Can the MMPI Predict Adult ADHD? An Approach Using Machine Learning Methods.

Area of Science:

• Psychiatry
• Psychology
• Computer Science

Background:

• Adult ADHD presents diagnostic challenges due to overlapping psychopathology.
• Impulsivity and attention deficits in adults significantly impact social functioning.

Purpose of the Study:

• To investigate the efficacy of machine learning (ML) techniques in predicting adult ADHD symptoms.
• To determine if the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) can be utilized for adult ADHD prediction via ML.

Main Methods:

• Analysis of data from 5726 college students.
• Utilized the MMPI-2-Restructured Form (MMPI-2-RF) and the Adult ADHD Self-Report Scale (ASRS).
• Applied K-nearest neighbors (KNN), linear discriminant analysis (LDA), and random forest algorithms.

Main Results:

• Achieved high prediction accuracies: KNN (93.1%), LDA (91.2%), and Random Forest (93.6%).
• Linear Discriminant Analysis (LDA) demonstrated the highest Area Under the Curve (AUC) at 0.806, indicating excellent diagnostic capability.
• Machine learning models showed significant potential in identifying adult ADHD symptoms.

Conclusions:

• Machine learning models, particularly LDA, show high accuracy in screening for adult ADHD using MMPI-2 data.
• The MMPI-2, when analyzed with ML, offers a reliable tool for identifying adult ADHD symptoms in large populations.
• This study supports the use of ML for efficient and accurate adult ADHD screening.