Quantifying abnormal emotion processing: A novel computational assessment method and application in schizophrenia

Area of Science:

• Neuroscience
• Psychiatry
• Computational Linguistics

Background:

• Abnormal emotion processing is central to schizophrenia spectrum disorders (SSDs).
• Emotion processing via language is crucial for social functioning but less understood in SSDs.
• Existing methods for assessing emotion processing are limited.

Purpose of the Study:

• To introduce and validate a novel deep learning method for assessing emotion processing in spoken language.
• To investigate emotional alignment (EA) in male-identified patients with SSDs compared to controls.
• To examine the effects of oxytocin on EA and formal thought disorder (FTD) in SSDs.

Main Methods:

• Developed a deep learning model to derive an 'emotional alignment' (EA) measure from spoken language.
• Tested the method in male-identified patients with SSDs (n=37) and healthy controls (n=51) using free responses.
• Administered oxytocin and assessed EA and FTD changes in patients.

Main Results:

• Patients with SSDs exhibited significantly impaired EA compared to controls.
• Impaired EA correlated with deficits in social cognition, motivation, and pleasure.
• EA improved classification of SSDs when added to FTD measures.
• Oxytocin administration enhanced EA in patients but did not affect FTD.

Conclusions:

• Automated assessment of emotional alignment from spoken language is a promising tool for SSDs.
• EA deficits are linked to core symptom domains in SSDs.
• Oxytocin may enhance socioemotional processing in SSDs, as measured by EA.