03 AI-driven modes revea cognitive dynamics of coaborative design innovation: A muti-moda approach to understanding designer interaction patterns

Abstract

Through the collaborative design process, professionals engage in complex cognitive activities, including creative generation and critical evaluation, to produce innovative solutions. However, the cognitive mechanisms that integrate these complementary processes (precisely conveying and distilling design knowledge through these processes) remain little known. Here, we use pre-trained deep learning artificial intelligence models in combination with multimodal cognitive monitoring to identify neural and behavioral signals that reflect design creation, evasion, and transformation among professionals in natural collaborative design meetings. The research results show that cognitive activities reflecting design creation and evaluation are widely distributed in the attention and working memory networks of various measurement methods. We also found that these activities are specific to the design concepts and solutions under development and are based on the combination of specific environments and design elements. Finally, we demonstrated that these cognitive models were overused during the design generation and evaluation stages, and that the transition from creator to evaluator was associated with specific, time-limited changes in cognitive activities. Furthermore, our research findings reveal the dynamic organization of cognitive activities that serve design creation and evasion in natural collaborative work, and utilize deep learning models to understand the cognitive mechanisms behind human design innovation.