As human-computer interaction (HCI) technology advances, the use of augmented reality (AR) in cognitive training is becoming more prevalent. However, traditional training methods often apply a one-size-fits-all approach, failing to accommodate the varied training needs of individuals with different cognitive levels. Additionally, most HCI systems use subjective questionnaires for evaluation, which can be influenced by the subjects' emotional and mental states. To overcome these challenges, this study developed an AR-based adaptive HCI cognitive training system that dynamically adjusts task difficulty based on real-time user performance. We used multi-source data to empirically validate the effectiveness of adaptive HCI in cognitive training. Specifically, we recorded functional Near-Infrared Spectroscopy (fNIRS) data, movement data, task performance, and subjective feedback from 22 elderly participants, dividing them into two groups—low cognitive group and normal cognitive group. The results showed that the system exerted a significant influence on brain functional connectivity (FC) associated with cognition, movement, and vision. Changes in FC may highlight the benefits of adaptive HCI training strategies. Furthermore, participants with normal cognitive abilities significantly outperformed their low cognitive counterparts in task performance. In conclusion, this study designed and evaluated an AR-based adaptive HCI cognitive training system that ensures personalized training. It demonstrated the feasibility of adaptive HCI strategies in cognitive rehabilitation by incorporating physiological and behavioral data, thereby enhancing the precision of quantitative assessments for HCI systems.

Automation has been extensively used in flight operations, so pilots are less involved in actual flight control. With the long idle time during cruising, pilots may have their vigilance level reduced and eventually become out-of-the-loop. This research proposes a two-stage explainable adaptive automation approach to keep pilots in the loop based on Convolutional Neural Networks, Long Short-Term Memory, and EEG data collected from 24 participants in a one-hour simulator-based flight task in each level of automation. Our proposed spatiotemporal model yields test accuracy of 0.9918 and 0.9907 in the first and second stages, respectively, outperforming other benchmarking models by 30.79% and 10.73%, respectively. Furthermore, the Shapley additive explanations are adopted to strengthen the model interpretability and trustworthiness for safety-critical applications. Our model successfully identified that high delta and theta waves with low beta and gamma waves contribute positively to the out-of-the-loop state. It indicates that the classification aligns with the theoretical background and is trustworthy. The trustworthy adaptive deep learning model supports the dynamical automation configuration for improving human-automation collaboration in cruising flights.