Structural imaging of schizophrenia: is the modern psychiatrist becoming a ‘disease scene investigator?’

Perlini et al. (1) provide an overview of how brain images can be read, explaining how structural abnormalities are being investigated in schizophrenia. This approach – along with modern genomics – promises to allow psychiatrists to become disease scene investigators analogous to crime scene investigators celebrated by the popular media. That is, practicing clinicians may soon be investigating the scene of the disease, namely the brain, with these structural imaging techniques (1), combining them with functional imaging techniques (2), while simultaneously assessing DNA from the individual patient found at the disease scene as well, sending that DNA for analysis in a modern library of genotypes linked to diagnoses and treatments (3, 4). Not the kit only of science fiction any longer, this can now be done in living patients and is rapidly extending from the research laboratory to the clinic. We clinicians should now acquaint ourselves with these techniques (1–4) so we will be ready to apply them, perhaps coming to think of ourselves in part as CNS-I s (central nervous system investigators) analogous to contemporary CSIs (crime scene investigators). How does this work? Perlini et al. (1) help us understand how to interpret existing findings of structural abnormalities in schizophrenia reported with one or another of the three basic methods: ROI (region of interest), VMB (voxel-based morphometry) and computational neuroscience. It is not important for clinicians to be able to read the brain scans themselves or to analyze personally the data from these techniques in a given patient. It is, however, important for clinicians to know what the results of these techniques mean when receiving a neuroimaging report on one of their patients. For example, ROI results are similar to how a clinical neuroradiologist reads brain scans, both qualitatively according to the individual s unique anatomy and quantitatively according to objective criteria. However, this technique is slow and subject to the reader s bias. VBM, on the other hand, is objective, automated and thus fast, but a bit dumb, leaving out judgment and blind to the individual s anatomy. Finally, there is computational neuroscience, which is potentially the most powerful but least intuitive and beyond the scope of humble clinicians. When applied to schizophrenia, what have these techniques told us? We have known for a long time that gross anatomical abnormalities are not consistently reported in schizophrenia and that even the three current techniques discussed here cannot image the single neuron or synapse where the genetically mediated structural brain abnormalities are thought to exist (3). However, these techniques have shown that certain brain areas – but not all brain areas – are altered with either increased or decreased gray matter. These changes have generated hypotheses but not individually reportable structural findings about how to diagnose schizophrenia at the earliest stages of the illness as well as how to monitor progression of anatomical changes over time caused by disease progression, antipsychotic medications or both. We are still, however, a long way from getting brain scans of patients to diagnose their schizophrenia or to monitor their course illness over time. In fact, it is beginning to look like no structural, functional or genetic test will ever diagnose schizophrenia or tell us what drug to use or avoid. It is not likely to be that simple. More likely, it will be a matter of equipoise or how we balance the evidence. Thus, the practice of psychiatry is not in immediate danger of losing its requirement for a unique combination of skills that balance the art with the science (equipoise). What seems very possible to emerge from all of this is for the modern psychiatrist to become a type of sleuth, namely what we could call a CNS-I or disease scene investigator, based on the notion that with more science, combined with continuing the art of Acta Psychiatr Scand 2012: 126: 231–232 All rights reserved DOI: 10.1111/j.1600-0447.2012.01904.x 2012 John Wiley & Sons A/S