Recent Advances in EEG-based Classification of Neuropsychiatric Disorders Using Deep Sparse Neural Networks with Gooseneck Barnacle Optimization Algorithm: A Systematic Review

Abstract

Neuropsychiatric disorders such as depression, schizophrenia, anxiety, and epilepsy pose significant diagnostic challenges due to their complex neurophysiological characteristics and subjective clinical assessment methods. Electroencephalography (EEG) has emerged as a promising non-invasive modality for capturing brain activity and identifying neural biomarkers associated with these disorders. However, EEG signals are highly non-linear, noisy, and high-dimensional, necessitating advanced computational approaches for effective classification.

Recent advancements in artificial intelligence, particularly deep learning and sparse neural networks, have significantly improved EEG-based classification performance. Deep sparse neural networks enable efficient feature extraction while reducing redundancy and computational complexity. Additionally, optimization algorithms such as the Gooseneck Barnacle Optimization Algorithm (GBOA) have been introduced to enhance model training, parameter tuning, and convergence.

This systematic review analyzes recent studies focusing on EEG-based classification of neuropsychiatric disorders using deep sparse neural networks and optimization techniques. Comparative analysis demonstrates that hybrid models combining deep learning with optimization strategies outperform traditional approaches, achieving classification accuracies above 90% in several cases.

The review also highlights key trends, challenges, and future directions, emphasizing the need for explainable AI, large-scale datasets, and real-time clinical deployment. These findings indicate that AI-driven EEG analysis has strong potential to revolutionize neuropsychiatric disorder diagnosis.