Decoding Consumer Decision-Making in Digital Advertising Through Eye-Tracking and Physiological Signals

Background: Consumer decision-making in digital advertising is a multifactorial process involving attention, emotional engagement, cognitive evaluation, and contextual influences. Traditional self-report methodologies fail to capture the automatic, unconscious, and real-time nature of these processes. Neuromarketing research, utilizing eye-tracking and physiological signals such as electroencephalography (EEG), electrodermal activity (EDA/GSR), heart rate, and event-related potentials (ERPs), offers objective, millisecond-resolution measurement of consumer responses during advertising exposure. Methods: Following systematic classification principles, we identified 19 empirical studies from a corpus of 69 papers as quantitatively eligible for meta-analytic synthesis. Studies were required to employ at least one eye-tracking or physiological measurement tool, examine consumer behavior or decision-making in digital or visual commercial contexts, and report sufficient quantitative data for synthesis (means, standard deviations, sample sizes, effect estimates, and/or p-values). A random-effects synthesis framework was adopted given the methodological heterogeneity across studies. Subgroup analyses were conducted by tool type (eye-tracking only, EEG-based, EDA/GSR, multimodal, machine learning pipeline, ERP), stimulus type (video, static/print, food-related, warning labels), and outcome type (attention, emotional arousal, neural decoding, preference/choice, effectiveness). Results: The 19 eligible studies (total N across studies > 1,700) encompass diverse designs including between-subjects experiments, within-subjects repeated-measures paradigms, mega-analyses of multiple datasets, and machine learning prediction pipelines. Key findings include: (1) gaze fixation duration strongly predicts final consumer choice (r ≈ 0.62–0.80, e.g., wine label design studies); (2) frontal EEG gamma activity predicts commercial media success beyond self-reported liking; (3) the Late Positive Potential (LPP) reliably indexes emotional processing intensity in response to warning labels and food stimuli; (4) Random Forest machine learning classifiers achieve 81% accuracy in predicting advertising preferences from EDA and facial expression data; (5) multimodal integration of eye-tracking, EEG, and GSR provides richer consumer behavior profiling than any single method alone. Heterogeneity (I² estimated moderate-to-high across subgroups) reflects genuine variation in stimuli, populations, and measurement tools. Conclusions: Eye-tracking and physiological signal methodologies offer convergent and complementary evidence for how consumers process, respond to, and decide about digital advertising stimuli. The synthesis indicates that attention, emotional arousal, and cognitive effort are meaningfully associated with advertising effectiveness and consumer choice. Future research should prioritize methodological standardization, increased reporting transparency, and cross-cultural validation to strengthen the evidence base for biometric-informed advertising research.