Evoked Electroencephalographic Responses to Naturalistic Speech: A Review of Analytic Methods, Current Insights, and Future Directions

Abstract The use of natural and naturalistic speech, including continuous narratives, is increasingly popular among researchers who use EEG to study speech and language processing. Despite bridging the gap between laboratory studies and real-life situations, its continuous, multidimensional, and overlapping nature poses several difficulties to the classic ERP analysis. This review provides a conceptual overview of more recent analytic methods developed to overcome those challenges. Subsequently, we summarize critical insights from these lines of research, extending from relationships between EEG and single-stimulus features to interdependencies among variables. These findings are also compared with those from the classic ERP literature. Finally, we discuss some unanswered questions rooted in the larger language and speech science literature, calling for tighter integration with existing theories of representation and processing. Collectively, the naturalistic paradigm offers a promising framework to generate and test hypotheses in cognitive neuroscience of language and speech.