par Capparini, Chiara 
;Fourdin, Lauréline ;Wens, Vincent ;De Tiege, Xavier ;Bertels, Julie
Référence International Congress of Infant Studies (ICIS) (Glasgow, UK)
Publication Non publié, 2024-07
;Fourdin, Lauréline ;Wens, Vincent ;De Tiege, Xavier ;Bertels, Julie Référence International Congress of Infant Studies (ICIS) (Glasgow, UK)
Publication Non publié, 2024-07
Communication à un colloque
| Résumé : | Visual statistical learning (VSL) refers to the ability of becoming sensitive to the visual regularities in our environment. This skill facilitates the detection of structure and emerges very early in development (Bulf et al., 2011; Kirkham et al., 2002). Thus far, developmental research has mostly investigated VSL with post-exposure behavioural tasks that only reveal the outcome of learning. Notably, behavioural tasks may lead to ambiguous interpretations since there is no clear consensus about the directionality of the expected learning outcome (i.e., novelty vs. familiarity effect). Steady-state visual evoked potentials (SSVEPs) can be acquired while learning occurs and can shed light onto the temporal course of learning. At present, SSVEPs investigations of the ongoing learning processes have been limited to the auditory domain (Choi et al., 2020). In this study, we used SSVEPs to investigate infants’ neural entrainment mechanisms in response to visual regularities. The final sample included 30 4- to 6-month-old infants (11 females, M age = 156.5 days). Participants were presented with 20s blocks of 8 colourful shapes appearing in the centre of the screen at a frequency of 6 Hz. These blocks were alternated with attention-getting animations to attract the infant attention to the screen. Participants were randomly assigned to one of three conditions: 1) the standard doublet condition, in which the 8 shapes were organised in 4 doublets, 2) the control doublet condition, in which 4 shapes could take the first position of the doublet and other 4 shapes the second position of the doublet, leading in turn to 16 possible doublets, and 3) the random condition, in which individual shapes were randomly presented. We compared the signal-to-noise ratio (SNR) at the base frequency of visual stimulation (6 Hz and its higher harmonics) and at the frequency of doublet presentation (3 Hz and its higher harmonics) across conditions. As long as participants were attending to the visual stimuli, we hypothesised a peak at the base frequency of visual stimulation across occipital channels in all conditions. At the same time, if the sequence included a regular organization in doublets, we hypothesised a progressive response at 3 Hz. Linear mixed-effects models revealed that the SNR at the base frequency was only explained by trial order, with a decreased SNR over time. Hence, neural entrainment at the base frequency did not differ across conditions. This confirmed that infants attended similarly to the visual stream of stimuli, independently of the condition. On the other side, activity at the doublet frequency was only explained by condition. In particular, the SNR at the doublet frequency and its harmonics was higher in the standard doublet condition compared to the random condition. This showed that the infant brain is sensitive to a regular doublet organization, especially in the standard doublet condition. Overall, these findings provide an early neural signature of VSL abilities and are crucial to better understand learning mechanisms as they unfold during stimulus exposure. |
