Emotion processing in schizophrenia and its relation to μ-opioid receptor availability

Emotion processing deficits are core features of schizophrenia, and there is interest in understanding their neurochemical basis. Preclinical evidence implicates opioid neurotransmission in emotional regulation, however the role of the μ-opioid receptor (MOR) in human emotion processing in schizophrenia remains unclear. Here, we examined the association between central MOR availability and emotion processing in patients with schizophrenia.

Nineteen patients with schizophrenia and 19 age- and sex-matched healthy controls underwent functional MRI (fMRI) while performing two emotion-processing tasks: viewing emotional pictures from the International Affective Picture System (IAPS) and a facial emotion recognition task (happy, sad, neutral faces). All participants also underwent a [¹¹C]-carfentanil positron emission tomography (PET) scan to measure MOR availability (binding potential, BP_ND). Group differences in brain activation and MOR binding were analysed, and associations between MOR availability and fMRI activation in regions of interest (ROIs) were assessed.

Patients with schizophrenia had mild-moderate negative symptoms (PANNS-negative 21.4 ± 1.0, SANS total 55.7 ± 5.0). Both the IAPS (pleasant + unpleasant > neutral) and facial-emotion (happy + sad > baseline) contrasts engaged the canonical cortico-limbic "hedonic" network (amygdala, anterior cingulate cortex (ACC), insula, orbitofrontal cortex and striatum) in each group. Multivariate ROI analyses showed no significant case-control difference for either task (IAPS: Pillai = 0.276, p = 0.055; Faces: Pillai = 0.277, p = 0.069), and no individual ROI survived false-discovery correction (all pFDR > 0.05). Mixed-effects models indicated that [¹¹C]-carfentanil BPND did not predict emotion-evoked BOLD signal across ROIs in either group for the IAPS task (controls β = -4.6 ± 3.0, p = 0.13; patients β = -1.0 ± 3.4, p = 0.76; group × PET interaction p = 0.20) or the facial-emotion task (controls β = 5.1 ± 4.3, p = 0.24; patients β = 0.7 ± 4.8, p = 0.89; group × PET interaction p = 0.28. Exploratory correlations identified a negative association between MOR and BOLD response related to viewing emotional faces in the ACC in controls only (sad + happy > baseline, r = -0.60, pFDR = 0.03).

We found no evidence of impaired emotion-related brain activation in the hedonic network in schizophrenia patients with moderate negative symptoms compared to healthy individuals, and baseline MOR availability did not predict the neural response to emotional stimuli in this network.