Cognitive decline is a critical area of research due to its profound impact on neural integrity, cognitive function, and its association with neurodegenerative diseases. Early identification of cognitive impairment is essential, as it often signals underlying neurological dysfunction, which, if left unaddressed, can lead to progressive mental deterioration. Moreover, cognitive decline extends beyond individual health, influencing high-demand environments where sustained cognitive performance is crucial for safety and decision-making. Heart rate variability (HRV), derived noninvasively from photoplethysmography (PPG), offers a real-time method for detecting autonomic dysregulation linked to cognitive fatigue. Continuous PPG monitoring under conditions of sleep deprivation, combined with machine learning algorithms such as Long Short-Term Memory (LSTM) networks, enabled accurate prediction of cognitive states based on HRV patterns through their ability to capture temporal dependencies. The findings reveal significant autonomic disturbances corresponding to mental fatigue, underscoring HRV’s potential as a sensitive biomarker for cognitive decline and its applicability in transfer learning frameworks.