From screens to live interactions: Considerations for naturalistic EEG paradigms in social and affective neuroscience.

Live, naturalistic, interactive electroencephalographic (EEG) paradigms have increased over the last 20 years, reflecting a movement towards ecologically valid measurement of the actual stimuli and responses theorized to support human brain and cognitive development. A challenge to this growing field is that many analytic decisions are at researchers' discretion, and there is a lack of clear standards for reporting and justifying decision-making processes, and a lack of understanding how different analytic decisions may affect results. These circumstances threaten study replicability, transparency, and accuracy. Thus, necessary objectives are to bring awareness to this existing analytic variability, empirically evaluate its effects, and ultimately develop guidelines for optimally applying different analytic approaches. The present study takes an important first step toward these objectives. We review studies using live, interactive EEG paradigms published between 2006 and 2026, and identify several analytic approaches that vary widely: 1) analyzing EEG across longer conditions versus shorter, discrete events, 2) considering influence of amount of usable EEG, 3) inclusion of covariates, 4) whether to examine all electrodes across the scalp versus targeted electrodes, and 5) determining appropriate baselines for comparison to target EEG activity. We discuss considerations for these different analytic approaches, and evaluate their effects in 4- to 6-month-old (N = 35) infant EEG data from a novel, live, naturalistic mother-infant interactive task. Results reveal that different, often arbitrarily selected analytic approaches can change EEG results. We recommend standards and best practices in reporting analytic decisions to increase transparency and replicability, and suggest strategies for determining optimal analytic approaches.