O8.1. ASSOCIATIONS BETWEEN REWARD ALTERATIONS AND THALAMIC GLUTAMATE LEVELS IN ANTIPSYCHOTIC-NAïVE FIRST-EPISODE PATIENTS WITH PSYCHOSES

Abstract

Background: Alterations of the brain reward system have been found in schizophrenia patients and prediction error coding has been suggested to underlie the formations of psychotic symptoms. Prediction errors are the discrepancy between what actually happens and what we expect to happen. This misprediction has been suggested to contribute to misallocation of attention leading to generation of psychotic symptoms. Recent data from our group have shown that thalamic glutamate levels are not only heritable and linked to liability for schizophrenia, but also increased in the thalamus at illness onset; the latter was driven by patients that subsequently did not respond to treatment. However, it remains unresolved if glutamatergic dysfunction is correlated to striatal reward processing and prediction errors. In the present study, we aim to explore the relationship between thalamic glutamate levels and striatal prediction error coding in psychosis before the patients first antipsychotic treatment. Method(s): 40 antipsychotic-naive patients with first-episode psychosis (age 22.8+/-5.7, 15(38%) males) and 38 healthy controls matched on age and gender (age 21.8+/-3.2, 17(45%) males) underwent functional Magnetic Resonance imaging (fMRI) and magnetic resonance spectroscopy (1HMRS) (3 Tesla scanner). The patients were moderately ill with a mean PANSS total score of 74 (15). Glutamate levels scaled to creatine was estimated using the PRESS sequence in a 2.0x1.5x2.0 cm voxel in the left thalamus and analyzed using LCModel. Reward related brain activity was examined using a Monetary Delay Task. Analyses were focused on evaluation of outcome during trials where the outcome was worse than expected indicating negative prediction error (miss-contrast). Tools from the FMRIB Software Library were used to extract ROI parameter estimates from the left nucleus caudatus. Psychopathology was assessed with the Positive and Negative Syndrome Scale (PANSS). Result(s): For reward related brain activity, significant group difference was observed in the miss-contrast (p=0.019) with patients having a higher contrast activity. For thalamic glutamate there was a trend towards higher levels in the patients (p=0.057). In patients, a significant negative correlation was observed between glutamate levels and activity in the miss-contrast (r=-0.41; p= 0.009), which was not present in healthy controls. There was no significant correlation between symptom severity and prediction error activity or glutamate levels. Discussion(s): Our findings support that patients have aberrant prediction error signaling and increased thalamic glutamate levels. However, patients with increased thalamic glutamate levels display a more normal prediction error signaling, whereas patients with lower thalamic glutamate levels display abnormal prediction error signaling. This could suggest a role of thalamic glutamate in the neural coding of prediction errors in antipsychotic- naive first-episode psychosis patients but not in healthy controls.