The terms “brain-computer interface” and “brain-computer interaction” are closely related, but they emphasize different aspects of brain-computer systems. The ongoing misinterpretation of these terms has impeded the accurate classification of applications and research studies, making this clarification essential for advancing the field. The primary aim of this study is to clarify the confusion in the literature by highlighting the distinctions between “brain-computer interface” and “brain-computer interaction”, as well as to explore the relationship between these two concepts. Clarifying these definitions will help establish a more consistent theoretical framework and improve the comparability of research findings. Moreover, it will support the development of user-centered systems that integrate both technical performance and experiential dimensions. In doing so, this study seeks to contribute to a more coherent understanding of how humans and computers can communicate through neural activity. Through a conceptual analysis supported by a structured review of recent literature, the findings demonstrate that while the two terms are frequently used interchangeably, they reflect distinct emphases in both system design and research orientation. Brain-computer interface traditionally denotes the technical mechanism enabling neural signal translation and device control, whereas brain-computer interaction encompasses a broader, bidirectional, and user-centered perspective that integrates feedback, adaptability, and experiential dimensions. In conclusion, establishing clear and consistent use of these terms will contribute to a more coherent scientific discourse and facilitate progress toward intelligent, responsive, and ethically grounded brain-computer systems.
Erman Çakıt, W. Karwowski· AHFE International· 1 citation
Risk prediction in conflict-affected environments requires models that can learn from evolving data streams. This study develops an incremental recurrent learning framework for adverse-event count prediction and evaluates it in a field study of human-social-infrastructure projects in Afghanistan. The framework implements GRU-Linear (GRU-L) and LSTM-Linear (LSTM-L) predictors that encode short historical sequences with a recurrent layer and map the learned representation to nonnegative count predictions through a fully connected regression head. For each arriving instance, the model first predicts the target and then updates its parameters after the true value becomes available, following a prequential (test-then-train) evaluation protocol. We compare these predictors with statistical, ensemble, gradient-boosted, recurrent batch, and rule-based stream baselines under a unified chronological protocol. Batch models use expanding-window refitting, whereas stream models use prequential test-then-train evaluation, ensuring identical information availability constraints across all model classes. The experiments report MAE, RMSE, paired 400-instance block-level significance tests, sliding-window error trajectories, Page-Hinkley prediction-error drift alarms, and runtime. The results show that LSTM-L obtains the lowest MAE for Events and Wounded, and GRU-L obtains the lowest MAE for Hijacked and Dead. Linear Regression (LR) and TargetMean provide the lowest runtimes, while the recurrent stream models offer stronger MAE performance, particularly when recent temporal patterns are predictive. The study demonstrates that incremental recurrent learning can support adaptive risk monitoring when adverse-event distributions change over time.
Donghui Shi, Jifei Cheng, Yakun Sun et al.· IEEE Access· 0 citations
