Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session A: Tuesday, August 12, 1:30 – 4:30 pm, de Brug & E‑Hall

Object Position, But Not Identity, Is Decodable During Object Permanence with MEG

Saskia Fohs1, Katharina Dobs1, Apurva Ratan Murty2; 1Justus Liebig Universität Gießen, 2Georgia Institute of Technology

Presenter: Apurva Ratan Murty

When objects are dynamically occluded, we maintain knowledge of their existence—a phenomenon known as object permanence. Despite extensive research on object permanence, the nature of neural object representations under occlusion remains unclear. Specifically, does the neural code maintain perceptual-rich features, or does it shift toward a sparse, conceptual format? Here, we measured dynamic neural representations using MEG in an ecological valid setting. Participants viewed one of five objects (e.g. bike, chair) moving either unobstructed or behind a wall, with the knowledge that the object would either disappear or reappear, while performing a speed-change detection task. Using MVPA decoding, we were able to decode object identity when the object was unoccluded. Throughout occlusion, however, we could decode object positions but not object identity. Notably, when the object was predicted to disappear during occlusion, the neural position signal was traceable only until the moment of vanishing. Our findings suggest that during occlusion, neural signals may reflect simulated spatial information rather than detailed perceptual features, thereby placing important constraints on computational models of object perception in the brain.

Topic Area: Object Recognition & Visual Attention

Extended Abstract: Full Text PDF