This study conducts a thorough examination of machine learning models within Artificial Intelligence (AI)- integrated mental fatigue research, focusing on their identification and assessment. Investigating various biomedical signals linked to mental fatigue-such as Electroencephalogram (EEG), Electrocardiogram (ECG), and Galvanic Skin Response (GSR) this research sheds light on their roles in deciphering cognitive states and emotional responses. Through integration, these signals facilitate the creation of network visualizations, unveiling the intricate relationship between physiological markers and cognitive aspects during mental fatigue scenarios. Additionally, this study identifies gaps in current mental fatigue research with AI, emphasizing the need for an integrated approach across diverse domains and populations. It outlines future directions, suggesting avenues like multi-domain integration, personalized interventions, ethical considerations, longitudinal studies, and interdisciplinary collaborations to propel advancements in mental fatigue research, offering promising prospects for innovative AI-driven solutions in mental health and well-being.