Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams
Sleep components involved in the consolidation of new vocabulary and morphological regularities
Poster A37 in Poster Session A, Thursday, October 6, 10:15 am - 12:00 pm EDT, Millennium Hall
Eva Kimel1,3, Dafna Ben Zion1, Anat Prior1, Gareth Gaskell3, Ilana S Hairston1,2, Tali Bitan1,4; 1University of Haifa, 2Tel Hai Academic College, 3University of York, 4University of Toronto
It was previously shown that sleep plays an active role in the consolidation of newly learned linguistic information. However, a more accurate mapping of consolidation for various learning types to specific underlying sleep properties, and its relation to the temporal dynamics of consolidation, is still under investigation. In the current study we are assessing the involvement of sleep in learning of novel vocabulary, and novel plural inflections based on implicit morpho-phonological regularities. We consider the temporal dynamics of these learning types and test their association with sleep spindles, which were shown to be connected to overnight consolidation. Participants were trained in the evening on the inflection of 36 novel words, in which morpho-phonological regularities were embedded, and were presented either frequently or infrequently during training. Training was followed by an immediate testing of acquisition and generalization, a night in a sleep lab with a polysomnography, a test in the morning, and additional testing of offline consolidation – 36 hours post-training and one-week post-training. Preliminary data analysis (N=29) indicated a high variability in performance change after sleep, allowing to test associations with sleep more effectively. Overall, memory for novel vocabulary improved over the first night after training, and declined to the initial post-training level throughout the following week. Consistent with previous studies, the dynamics was different for frequent and infrequent items. Both the improvement and the deterioration were significant for the infrequent items (immediate vs. morning: t = 2.5, p = 0.017; 36 hrs. post-training vs. week after: t = 3.1, p = 0.004) but not for the frequent items, the performance for which did not significantly change throughout the 4 tests. Both the absolute score in the morning and the overnight improvement for infrequent items were associated with sleep spindle density during slow-wave sleep (SWS; correlation with morning score: Spearman’s rho 0.606, p = 0.001; correlation with overnight improvement: 0.602, p = 0.001). Interestingly, while performance one week post training was not correlated with sleep spindles during SWS it was strongly associated with sleep spindles during the N2 sleep stage (Spearman’s rho: 0.583, p = 0.001), in agreement with previous suggestions on a different role of spindles during N2 sleep vs. spindles during SWS. We did not find a strong association of sleep spindles with generalization, which relies on regularity learning in this task. To conclude, our results show that distinct sleep components are associated with learning novel items and learning a regularity. Specifically, the temporal dynamics of learning vocabulary vs. learning morphological regularities seem to be different, and accordingly their dependence on sleep differs as well. Spindles occurring during SWS are associated with item learning (vocabulary) but not with regularity acquisition (morphological inflections), and are not associated with the efficacy of the learning process before sleep, but are strongly associated with the overnight change in performance. Spindles during N2 sleep might be associated with long-term retention of novel items.
Topic Areas: Morphology, Computational Approaches