Dependence Graphs Based on Association Rules to Explore Delusional Experiences.

Area of Science:

• Psychopathology
• Network Science
• Data Mining

Background:

• Current methods for estimating dependence graphs in psychopathology have limitations.
• Existing approaches often fail to align with data characteristics and disregard individual patient dynamics.
• Traditional methods overlook complex, higher-order interactions between symptoms.

Purpose of the Study:

• To develop a novel method for constructing dependence graphs in psychopathology research.
• To address limitations of existing methods by incorporating individual-level dynamics and higher-order interactions.
• To apply Association Rules to binary psychopathology data for more accurate network construction.

Main Methods:

• Utilized Association Rules, specifically the apriori algorithm, for constructing dependence graphs.
• Applied the method to binary records of 12 delusional experiences in 1423 subjects with psychotic disorders.
• Focused on identifying higher-order interactions beyond pairwise symptom relationships.

Main Results:

• The Association Rule-based method facilitated intuitive interpretation of symptom dependencies.
• Revealed previously undetected relevant dependencies among psychotic experiences.
• Identified unique interaction patterns within subgroups of patients.

Conclusions:

• The novel method offers a more comprehensive approach to understanding symptom networks in psychopathology.
• Association Rules effectively uncover complex, higher-order interactions in heterogeneous datasets.
• This approach enhances the validity and interpretability of dependence graphs in clinical research.