Contribution of substantia nigra glutamate to prediction error signals in schizophrenia: a combined magnetic resonance spectroscopy/functional imaging study

BACKGROUND: Because dopamine neurons signal a mismatch between expected and actual reward called prediction error (PE), aberrant PE signals in schizophrenia have been attributed to known dopaminergic abnormalities. However, dysfunction of N-methyl-D-aspartate receptors on cortical γ-aminobutyric acid neurons, as hypothesized in schizophrenia, could lead to excess glutamate release in the substantia nigra (SN) and affect reward processing. AIMS: The aim of this study was to investigate the contribution of SN glutamate to PE signals in healthy controls (HC) and patients with schizophrenia (SZ). METHODS: We recruited 22 medicated SZ and 19 HC. We obtained (1) functional magnetic resonance imaging during a probabilistic monetary reward task to assess PE-related blood oxygen level-dependent (BOLD) signal and (2) magnetic resonance spectroscopy to measure Glx (glutamate+glutamine) in the SN. To identify group differences in regions where the BOLD signal varies as a function of PE, we analyzed PEs generated during the task as parametric modulators of reward delivery. Finally, we examined the correlation of PE-related BOLD signal and SN Glx in each group. RESULTS: Relative to HC, PE-related BOLD signals in SZ were significantly different in the midbrain/SN and ventral striatum. In SZ, SN Glx was significantly elevated. In HC, but not in SZ, PE-related BOLD signal in SN was positively correlated with SN Glx. CONCLUSIONS: These results suggest a role of glutamate in the neural coding of PE in controls. They indicate that glutamatergic dysfunction might contribute to abnormal PE coding in schizophrenia, suggesting the use of glutamate-targeted approaches to improve these deficits.