<p>Assessing students' comprehension and identifying learning challenges in traditional physical courses is a challenging task, as it often lacks the data-driven techniques used in online learning environments. This study introduces a novel approach to address this issue by leveraging physiological signals and accounting for individual differences in learning-related personality traits. Previous research has shown that emotional responses are linked to learning challenges. Physiological signals, such as electrocardiogram (ECG), galvanic skin response (GSR), and heart rate (HR), offer a potential avenue to detect these emotional responses. The study involved two phases. In the first phase, participants' physiological signals were recorded while they answered a quiz, and their perceived difficulty levels were self-assessed. To account for individual differences in learning-related personality traits, participants were categorized into four groups based on their conscientiousness and openness to experience, both of which have been identified in prior literature as significant predictors of academic performance. Random Forest models were then trained for each group, using physiological signal features as input and perceived difficulty levels as output. This approach significantly improved accuracy compared to a single, generalized model. In the second phase, the developed models were tested in a “Discrete Mathematics" course with 48 enrolled students, and the results provided a preliminary indication of the system's feasibility in a real classroom setting, with students reporting that the system's difficulty assessments were broadly plausible. The proposed approach is adaptable to both physical and online courses, automated, and capable of real-time assessment. Future work may focus on refining and expanding the model to encompass a broader range of personality traits and learning contexts.</p>