Roux, F. 1 , Frost, R. 1, 2 & Carreiras, M. 1, 3, 4
1 BCBL. Basque Center on Cognition, Brain and Language, San Sebastian, Spain
2 Department of Psychology, The Hebrew University Jerusalem, Israel
3 Ikerbasque, Basque Foundation for Science, Bilbao, Spain
4 UPV/EHU, Universidad del Pais Vasco, Spain
Many complex behaviours involve the ability to learn and reproduce sequences of events in their correct temporal order. Some of these sequences can occur embedded in continuous streams of information. Therefore, we often extract sequences from background information without any prior knowledge about the timing of sequence on- and offsets. While recent theoretical and empirical work indicate that the sequential structure of inputs may be inferred from perceived transitional probabilities, the ability to exploit such statistical information seems to be highly variable. A mechanistic account of what is causing such differences in behaviour, however, is still missing. Here we address how the brain extracts the statistical regularities embedded in temporal sequences and whether inter-individual differences in sequence learning can be predicted from fluctuations of brain activity. Building on previous electrophysiological studies, we hypothesised that rhythmic brain activity supports the representation of behaviourally relevant sequences in neuronal activity as well as their consolidation in memory. Our findings suggest that transitional probabilities embedded in temporal sequences are tracked by oscillatory activity in multiple frequency bands (1-50Hz). Finally, we observed that individual differences in sequence learning can be predicted with high accuracy from delta (1-4Hz) coupled beta (15-20Hz) oscillations.