HAL will be down for maintenance from Friday, June 10 at 4pm through Monday, June 13 at 9am. More information
Skip to Main content Skip to Navigation

Probabilités transitionnelles de second-ordre dans l’apprentissage statistique de triplets

Abstract : Sequential processing is a fundamental cognitive mechanism allowing extracting regularities to facilitate learning. The literature in cognitive psychology shows that the extraction of these regularities can operate through statistical learning. Most of the previous work has focused on first-order transitional probability (TP), which corresponds to the probability of a stimulus based on one previous stimulus. Nevertheless, within triplets, second-order TPs (probability of a stimulus based on two previous ones) must be taken into account, both for its specific effects and in interaction with other factors involved in statistical learning. The objective of this thesis was to determine the effects of three factors that may underlie sequence learning. A serie of experiments in learning aimed at manipulating within a set of sequences : i the first-order TPs (one stimulus predicting another), as well as the second-order TPs (a combination of stimuli predicting another stimulus), ii the structure of dependency characterizing the TPs (adjacent vs. non-adjacent) and, iii the frequency of the sequences. Our experiments have highlighted a present but difficult learning of second-order TPs within visuomotor and language sequences to the benefit of a prevalence of first-order TPs between adjacent elements. Our results also showed complex interactions between first-order TPs and second-order TPs. The discussion focuses on the adaptive capacity of participants to preferentially use some factors depending on the type of task, in this case, TP for the prediction of a stimulus and the use of frequency for the memorization of sequences. A last study suggested an opening relative to the ubiquity of sequential processing, underlining the preponderant role of sequential processing even in cognitive tasks that are non-sequential in nature.
Document type :
Complete list of metadata

Contributor : Abes Star :  Contact
Submitted on : Tuesday, May 10, 2022 - 11:24:21 AM
Last modification on : Wednesday, May 11, 2022 - 3:47:57 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03663516, version 1



Laura Lazartigues. Probabilités transitionnelles de second-ordre dans l’apprentissage statistique de triplets. Psychologie. Université Côte d'Azur, 2022. Français. ⟨NNT : 2022COAZ2003⟩. ⟨tel-03663516⟩



Record views


Files downloads