Enhancing Engineering Design Through Neurofeedback: A Brain and Behavior Analysis

Abstract

This dissertation investigated the effect of real-time functional near-infrared spectroscopy (fNIRS)-based neurofeedback on engineering design ideation in professional practice. One hundred twenty practicing civil engineers across ten firms in the mid-Atlantic United States were randomly assigned to one of three conditions: real-time neurofeedback, sham placebo feedback, or no feedback. All participants completed six open-ended design ideation prompts while prefrontal cortex (PFC) activation was continuously recorded. Behavioral, neural, and semantic outcomes were analyzed using network analysis, natural language processing, semantic embedding, and machine learning classification. Network-level analysis revealed lower functional connectivity density and a shift in hub activity toward the dorsolateral prefrontal cortex (DLPFC) in the feedback group, suggesting a reorganization of neural resources toward regions associated with working memory and attentional control. The neurofeedback group also produced more ideas and more novel ideas than both comparison groups, with effects concentrated on divergent ideation prompts and absent on process-oriented tasks. Semantic trajectory analyses showed that neurofeedback participants sustained divergence from their initial ideas over longer ideation sequences, though effects on global semantic reach were selective and context-dependent. Machine learning classification of fNIRS featured failed to reliably distinguish group membership above chance, establishing an important boundary condition on the detectability of neurofeedback effects from neural data alone. Together, these findings position neurofeedback as a mechanism for supporting sustained, efficient cognitive engagement during demanding creative work, and demonstrate the feasibility of deploying neuroadaptive tools in real-world professional engineering settings.