Human-computer cooperation is emerging as a key trend in human-computer interaction. Humans and AI communicate through interface layouts structured by semantic dialogue logic. This study investigates the cognitive characteristics of humans when interacting with different interface layouts (Copilot and OS agent) and AI roles under varying modes of human-computer cooperation. By integrating behavioral and eye-tracking data, this research examines how Copilot and OS agents and interfaces are influenced by semantic logic (generative and task-oriented) respectively. The leading role of either humans + AI or AI in cooperation determines the appropriate semantic logic, which subsequently shapes the corresponding interface layout. This study further clarifies how different modes of human-computer cooperation, interface layouts, and dialogue logics are interrelated, and how these factors collectively influence cooperative behavior. Additionally, it explores the underlying cognitive mechanisms of interface design through the lens of semantic dialogue logic and the cooperative distribution between humans and AI.

The human–computer cooperation process guided by natural interaction, intelligent interaction, and human–computer integration is gradually becoming a new trend in human–computer interaction. Cooperative scenarios of human–computer interaction systems often contain multi-interface and multi-device results in edges often interrupt the cognitive ergonomics of interface layout. This research takes typical areas as an example to establish a stepwise regression model to predict reaction time at an arbitrary position on the left interface. It uses a foveal region to position the starting point of attention and a parafoveal region to calculate the radius of each objective area, and design 10 similar tasks to analyze eye-tracking indexes through physiological assessment. Unlike fixed thinking such as spatial proximity on multi-interfaces, this research summarises cognitive features of layout based on the positive and negative effects of edge impact through eye-tracking analysis. It analyzes cognition including input, process, and output in human–computer cooperation from human intelligence and artificial intelligence respectively, and visualises the mapping relationship between these indexes and specific stages of cognition. Besides, the quantitative evaluation of the regression equation and qualitative analysis of the eye-tracking indexes provide a reference for other interfaces around the front interface.

While considerable research has been conducted, the field of neurodesign and human-computer interaction (N-HCI) has yet to be methodically delineated in light of contemporary studies. In this context, this paper undertakes a systematic review with dual objectives: (a) to uncover emerging research themes within N-HCI, and (b) to provide future directions for research on N-HCI. Employing robust systematic review methodologies, this paper scrutinized 122 pertinent documents, revealing four principal emerging research themes and constructing a conceptual framework that synthesized key concepts in N-HCI. The results underscored the relative underdevelopment in research that leverages neural evidence to enhance general human-computer interface design, and the notable absence of research at the intersections of the “user” and “environment” aspects of neurodesign with HCI. Accordingly, this paper identifies underdeveloped areas in existing research that require more depth and blank areas in the conceptual framework that call for exploration as future directions.

Given that the unreasonable layout caused by the surge of map information will dramatically reduce the user's operation performance, this paper proposed a 3 V layout method based on human visual motion cognition. According to the visual flow method, two visual search experiments were designed to study the influence of label layouts and graphic overlap rates on users’ search efficiency. By testing users’ reaction time and gaze behaviour, it was found that L-LR layouts arranged in a vertically symmetrical way had the best search performance among the four layouts. The overall performance of L-LR and L-TB layouts arranged in a unidirectional and symmetrical way was better than that of L-UG and L-EP layouts in the graphic and number search task. However, L-UG and L-EP layouts performed better than L-LR and L-TB ones in the graphic search under lower overlap rates. Moreover, it is also worth noting that the performance of graphic search was significantly higher than that of number search, indicating that users distinguished the location-based information represented by graphics better than that by text. The results provide a certain theoretical basis and application value for the map information layout of the GIS interface.

This paper proposes a B2C Web site emotion feedback system based on emotion recognition, which helps users to regulate their emotions by providing relevant suggestions, thereby improving user Emotional Experience (EE) in the process of Human-Computer Interaction (HCI). Adjust the aesthetics and usability of the B2C Web site to change the emotions of 16 users, combining subjective evaluation and Baidu’s facial expression recognition service to measure the user’s EE. After that, the emotion feedback system is generated in combination with the service, and the user’s emotions are displayed in real time through color symbol expressions. When negative emotions are recognized, a dialog box will pop up to remind users to relax. The results show that this emotion feedback system aimed at regulating users' emotions can be applied to B2C Web sites. This study provides more directions for the design research of user EE in the future.