Fig. 3: Model-based predictions compared to novice participants and experts in the field.
From: Computational models of category-selective brain regions enable high-throughput tests of selectivity
a Schematic of the comparison between neural network models for each fROI (left) and human participants who were either Professors with several published papers on the human ventral visual cortex (experts) or participants from a crowd-sourced experimental platform (novices), and b Performance of the ANN-based encoding models (left column), experts (middle column), and novice participants (right column) at predicting the observed responses to stimuli. The connected lines indicate each fROI. For the behavioral data, each dot indicates the mean prediction and the errorbars indicate the s.e.m. across participants. Source data are provided as a Source Data file. c Comparison between ANN models (A), experts (E), and novice (N) participants at predicting the observed responses at the grain of individual stimuli within each of the four categories of stimuli used in the experiment for the FFA (left column), the EBA (middle column), and the PPA (right column). For the behavioral data, each dot indicates the mean and the errorbars indicate the s.e.m. across participants. The line and shaded regions indicate the ceiling for how good the models can be, based on the reliability of the data. Source data are provided as a Source Data file.
