146. Glutamatergic Neurotransmission Modulates Cortico-Striatal Reward Networks Differentially for Treatment-Resistant and Treatment-Responsive Schizophrenia Patients: Evidence from a Multimodal fMRI and 1H-MRS Study

摘要

>Background: Positive symptoms and specifically paranoid delusions in psychosis are characterized by a fundamental luck of trust. A number of studies show that dopaminergic dysfunction modulating the cortico-striatal reward network underlies psychosis. However, antipsychotic medication is not effective in approximately a third of schizophrenia patients. Recent evidence suggest that their dysfunction might be primarily glutamatergic, hence might affect reward networks differentially. >Methods: We acquired event-related fMRI data from from 21 treatment-resistant schizophrenia patients, 20 responsive, and 21 matched healthy controls who performed a trust game. It requires them to invest an amount between 0 and 10 pounds. The computer acted as a cooperative player and returned 100%, 150%, and 200% for each trial (with probability 33% each). Increased investment over time indicates increased trust to the other player. 1H-MRS data were acquired from the anterior cingulate to measure glutamate levels. >Results: There were no differences across the groups in initial investement, F(2, 60) = 1.28, P > .2. For mean investment, compared to controls, responsive patients showed decreased investment T(39) = 2.6, P < .05; resistant patients showed decreased investment as well that was at trend level T(40) = 1.91, P = .1, but the 2 patient groups did not differ between them T(39) = −0.48, P > .5. Whole-brain analysis of the fMRI data showed that compared to controls, responsive patients showed decreased activation in the right anterior caudate, while resistant patients showed decreased activation in the lingual gyrus and right insular cortex. No differences were found between healthy controls and resistant patients in the caudate. Caudate function and Glu/Cre levels correlated significantly in responsive patients, R = −.41, P < .05, but not resistant patients or healthy controls, P > .2. For the repayment phase, compared to controls, both patient groups showed activation reductions in the anterior cingulate, left precentral gyrus and right medial frontal gyrus. Resistant patients showed significantly increased medial prefrontal activation compared to responsive, which correlated significantly with Glu/Cre in resistant patients, R = −.47, P < .05, but not in responsive patients or healthy controls, P > .15. >Conclusion: We show for the first time that for similar behavioral deficits, resistant and responsive patients show the engagement of different neural systems associated with social cognition and reward. Crucially, glutamatergic neurotransmision modulates differentially the corticostriatal network for the 2 groups. This indicates that dissociable neurochemical circuitry and differential dopamine and glutamate interactions in reward processing might underlie antipsychotic treatment-resistance. The elucidation of these differences is crucial for informing novel treatments.