Chronic schizophrenia shows a distinct MRI signal in brain waste-clearance pathways
Enlarged perivascular spaces on MRI were more common in people with chronic schizophrenia than in healthy controls, while the same signal did not separate first-episode psychosis from control participants. In a cross-sectional cohort of 174 subjects, the difference appeared in the centrum semiovale rather than the basal ganglia, pointing to a regional pattern rather than a diffuse brain-wide effect.
The finding matters because perivascular spaces are not just incidental MRI cavities. They are closely tied to the brain’s fluid exchange and waste-clearance systems, including the glymphatic pathway, and have been linked in other settings to neurovascular stress, neuroinflammation, and aging-related change. In schizophrenia, those biological themes have long hovered around the disease without a clear structural marker that tracks with stage. This study places enlarged perivascular spaces into that conversation, but only for the chronic phase.
A stage-specific pattern emerged in the centrum semiovale
The imaging analysis used high-resolution 3T MRI and a standardized rating scale, with three independent raters assessing the scans. Ordinal logistic regression showed that chronic schizophrenia carried higher odds of being in a higher centrum semiovale EPVS category than healthy controls, with an odds ratio of 2.31 and a 95% confidence interval of 1.15 to 4.62. Total EPVS burden was also higher in chronic schizophrenia, with an odds ratio of 2.02 and a 95% confidence interval of 1.04 to 3.93.
The basal ganglia did not show the same pattern. Nor did first-episode psychosis differ significantly from healthy controls in either region. That split is clinically interesting because it argues against a simple “schizophrenia equals EPVS expansion” model. Instead, the MRI signal appears to intensify with illness duration or downstream disease effects, while remaining absent, at least in this cohort, at the first psychotic presentation.
One sentence captures the most direct bedside implication: chronic illness, not first-episode psychosis, carried the structural burden.
Why the anatomy points toward vascular and inflammatory biology
The centrum semiovale and basal ganglia are not interchangeable locations for these spaces. Enlarged spaces in the centrum semiovale are often discussed in relation to impaired interstitial fluid drainage and perivascular clearance, while basal ganglia EPVS have been associated more often with small-vessel disease patterns in other neurologic contexts. A selective rise in centrum semiovale burden fits with the study’s framing around neurovascular dysfunction, neuroinflammatory activity, and accelerated brain aging in schizophrenia.
That biological interpretation remains inferential, but it is plausible. If perivascular drainage becomes less efficient, waste products and fluid may accumulate around penetrating vessels, leaving a visible footprint on MRI. In a disorder already associated with inflammation-related hypotheses and long-term brain structural change, enlarged spaces could represent one more downstream marker of chronic physiologic strain rather than a primary lesion. The data do not prove mechanism, yet the pattern is aligned with the idea that late-stage schizophrenia may involve cumulative vascular and glial stress.
First-episode psychosis did not carry the same MRI burden
The absence of a significant difference in first-episode psychosis is as important as the chronic schizophrenia signal. Early psychosis is often where researchers hope to identify biomarkers that precede or predict illness consolidation. Here, the MRI measure did not separate that group from controls, suggesting that EPVS burden may not be useful as an early diagnostic marker for psychosis broadly, at least not in this form and not at a single time point.
That negative result narrows the interpretation. It argues against EPVS as a general marker of psychosis onset and instead places it closer to a marker of chronic disease burden, illness duration, or accumulated neurobiological injury. Whether antipsychotic exposure, metabolic comorbidity, vascular risk, or other long-term disease correlates contribute to the pattern cannot be resolved from a cross-sectional design. The study design does not permit causal separation of those factors.
What the numbers do and do not support
Three independent raters and a standardized MRI scale strengthen the technical credibility of the imaging assessment. The sample included schizophrenia, first-episode psychosis, and healthy control groups, allowing stage comparison rather than a single-diagnosis snapshot. Even so, the work remains cross-sectional, so it captures association rather than trajectory.
No group difference was seen for basal ganglia EPVS. No significant difference separated first-episode psychosis from healthy controls. The signal was concentrated in chronic schizophrenia and in the centrum semiovale, which means the result is specific, not sweeping. For clinicians and researchers, that specificity is the most useful part of the paper, because it suggests future studies should not treat all perivascular spaces as one uniform biomarker.
The study was approved by the Ethical Committee of the NIMH, and all participants gave written informed consent. Funding came from the Ministry of Health of the Czech Republic and from a project financed by the European Union-Next Generation EU.
A biomarker candidate, but not a clinical tool yet
Enlarged perivascular spaces are visible on routine MRI, which makes them appealing as a potential research marker. Yet visibility is not the same as clinical readiness. Before EPVS can be used in psychiatric practice, larger studies will need to determine whether the finding replicates across populations, whether it tracks illness duration or treatment exposure, and whether it relates to cognition, negative symptoms, metabolic burden, or functional decline.
For now, the most defensible reading is narrower: chronic schizophrenia was associated with greater centrum semiovale EPVS burden, while first-episode psychosis was not. That pattern supports a model in which vascular or clearance-related brain changes accumulate over the course of illness rather than appearing at the first break from reality. The unanswered question is whether the MRI signal reflects disease biology itself, the consequences of long-term treatment and comorbidity, or a mixture of both.