Seen Together, Activating Together: How object co-occurrence statistics shape object representation and visual cognition

Description du projet

Zoom sur la vision et le comportement humains

Le décryptage des objets visuels ne se résume pas à une simple coup d’œil. Si le rôle de la taxonomie est clair, l’impact de la cooccurrence des objets demeure difficile à cerner. Cette dualité de la sémantique complique notre compréhension de la cognition visuelle. Soutenu par le programme Actions Marie Skłodowska-Curie (MSCA), le projet STAT se propose de faire la lumière sur la manière dont la cooccurrence des objets influence notre compréhension visuelle. Les chercheurs cartographieront, à l’aide de l’IRMf et de l’EEG, l’activité cérébrale afin de découvrir comment la cooccurrence façonne la représentation des objets et la cognition. Les quatre objectifs du projet vont de l’identification des représentations neuronales à la compréhension de leur impact sur la perception consciente et les tâches quotidiennes telles que la recherche visuelle. En décodant la façon dont les objets interagissent dans notre environnement, STAT pourrait redéfinir notre compréhension de la vision et du comportement humains.

Objectif

Visual objects in real life carry both perceptual and semantic information, which cannot readily be inferred from simple visual features and is based on prior knowledge. However, semantics is not a unitary concept; it can be separated into taxonomy (e.g. ducks and frogs are both animals) and co-occurrence (e.g. candles appear together with cakes). Previous studies have extensively characterized the role of taxonomy in visual cognition, while co-occurrence is less well-understood and more disputed. The aim of this proposal is to investigate the impact of object co-occurrence on object representation and visual cognition. The project has four objectives: First, I will use fMRI and EEG to identify where and when object co-occurrence statistics are represented in the brain, how they rival other competing organizations, and which mechanisms enable these representations. Objective 1 will provide a characterization of co-occurrence as a key organizational principle of object representation. Second, I will investigate how co-occurrence modulates neural competition between objects using fMRI. Objective 2 will clarify whether organizing objects based on co-occurrence can overcome information processing bottlenecks in the visual brain. Third, I will examine whether co-occurrence can facilitate conscious perception using the attentional blink paradigm and explore where and when neural representations of co-occurrence predict AB magnitudes. Fourth, I will investigate how and when co-occurrence impacts performance and neural processing in visual search using EEG. Objectives 3 and 4 will aid our understanding of how object co-occurrence contributes to human perception and cognition and thereby helps people to perform everyday tasks successfully. Overall, this research will provide a comprehensive characterization of how object co-occurrence in the world shapes human vision and open new avenues for explaining visual behaviors in natural environments.