One of the oldest scientific questions about schizophrenia is whether the brain looks different — and how. Beginning with crude pneumoencephalography in the 1920s and continuing through CT in the 1970s and MRI from the 1980s onward, researchers have built a remarkably consistent body of evidence on structural brain differences in schizophrenia. This article summarises what large-scale studies, particularly the ENIGMA consortium, have shown.
Group-level structural MRI studies in schizophrenia consistently show modest reductions in total brain and cortical volume, hippocampal volume, and several specific cortical regions, alongside modest enlargement of the ventricles — but the differences are small relative to within-group variation.
The classical findings
The earliest reliable finding, replicated since the 1970s, was enlargement of the lateral ventricles in schizophrenia. Ventricular enlargement reflects loss of surrounding brain tissue and was visible even on early CT scans. Decades of MRI have confirmed and extended this finding. By the 2000s, structural MRI had also documented:
- Modestly reduced total brain volume
- Reduced cortical grey matter, especially in temporal and frontal regions
- Reduced hippocampal volume
- Reduced thalamic volume
- Subtle white matter abnormalities, especially in tracts connecting frontal and temporal regions
The ENIGMA consortium
Most of what we now know with confidence comes from very large pooled studies. The Enhancing Neuro Imaging Genetics through Meta-Analysis (ENIGMA) consortium has aggregated structural MRI data from tens of thousands of participants worldwide, reporting standardised analyses with vastly more statistical power than any single study.
Key ENIGMA schizophrenia findings include:
- The 2016 paper by van Erp and colleagues in Molecular Psychiatry analysed roughly 2,000 patients and 2,500 controls and confirmed reduced hippocampus, amygdala, thalamus, and accumbens, with enlarged pallidum and lateral ventricles.
- The 2018 ENIGMA cortical paper in Biological Psychiatry showed widespread reductions in cortical thickness and surface area, with the largest cortical thinning in frontal and temporal regions.
- White matter studies have shown decreased fractional anisotropy in major tracts, reflecting altered microstructure.
How big are the differences?
Most reported group differences are small — typically effect sizes around 0.2 to 0.4 standard deviations. To put that in perspective: there is substantial overlap between schizophrenia and healthy control distributions on every standard structural measure. A given brain scan cannot reliably classify an individual as having or not having schizophrenia. Group-level differences are real and reproducible; individual prediction is not yet possible.
Are the changes there from the start?
Several lines of evidence suggest that some structural differences are present early — even before the first psychotic episode:
- Children at familial high risk for schizophrenia show subtle volume reductions before any symptoms
- Clinical high-risk individuals who go on to develop psychosis tend to have greater baseline cortical thinning than those who do not
- First-episode patients already show volume differences compared to age-matched controls
This pattern supports a neurodevelopmental component to schizophrenia.
Do brains change after illness onset?
Longitudinal MRI studies have documented progressive grey matter loss after the first episode, particularly in the first few years. The Northwestern early-psychosis cohort, the Iowa Longitudinal Study, and other groups have all shown trajectories of greater grey matter decline in patients than in healthy controls over follow-up periods of several years. Some of this trajectory may slow or stabilise with successful treatment.
Importantly, structural changes do not march in one direction in everyone — some patients show stable volumes over time, and recovery is consistent with structural stabilisation in many cases.
The medication question
One persistent question is whether antipsychotic medication itself contributes to grey matter loss. Studies of cumulative antipsychotic exposure have produced mixed findings. Higher-dose, longer-duration typical antipsychotic exposure has been associated with greater grey matter reductions in some studies. Atypical antipsychotics may have a smaller effect. However, untreated psychosis is also associated with worse outcomes, including possibly more brain change. The current consensus is that the modest contribution of medication to structural change is far outweighed by the benefits of preventing relapse — but the question is real and continues to be studied.
Symptom correlates
Structural findings have been linked to symptom domains:
- Hippocampal and temporal lobe changes have been associated with positive symptoms, including hallucinations.
- Frontal grey matter reductions have been associated with cognitive symptoms and negative symptoms.
- Thalamic changes have been associated with sensory gating abnormalities and cognitive impairment.
None of these correlations are strong enough to predict symptom domains in an individual, but they help build mechanistic models.
What standard clinical MRI looks like
When clinicians order brain MRI in a person with new psychosis, the goal is usually to rule out other causes — tumour, demyelinating disease, hydrocephalus, autoimmune encephalitis. Standard clinical MRI does not measure cortical thickness or hippocampal volume the way research scans do. It is a structural overview, read for gross abnormalities. In most cases of schizophrenia, the clinical MRI looks unremarkable.
Many guidelines recommend a brain MRI at first presentation of psychosis, especially in atypical presentations — late onset, focal neurological signs, rapid cognitive decline, or autoimmune features. The yield in typical schizophrenia is low but not zero.
What this means for patients and families
A few practical takeaways:
- Group-level brain differences are real but cannot be used to diagnose any one person.
- If your clinician orders a brain MRI, the goal is usually to rule out other causes — not to confirm schizophrenia.
- Early treatment may help limit progressive structural change, which is one of several reasons clinicians push for prompt and consistent treatment.
- Recovery is consistent with structural stabilisation. Brain changes do not march inevitably forward in every patient.
Where the field is going
Active directions include:
- Multimodal imaging combining structural MRI with fMRI, PET, MRS, and DTI
- Cortical surface analyses and gyrification measures
- Machine-learning models attempting individual prediction from multimodal data
- Cross-disorder ENIGMA studies comparing schizophrenia structural patterns with bipolar, autism, and depression
- Connectome-based approaches integrating structure and function
For more imaging perspectives, see our companion pieces on fMRI, PET imaging, and MR spectroscopy.
This article is for educational purposes only and is not medical advice, diagnosis, or treatment. Always consult a qualified mental health professional. If you or someone you know is in crisis, call or text 988 in the US, or your local emergency number.