ABSTRACT

As with all messages from the grave, we have to be sceptical when discussing the neuropathology of bipolar disorder. So far, there have been some intriguing whispers, but they remain difficult to hear, and we still cannot be sure how best to decipher them. In this chapter, I cover the main findings from postmortem studies of bipolar disorder, and try to link them with some other themes. I have not attempted to provide a systematic overview of the subject; for this see two recent reviews.1,2

At the start it is important to point out that neuropathological studies of bipolar disorder only began a few years ago. It is no exaggeration to say there were no data worth discussing until a study was published in 1998.3

The reason why the field is so recent is largely a practical one: people simply did not collect enough brains from patients who had suffered from bipolar disorder, with adequate characterization and appropriate controls,4 to carry out any kind of quantitative, or molecular, studies. It is thanks to the Stanley Foundation (now the Stanley Medical Research Institute) that the situation changed in the mid-1990s, because of their funding of an autopsy series of brains from patients (15 in each group) with bipolar disorder as well as schizophrenia, major depression and control subjects.5 This series was made available to investigators across the world. Part of the deal was that one had to study all 60 subjects in all experiments (the material was coded), and so people began to study bipolar disorder. A significant proportion of the world literature on the neuropathology of bipolar disorder now comes from this series. The gene expression component of this work has been meta-analysed on a region-by-region basis.6,7

A few other practical issues are also worth noting. First, the concepts, methods, targets and interpretation of the research have been much influenced by the experience with schizophrenia which had emerged over the preceding 10-15 years, and by the fact that it is the latter disorder that is often of primary interest. The majority of papers therefore discuss bipolar disorder findings in terms of similarities (of which there are many) and differences (of which there are relatively few) with those of schizophrenia. At the same time, however, comparison with other mood disorders cannot be neglected, and this is emphasized in some publications. Thus, for example, some papers concentrate on bipolar versus unipolar mood disorder, or emphasize the importance of familiality not polarity.3 These different ways of cutting the cake (or comparing two cakes) makes the literature somewhat confusing. The second point is that neuropathologists have to pick a particular part of the brain to study: the brain is too big (and too structurally complex even within a single area) to examine the whole thing. Because one has to start somewhere, research has naturally focused on where researchers think the positive findings are most likely to be. (As the man said when asked why he robbed banks: ‘they’re where the money is’.)

The most relevant clues have been those from brain imaging, and hence many of the neuropathological data are in the anterior cingulate cortex and orbitofrontal cortex, because of the many data implicating these regions (Figure 7.1). The best example of this kind is the combined study of magnetic resonance imaging (MRI) and positron emission tomography (PET) showing smaller volumes and reduced metabolic activity in mood disorder in the ‘subgenual’ part of the cingulate cortex,8 which led the same group to carry out the landmark postmortem study mentioned.3 Whilst this ‘candidate region’ approach may have been inevitable, it does of course lead to a circularity in the argument – it means we do not know where the pathology of bipolar disorder might be centred because people have not looked, to anything like the same extent, in many other areas of the brain – compare the cerebellum in schizophrenia: hardly an area which many people implicated, nor were interested in, a decade ago, yet where there is now good evidence for structural and functional involvement.