Visual blends is a design technique that combines elements from multiple images into harmonious compositions and has been increasingly explored as a means to support early-stage ideation in engineering design. However, existing blending workflows rely heavily on manual image selection and composition, making the process difficult, time-consuming, and skill-intensive for designers. In this work, we present a proof-of-concept brain-guided visual blends technique that integrates an EEG-to-image model to simplify the image acquisition process and a local image editing model to enable automated and controllable image composition. Our EEG-to-image model employs a two-stage training strategy, combining pretraining on large-scale unlabelled EEG data with fine-tuning in an EEG-conditioned diffusion model, achieving state-of-the-art performance in reconstructing visual stimuli. To support visual blending tasks, we incorporate a local editing model (Paint-by-Example) that generates coherent blends using user-provided masks, reference images, and backgrounds. A user study with 15 participants demonstrated that the model effectively supported the creation of visual blends that aligned with users' design vision, even without artistic skills. The results suggest that brain-guided blending can serve as a early-stage ideation interface in engineering design, helping designers iterate on mental concepts before formal modelling and evaluation.

Generative AI reshapes creative workflows through mixed-initiative collaboration, requiring users to rapidly make decisions among alternatives and fluidly shift roles as creators and critics. During exploration, users often rely on emotional intuition for swift judgments. However, most current systems depend on explicit prompts, limiting their adaptivity across iterations. Meanwhile, EEG-based neuroadaptive systems dynamically adjust interactions, improving human–machine alignment. EmotivChat addresses this gap by continuously monitoring users’ immediate emotional responses. Within a closed-loop architecture, the system optimizes prompting strategies across rounds. The system was evaluated across two paired studies on image co-creation(N=31) and text ideation (N=29) each contrasting the EEG-adaptive system with a rigorously matched non-adaptive baseline. Results indicate EmotivChat improved perceived teaming and iteration efficiency in image co-creation, while in text ideation, it improved perceived collaboration but produced limited changes in idea outcomes. These findings validate real-time emotional feedback in LLM-based interactions, pointing the way toward cross-modal, emotion-adaptive agents.