OS_42.2 - Comparing measures of consciousness in an artificial grammar learning task

Wierzchon, M. 1, 2 , Asanowicz, D. 1 & Cleeremans, A. 2

1 Institute of Psychology, Jagiellonian University, Krakow, Poland
2 Consciousness, Cognition & Computation Group, Université Libre de Bruxelles, Belgium

Consciousness can be measured in different ways, and different measures unfortunately often yield different conclusions about the extent to which awareness relates to performance. The challenge of correctly identifying which measure is best is thus substantial. Here, we compare five different subjective measures of rule awareness in the context of an artificial grammar learning task. Participants (N=217) had to express their rule awareness by means of one of five different scales: confidence rating (CR), post-decision wagering (PDW), rule awareness (RAS, a modified PAS scale), the Sergent-Dehaene continuous scale (SDS), and feeling of warmth (FOW, a new measure). All scales were found equally sensitive to conscious knowledge, but PDW and SDS are affected by risk aversion (suggesting that CR, RAS, and our new scale FOW should be preferred). We observed that CR captures the largest range of states of consciousness (yielding the largest difference in accuracy between the highest and lowest scale points), but also that only CR fails to indicate unconscious knowledge by means of the guessing criterion (chance performance when guessing). CR’s unique features suggest that it may be used in conjunction with RAS or FOW to enable finer assessment of subjective states of awareness.