Nicolas Goupil and Dora Kampis
Affiliation: Center for Småbørns Kognition - Institut for Psykologi of Copenhagen
Category: Psychology
Keywords: infancy, episodic memory, relational memory, binding, development
Date: Thursday 4th of September
Time: 16:00
Location: GSSR Plenary Hall (268)
View the full session: Memory
Young infants learn very fast, prompted by attentional adaptations to learn from most significant stimuli around them, such as people and animate beings. In the first year, they extract generic rules or structures of the world, like object categories, statistical regularities, naïve physics. Memory, in that process, stores decontextualized representations, stripped from the anecdotic variations of events. But memory, as we know it as adults, can also encode contextually rich events, binding objects, people and their relationships often without noticeable causal or semantic link, in unique episodes. Episodic memory, a fortiori episodes about oneself, is characterized by such specificity, binding objects or knowledge to precise contexts or sources (Olson & Newcombe, 2013).
Various components have been suggested to contribute to episodic memory, including conscious awareness of an episode to which one can mentally travel back in time to reconstruct the episode (Tulving, 2005), and brain structures that begin to substantially develop after 2 years of age (Newcombe et al., 2007). Consequently, episodic memory has been suggested to have a protracted development. However, it is meaningful to ask what the building blocks of this capacity are, some of which may be present earlier in ontogeny.
Flexible and rich episodic memories rarely show before three years of life, but children’s memory already shows specificity during their first two years. Around one year, infants remember objects at precise locations (Oakes et al., 2017), specific associations between object pairs (Johnson et al., 2020), or people and scenes (Richmond & Nelson, 2009). Over the second year, infants participate in increasingly rich social interactions, which confront them with the perspectives of other people. Driving their attention towards new aspects of the world, these perspectives may also conflict with their own views, bringing about the need to dissociate these views, and to assign each view to a specific person –that is, to track sources of knowledge. We hypothesize this new challenge accompanies transformations in memory, balancing the need for generalizable knowledge with the need for source specificity. We investigated that in a longitudinal study at 12, 18, 30 and 36 months (planned N=125).
The current presentation focuses on eye-tracking experiments measuring binding in 12 and 18-month-old infants’ memory: the ability to form unique and arbitrary associations between items or elements presented together, or items related to an aspect of the scene such as their location. Toddlers were tested on 1) items matching (vs. non-matching) their previous association with another item (item-item binding), and 2) faces matching (vs. non-matching) their previous scene context (face-scene binding). Tackling each age group’s memory capacity, two experiments additionally tested 3) 12-month-olds’ memory for colors matching (vs. non-matching) their previous spatial location (feature-location binding), 4) 18-month-olds’ memory for boxes containing (vs. non-containing) toys in two different rooms, this time during an interactive task. The contribution of these early binding mechanisms to social cognition will be investigated in additional experiments on infants’ ability to track social agents’ perspectives, not presented here. Across tasks, results at 12 months showed a consistent novelty effect for non-matching stimuli, with longer looking times on new item-item binding (N=120, p = .048), new scene-face binding (N=120, ps < .033), and new item-location binding (N=116, p < .001). Results at 18 months are still under collection until May 22nd of 2025 (N=64 at time of submission).
Still in progress, these works already show that, in different paradigms taxing distinct perceptual mechanisms, one-year-old infants spontaneously look longer at items that don’t match the spatial location, item or scene they were previously associated with, even though none of the target items themselves were new. That is, they kept track of arbitrary associations they only observed once. Highlighting changes in contextually rich environments, such a mechanism may allow infants to discriminate distinct episodes from each other, in order to build their first episodic memories. Results at 18 months will tell how this develops and supports memory for object location in interactive tasks. Further works on agent belief/preference binding will attempt to link early binding mechanisms to children’s developing specificity in associating beliefs or preferences to people.