Artificial Intelligence and Neuroscience in Support of Economic and Managerial Decision-Making: A Systematic Review of Machine Learning-Based EEG/fNIRS Approaches

Abstract

This systematic review examines how artificial intelligence (AI), when applied to neurophysiological signals – primarily electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) – can support the prediction and interpretation of economic and managerial decisions such as preference, purchase intention, willingness to pay, credit decisions, and advertising effectiveness. It addresses a specific gap in the literature: previous reviews have mainly focused on EEG-based neuromarketing or on signal-analysis techniques, whereas few studies jointly examine machine-learning models, EEG/fNIRS measures, economic and financial applications, external validity, and ethical governance. Following PRISMA 2020, a structured English-language search was conducted for publications issued over the last ten years across PubMed Central, Frontiers, ScienceDirect, Springer Nature, and the ACM Digital Library. Out of 28 records identified, 6 duplicates were removed, 22 titles and abstracts were screened, 14 full texts were assessed, and 12 empirical studies were ultimately included. The synthesis shows a recurrent predictive gain when EEG signals complement self-reports or behavioral measures, but it also highlights substantial heterogeneity in reported performance, depending on task type, sample size, metrics, and validation protocols. The review further underlines the emerging role of interpretable approaches, especially SHAP-based fNIRS/finance analyses and EEG regional analyses linking predictions to prefrontal markers. Beyond descriptive synthesis, the article proposes an integrative “Neuro-AI-Decision” framework and outlines a research agenda centered on external validity, reproducibility, and the ethical governance of neural data.

JEL Classification : C45, D91, G41, M31

Paper type: Theoretical research (systematic literature review)