. 2007;20(4):365–379. doi:10.1097/YCO.0b013e3281a305ab [CrossRef]
- Pennington K, Dicker P, Hudson L, Cotter DR. Evidence for reduced neuronal somal size within the insular cortex in schizophrenia, but not in affective disorders. Schizophr Res. 2008;106(2–3):164–171. doi:10.1016/j.schres.2008.08.022 [CrossRef]
- Sharma R, Venkatasubramanian PN, Barany M, Davis JM. Proton magnetic resonance spectroscopy of the brain in schizophrenic and affective patients. Schizophr Res. 1992;8(1):43–49. doi:10.1016/0920-9964(92)90059-E [CrossRef]
- Zubieta JK, Taylor SF, Huguelet P, Koeppe RA, Kilbourn MR, Frey KA. Vesicular monoamine transporter concentrations in bipolar disorder type I, schizophrenia, and healthy subjects. Biol Psychiatry. 2001;49(2):110–116. doi:10.1016/S0006-3223(00)00981-1 [CrossRef]
- Zubieta JK, Huguelet P, Ohl LE, et al. High vesicular monoamine transporter binding in asymptomatic bipolar I disorder: sex differences and cognitive correlates. Am J Psychiatry. 2000;157(10):1619–1628. doi:10.1176/appi.ajp.157.10.1619 [CrossRef]
- Dye SM, Spence SA, Bench CJ, et al. No evidence for left superior temporal dysfunction in asymptomatic schizophrenia and bipolar disorder. PET study of verbal fluency. Br J Psychiatry. 1999;175:367–374. doi:10.1192/bjp.175.4.367 [CrossRef]
- Sussmann JE, Lymer GK, McKirdy J, et al. White matter abnormalities in bipolar disorder and schizophrenia detected using diffusion tensor magnetic resonance imaging. Bipolar Disord. 2009;11(1):11–18. doi:10.1111/j.1399-5618.2008.00646.x [CrossRef]
- Killgore WD, Rosso IM, Gruber SA, Yurgelun-Todd DA. Amygdala volume and verbal memory performance in schizophrenia and bipolar disorder. Cogn Behav Neurol. 2009;22(1):28–37. doi:10.1097/WNN.0b013e318192cc67 [CrossRef]
A Review of Neuroimaging Studies Comparing Schizophrenia with Bipolar Disorder
|Pennington K, et al||2008||Evidence for reduced neuronal somal size within the insular cortex in schizophrenia, but not in affective disorders||Schizophr Res.||Postmortem||Insular cortex||Cytoarchitecture of insula is abnormal in schizophrenia but not in mood disorders.|
|Dasari M, et al||1999||A magnetic resonance imaging study of thalamic area in adolescent patients with either schizophrenia or bipolar disorder as compared to healthy controls||Psychiatry Res.||MRI||Thalamus||Thalamic area in bipolars is similar to that of schizophrenics. Both patient groups have less thalamic area than controls.|
|McDonald C, et al||2005||Regional volume deviations of brain structure in schizophrenia and psychotic bipolar disorder: computational morphometry study||Br J Psychiatry.||MRI||Frontotemporal neocortex, medial temporal lobe, insula, thalamus, and cerebellum; major longitudinal and interhemispheric tract||Grey matter deficits (frontotemporal neocortex, medial temporal lobe, insula, thalamus, and cerebellum) are greater than in bipolars or controls. Major longitudinal and interhemispheric tract deficits in shizophrenics and bipolars are similar and greater than controls.|
|McDonald C, et al||2004||Association of genetic risks for schizophrenia and bipolar disorder with specific and generic brain structural endophenotypes||Arch Gen Psychiatry.||MRI||Frontostriato-thalamic, temporal, anterior cingulate gyrus, ventral striatum||Genetic risk for schizophrenias assoc. w/distributed gray deficits (bilateral frontostriatothalamic, left lateral temporal), genetic risk for bipolar assoc. w/gray deficits (right anterior cingulate gyrus and ventral striatum); left frontotemporal disconnectivity as a genetically controlled brain structural abnormality common to both disorders.|
|McIntosh AM, et al||2004||Voxel-based morphometry of patients with schizophrenia or bipolar disorder and their unaffected relatives||Biol Psychiatry.||MRI||Anterior thalamus (AT); middle prefrontal gyrus (mPFG); dorsal medial thalamus (DMT)||Anterior thalamus is similar in schizophrenics, bipolars, and their relatives but smaller than controls. Schizophrenics have smaller mPFG and DMT than bipolars and controls.|
|McIntosh AM, et al||2008||White matter tractography in bipolar disorder and schizophrenia||Biol Psychiatry.||MRI||Uncinate fasciculus, anterior thalamic radiation||Uncinate fasciculus and anterior thalamic radiation is similar in schizophrenics and bipolars, both less than controls.|
|McIntosh AM, et al||2009||Prefrontal gyral folding and its cognitive correlates in bipolar disorder and schizophrenia||Acta Psychiatr Scand.||MRI||Prefrontal gyrification||Prefrontal gyrification is greater in controls than in either bipolars or schizophrnics (both with similar results).|
|McIntosh AM, et al||2006||Genetic liability to schizophrenia or bipolar disorder and its relationship to brain structure||Am J Med Genet B Neuropsychiatr Genet.||MRI||Prefrontal cortex (dorsolateral and ventrolateral)||Genetic risk associated with decreased volume in schizophrenia but not in bipolar disorder.|
|Velakoulis D, et al||2006||Hippocampal and amygdala volumes according to psychosis stage and diagnosis: a magnetic resonance imaging study of chronic schizophrenia, first-episode psychosis, and ultra-high-risk individuals||Arch Gen Psychiatry.||MRI||Hippocampus||Smaller bilaterally in schizophrenics than in bipolars or controls.|
|Connor SE, et al||2004||A study of hippocampal shape anomaly in schizophrenia and in families multiply affected by schizophrenia or bipolar disorder||Neuroradiology.||MRI||Hippocampal shape anomaly (HAS)||Present in schizophrenics, but not in their relatives, bipolars, or controls.|
|Strasser HC, et al||2005||Hippocampal and ventricular volumes in psychotic and non-psychotic bipolar patients compared with schizophrenia patients and community control subjects: a pilot study||Biol Psychiatry.||MRI||Hippocampus, ventricles||Ventricle volume greater in psychotic bipolars than in non-psychotic bipolars or controls. Hippocampal volume less in schizophrenics than in bipolars or controls.|
|McDonald C, et al||2006||Regional brain morphometry in patients with schizophrenia or bipolar disorder and their unaffected relatives||Am J Psychiatry.||MRI||Ventricles, hippocampus||Lateral and 3rd ventricular volumes greater in schizophrenics (and relatives) than in bipolars or controls.|
|Sharma R, et al||1992||Proton magnetic resonance spectroscopy of the brain in schizophrenic and affective patients||Schizophr Res.||MRS||Basal ganglia||Bipolar patients had greater elevated NAA/PCr-Cr, choline/PCr-Cr, and inositol/PCr-Cr ratios than schizophrenics.|
|Zubieta JK, et al||2001||Vesicular monoamine transporter concentrations in bipolar disorder type I, schizophrenia, and healthy subjects||Biol Psychiatry.||PET [11C] DTBZ||VMAT2 binding in thalamus, ventral midbrain||Bipolars had greater VMAT2 binding in the thalamus but similar binding in the ventral brainstem as compared with schizophrenics.|
|Dye SM, et al||1999||No evidence for left superior temporal dysfunction in asymptomatic schizophrenia and bipolar disorder. PET study of verbal fluency||Br J Psychiatry.||PET (rCBF)||Temporal||rCBF is reduced in the STC in all groups.|
|Mechelli A, et al||2008||The effects of neuregulin1 on brain function in controls and patients with schizophrenia and bipolar disorder||Neuroimage.||Fmri, genetic||Relation of PFC activation to neuroregulin1||Interaction found (rIFG, high risk, schizophrenia) (rPOG, high risk, bipolar disorder).|
|McIntosh AM, et al||2008||Prefrontal function and activation in bipolar disorder and schizophrenia||Am J Psychiatry.||fMRI||Insula and dorsal prefrontal cortex activation||Difference in OFC and ventral striatum recruitment in bipolars and schizophrenics.|
|Curtis VA, et al||2001||Differential frontal activation in schizophrenia and bipolar illness during verbal fluency||J Affect Disord.||fMRI||Frontal lobe||Bipolars have different frontal response patterns (during verbal fluency tasks) than schizophrenics.|
|McIntosh AM, et al||2008||Prefrontal function and activation in bipolar disorder and schizophrenia||Am J Psychiatry.||fMRI||Insular cortex, dorsal prefrontal cortex||Activation during Hayling task differentiated bipolars from schizophrenics and controls|
|Mitchell RL, et al||2004||Neural response to emotional prosody in schizophrenia and in bipolar affective disorder||Br J Psychiatry.||fMRI||Temporal lobe||Bipolars and schizophrenics show left-lateralization of temporal lobe response to emotional prosody|
|Sussmann JE, et al||2009||White matter abnormalities in bipolar disorder and schizophrenia detected using diffusion tensor magnetic resonance imaging||Bipolar Disord.||DTI||Anterior limb of internal capsule (ALIC), anterior thalamic radiation (ATR), uncinate fasciculus (UF)||Reduced integrity of the ALIC, ATR, UF regions in both bipolars and schizophrenics as compared with controls.|
|Fornito A, et al||2009||Reconciling neuroimaging and neuropathological findings in schizophrenia and bipolar disorder||Curr Opin Psychiatry.||MRI review.||Results point to the gross complexity, regional heterogeneity present in both bipolar disorder and schizophrenia.|