In contemporary operational environments, increasingly shaped by automation, task variability, and decision-making complexity, the ability to monitor stress and cognitive workload represents a critical challenge for both human performance and system safety. Traditional assessment approaches, largely grounded in subjective self-report instruments such as the NASA Task Load Index, provide valuable insights but are inherently retrospective and discontinuous, thus revealing clear limitations in dynamic settings. Against this background, this thesis investigates the use of physiological signals to enable an objective and near real-time assessment of stress, with particular attention to photoplethysmography (PPG). The research is embedded within a multimodal experimental framework integrating electroencephalography, eye tracking, electrodermal activity, and cardiovascular monitoring.