Leveraging Stacked Classifiers for Multi-task Executive Function in Schizophrenia Yields Diagnostic and Prognostic...

Area of Science:

• Neuroscience
• Psychiatry
• Cognitive Psychology

Background:

• Cognitive impairment, especially executive functioning (EF) deficits, is a hallmark of schizophrenia.
• Understanding specific EF dimensions' roles in diagnosis and prognosis is crucial due to schizophrenia's heterogeneity.

Purpose of the Study:

• To investigate how specific dimensions of EF can discriminate individuals with schizophrenia from healthy controls.
• To predict remission status in schizophrenia patients based on baseline EF measurements.

Main Methods:

• Utilized stacking classification models on systematically measured EF data from 195 schizophrenia patients and healthy individuals.
• Assessed feature importance of EF dimensions for diagnosis and prognosis, linking individual scores to symptom dimensions.

Main Results:

• Executive functioning assessments, specifically inhibitory control and working memory, accurately predicted schizophrenia diagnosis (AUC=0.87) and remission status (AUC=0.81).
• Interference inhibition and working memory updating were key predictors for identifying individuals with schizophrenia and those in remission.
• Patients identified via EF had high baseline negative symptoms; those who remitted showed milder follow-up cognitive symptoms.

Conclusions:

• Specific EF dimensions are differentially associated with symptom load and prognostic outcomes in schizophrenia.
• EF assessments and predictive models show promise for aiding clinical evaluation and understanding schizophrenia's complex presentation.