I thought it might be of interest to summarize recent neuroimaging research in the schizophrenias. These findings are reported in the scholarly volume, Schizophrenia: From Neuroimaging to Neuroscience by Stephen Lawrie, Eve Johnstone, and Daniel Weinberger published in 2004 for Oxford University Press.

The material for this summary of the research is taken from a chapter in this volume entitled “Structural magnetic resonance imaging” by Andrew McIntosh and Stephen Lawrie. This chapter is a good source of understanding the methodological considerations and types of image analysis used in current sMRI (structural as opposed to functional) research.

Differences between patients diagnosed with schizophrenia and controls

Several systematic reviews consisting of more than 100 controlled studies exist describing findings differentiating persons with schizophrenia and controls. The review by Shenton et al (2001) is the most extensive.

Whole brain (2% reduction in white matter-4% reduction in grey matter)

Whole brain volume is reduced compared to controls only when studies are combined. Most individual studies fail to find significant differences. The effect size for grey and white matter correspond to a volume reduction of 4% and 2%, respectively. The majority of first-episode studies are similarly equivocal.

Ventricles and CSF (20-30% enlargement in lateral & third ventricles)

In summary, persons with schizophrenia have 20-30% larger lateral and third ventricles than controls. However, as I pointed out in many previous papers, ventriculomegaly is demonstrated to be non-specific in schizophrenia. It is observed in bipolar patients, affective disorders, PTSD as well as in the normal aging brain. Ventriculomegaly is usually attributed to atrophy of surrounding neural tissue, particularly in temporal regions (e.g., hippocampus, parahippocampus, amygdala).

Prefrontal cortex (2-8% reduction)

Most studies have observed lower prefrontal volumes in persons with schizophrenia, regardless of whether chronic or acute presentations. Overall, the reductions appear to be from 2 to 8% (the prefrontal cortex comprises roughly 30% of total cerebral volume)

Temporal Lobe (2-4% reduction)

Volume reductions in temporal regions are well described in the research literature. Of 51 sMRI studies of whole temporal lobe volumes in schizophrenia, 61% showed reductions compared to controls (Shenton et al 2001). Wright et al (2000) pooled the data from 25 studies and found a reduction of between 2 and 4% in patients with schizophrenia. Temporal reductions are less consistently found in first-episode patients and there is some evidence that temporal lobe volume reductions may be progressive over time. Superior temporal lobe (STG) and fusiform gyri have been observed to be reduced, the former associated with language and the latter with facial recognition [perhaps both reduced for defensive purposes].

Amygdala (10% reduction)

Large volume reductions have been observed in the left and right amygdalae in persons with schizophrenia of about 10%.

Hippocampus and parahippocampus (6% reduction)

More than 30 well controlled studies demonstrated volume reductions of about 6% in both left and right hippocampi in persons with schizophrenia compared to controls (the reductions appear at first presentation), with a similar pattern in the left and right parahippocampus.

Thalamus (reductions)

Recent evidence points to volume reductions in the thalamus. It is difficult to separate the whole structure, because the thalamus is comprised of several nuclei.

Basal ganglia and nucleus accumbens (20% increase in globus pallidus)

In contrast to other nuclei, persons with schizophrenia, demonstrate larger left and right globus pallidus than controls. Several studies have found increases of about 20% overall, including some studies of first episode patients (prior to neuroleptic exposure). Antipsychotic agents are associated with larger caudate volume. The studies remain inconsistent whether the basal ganglia are larger, smaller or similar in size prior to neuroleptic exposure. There is no convincing evidence that the nucleus accumbens differs in size compared to controls.

Insula (grey matter reductions)

There is some evidence for grey matter reductions in the insula (a structure buried deep within the lateral sulcus, concealed by the frontal, temporal and parietal lobes) in persons with schizophrenia.

McIntosh and Lawrie (2004) concluded:

“Although a neuroanatomical profile of schizophrenia is beginning to emerge, the precise nature of the findings requires clarification...Many twin and family studies assume that equivalence in a regional volume represents a genetic influence, but rarely are the studies sufficiently sized to make this a strong conclusion...On the other hand, given that symptoms and cognitive function are likely to be mediated by networks rather than circumscribed regions of neurons, and potentially affected by pathoplastic effects (e.g., ‘personality,’ life events, and other historical factors), functional imaging [fMRI] is likely to prove more fruitful. The timing of many findings and the possibility of their progression before and after illness onset is the most controversial issue in sMRI at the present time” (p.46).

As I noted in my volume in progress, The Schizophrenias: Brain, Mind and Culture: ”A comprehensive model of schizophrenia must address the following factors in a coherent manner: neurogenesis, symptom formation, etiology, onset, and heterogeneity ( Heinrichs 2001). To this list I would add variable courses and outcomes.” I believe that these factors are most logically and comprehensively explained by the following processes. The sMRI findings are perhaps more associated with the neural effects of fear/anxiety than the symptoms of schizophrenia, e.g., hallucinations and delusions. I have amply demonstrated this in my overview of the neuroscience research in stress outlined in my book in progress noted above. The symptoms, rather than being neurotoxic in themselves, may be the individual’s attempts to deal with the vacuum of personal existence (what Gaetano Benedetti referred to as “negative existence”). The overlap between the neuroscience research in schizophrenia, as well as in bipolar disorder, and the neuroscience research in stress/fear is so significant that it is quite puzzling to me as to why more reductionistically-oriented researchers do not attempt to incorporate this into their theories of the etiopathogenesis of the schizophrenias (as well as bipolar disorder).

Brian Koehler
New York University