Eye-Tracking and EEG in Neuro-Tourism: A Systematic Review of Machine Learning Approaches for Visual Attention and Cognitive Load Modeling

Background Neuro-tourism represents an emerging interdisciplinary field applying neuroscientific measurement techniques to understand tourist behavior through objective physiological data. Despite growing interest in eye-tracking and electroencephalography (EEG) applications in tourism contexts, the integration of machine learning approaches for analyzing multimodal neurophysiological data remains fragmented and methodologically inconsistent. This systematic review examines how machine learning algorithms can optimize the analysis of combined eye-tracking and EEG data to understand visual attention patterns and cognitive load in tourist experiences. Methods Following PRISMA guidelines, we systematically searched seven academic databases for peer-reviewed studies published between 2020 and 2025. Inclusion criteria required empirical studies employing both eye-tracking and EEG measurements in tourism contexts with machine learning analysis. From initial database searches, 56 studies met all inclusion criteria and underwent qualitative synthesis organized around four research questions addressing algorithm optimization, feature integration, experiential correlations, and predictive performance across tourist segments. Results Traditional statistical methods dominated the majority of studies, while machine learning implementations demonstrated consistently superior performance across all evaluation metrics. Multimodal approaches combining eye-tracking and EEG data showed substantial accuracy improvements over single-modality analyses. Deep learning architectures, particularly hybrid models, achieved the highest predictive performance. Fixation duration emerged as the primary eye-tracking metric, while frontal brain asymmetry best indicated emotional responses. However, ecological validity remained limited, with most studies conducted in laboratory settings rather than authentic tourism environments. Cross-cultural validity and sample diversity presented significant limitations. Conclusions Machine learning approaches have fundamentally transformed neuro-tourism research capabilities, enabling objective measurement beyond traditional self-report methods. However, significant barriers including technical expertise requirements, limited real-world applicability, and minimal ethical framework development constrain practical implementation. Future advancement requires establishing standardized protocols, interdisciplinary collaboration, and culturally sensitive research designs that balance technological sophistication with ethical responsibility.