Neuroimaging in schizophrenia: an overview of findings and their implications for synaptic changes

Abstract

Over the last five decades, a large body of evidence has accrued for structural and metabolic brain alterations in schizophrenia. Here we provide an overview of these findings, focusing on measures that have traditionally been thought to reflect synaptic spine density or synaptic activity and that are relevant for understanding if there is lower synaptic density in the disorder. We conducted literature searches to identify meta-analyses or other relevant studies in patients with chronic or first-episode schizophrenia, or in people at high genetic or clinical risk for psychosis. We identified 18 meta-analyses including over 50,000 subjects in total, covering: structural MRI measures of gyrification index, grey matter volume, grey matter density and cortical thickness, neurite orientation dispersion and density imaging, PET imaging of regional glucose metabolism and magnetic resonance spectroscopy measures of N-acetylaspartate. We also review preclinical evidence on the relationship between ex vivo synaptic measures and structural MRI imaging, and PET imaging of synaptic protein 2A (SV2A). These studies show that schizophrenia is associated with lower grey matter volumes and cortical thickness, accelerated grey matter loss over time, abnormal gyrification patterns, and lower regional SV2A levels and metabolic markers in comparison to controls (effect sizes from ~ -0.11 to -1.0). Key regions affected include frontal, anterior cingulate and temporal cortices and the hippocampi. We identify several limitations for the interpretation of these findings in terms of understanding synaptic alterations. Nevertheless, taken with post-mortem findings, they suggest that schizophrenia is associated with lower synaptic density in some brain regions. However, there are several gaps in evidence, in particular whether SV2A findings generalise to other cohorts.

Fig. 1. Overview of in vivo neuroimaging methods, summarising the main target and advantages and disadvantages in terms of capturing information related to synapses.

The specificity line is not linear but illustrates the rank order of specificity for synapses. PET positron emission tomography, SV2A synaptic vesicle protein 2 A.

Fig. 2. Brain volume differences in patients with first episode psychosis relative to controls.

Meta-analytic results of regional brain volume differences between patients with first episode psychosis and healthy control subjects, adapted with permission from Brugger and Howes (2017). ACC anterior cingulate cortex.

Fig. 3. Estimated percentage change in regional grey matter volumes per year in patients with schizophrenia versus healthy controls in a meta-analysis of 19 longitudinal studies, including data from 813 patients and 718 controls.

Copied with permission from Vita et al., 2012.

Fig. 4. Illustrating the estimated relative composition by volume of 1 mm³ of human cortical grey matter, showing that the major constituents are axons and dendrites, with synapses contributing relatively little.

Adapted with permission from Bennett 2011.

Fig. 5. The relationship between spine density and grey matter volume in a rodent experiment.

Scatter plot showing a significant correlation between spine density values with the voxel-based morphometry signal (Grey Matter Voxel Intensity (GMVI) x Jacobian) in the auditory cortex of mice in a fear-conditioning experiment (n = 9 per group p = 0.027, Pearson’s correlation R² = 0.2042, blue = control group, red = auditory fear-conditioned group). Adapted with permission from Keifer et al., 2015.

Fig. 6. Whole brain maps of the mean [11 C]UCB-J volume of distribution (V_T) in schizophrenia (left) and healthy control (right) groups, indicating lower levels of the synaptic vesicle protein 2 A (SV2A) marker in schizophrenia particularly in fronto-temporal regions, consistent with lower synaptic density in schizophrenia (based on data from Onwordi et al. 2020).

Low [11 C]UCB-J uptake (dark blue) corresponds to V_T values in the 10–15 ml/cm³ range, high [11 C]UCB-J (orange-red) corresponds to V_T values in the 20–25 ml/cm³ range. Image credit: Onwordi and Howes (2022).