Psychological disorder detection: A multimodal approach using a transformer-based hybrid model.

Area of Science:

• Psychology
• Computer Science
• Artificial Intelligence

Background:

• Detecting psychological disorders like depression is challenging due to subjectivity and bias in traditional methods.
• Existing diagnostic techniques often lack comprehensive assessment of psychological distress.

Purpose of the Study:

• To develop a novel approach for improved identification of psychological distress, focusing on depression.
• To address limitations of traditional diagnostic methods by using multimodal data.

Main Methods:

• Developed a Transformer-Based Hybrid Model integrating advanced natural language processing and deep learning.
• Utilized multimodal data, including voice characteristics (aural cues) and linguistic content (textual data) from participant interviews.

Main Results:

• The model offers a comprehensive assessment of psychological well-being by merging aural and textual data.
• Demonstrated a novel method for strengthening emotional ailment detection, particularly for depression.

Conclusions:

• The Transformer-Based Hybrid Model provides a promising, replicable approach for psychological disorder detection.
• This multimodal strategy enhances the accuracy and reduces bias in identifying conditions like depression.