Decoding the Future:Single Trial Analysis of Anticipatory Responses in Prestimulus EEG Activity

Can people ‘feel the future’? According to one of the most controversial papers published in the past decade, Daryl Bem’s ‘Feeling the Future’ (Bem, 2011), participants are in some circumstances able to anticipate on future events. Bem’s paper, though, is only one example of a considerable series of experiments claiming anomalous anticipatory responses to future events, in behaviour, but also in physiological responses. In a recent study using single-trial EEG decoding, we observed a similar effect: we were able to decode the identity of an upcoming stimulus based on unfiltered baseline activity alone. Subsequent analysis, however, revealed that this effect was most likely due to improper randomization. This led us to hypothesize that the anomalous baseline effects reported by others may be due to weak implicit learning in insufficiently randomized trial sequences. In the present study, we contrast pseudo-randomized (randomized with the Mersenne Twister algorithm) versus truly randomized (randomized with a Zener-diode hardware RNG and a Geiger-counter based RNG) sequences. Contrary to our hypothesis, we were able to predict stimulus identity significantly above chance level on basis of single trial prestimulus activity for both pseudo- and truly randomized sequences, despite our best efforts to exclude alternative explanations, such as filter artefacts or participant expectation. Although a direct replication of our study is warranted, our results cannot confirm our hypothesis that anomalous anticipatory activity is the result of improper randomization. In stead, participants seemed to able to ‘feel the future’ even for truly random events.