AI-driven early diagnosis of specific mental disorders: a comprehensive study

Area of Science:

• Computational psychiatry and artificial intelligence in healthcare.
• Application of machine learning and deep learning in mental health diagnostics.

Background:

• Mental disorders pose significant global health challenges, necessitating improved diagnostic tools.
• Artificial intelligence (AI) offers potential for early detection and accurate diagnosis of various mental health conditions.
• Existing diagnostic methods can be subjective and time-consuming, highlighting the need for objective, data-driven approaches.

Purpose of the Study:

• To conduct a comprehensive literature review on AI approaches for mental disorder detection.
• To analyze the effectiveness of machine learning (ML) and deep learning (DL) models across different data modalities.
• To identify key AI techniques and data preprocessing methods that enhance diagnostic accuracy in mental healthcare.

Main Methods:

• Literature review of AI applications in detecting bipolar disorder, schizophrenia, autism spectrum disorder, depression, suicidality, and dementia.
• Analysis of ML models (XGBoost, LightGBM, RF, SVM, KNN) and DL models (CNN, LSTM, GRU) applied to survey, EEG, text, and image data.
• Evaluation of data preprocessing techniques (wavelet transforms, normalization, clustering) and hyperparameter optimization for model performance enhancement.

Main Results:

• High accuracies achieved: LightGBM (96%) and SVM (97%) for anxiety/depression; XGBoost, RF, LightGBM (98%) for autism spectrum disorder.
• Deep learning models demonstrated strong performance: LSTM (83%) for schizophrenia, GRU (93%) for text-based suicide/depression detection, LSTM/GRU (99%) for dementia detection.
• Effectiveness of LSTM/GRU for sequential data and medical imaging/NLP; XGBoost/LightGBM as accurate ML tools for clinical diagnosis.

Conclusions:

• AI, particularly ML and DL, shows significant potential in improving the accuracy and timeliness of mental disorder diagnosis.
• Advanced data preprocessing and hyperparameter optimization are crucial for maximizing AI model performance in clinical settings.
• AI-driven clinical decision support systems can facilitate early diagnosis and personalized treatment strategies for mental health conditions.