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

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

Latent dimensions in neural representations predict choice context effects

Asaf Madar¹, Tom Zemer, Ido Tavor, Dino Levy; ¹Tel Aviv University

Presenter: Asaf Madar

Human choice is often affected by the context of available alternatives, a phenomenon known as choice context effects. To explain context effects, current models require the choice options to be described by two numerical attributes. However, decision-makers are not restricted by these attributes and might represent the options by additional latent attributes. Here, we propose using participants’ neural representations to gain access to the full attribute set they consider, while relaxing the assumptions regarding their attribute space. We aimed to use these representations to predict the context effects in participants’ choices. We estimated the context effects elicited by lottery stimuli using one behavioral sample (n=122) and then recruited two independent fMRI samples in a preregistered design (n_first=28,n_replication=34) to estimate the neural representations of each lottery without the context of choice. We predicted the context effects based only on the neural similarity between the individual lotteries, improving out-of-sample predictions by 14% and explained variance by 20% compared to traditional methods. This framework can be generalized to any stimulus type and help extend the study of context effects to more naturalistic stimuli.

Topic Area: Reward, Value & Social Decision Making

Extended Abstract: Full Text PDF