Fig. 7
From: An fMRI dataset in response to large-scale short natural dynamic facial expression videos
Cortical predictability based on encoding models. (a) Model-predicted accuracy in cortical responses to facial expression video stimuli is assessed by measuring the Pearson correlation between the measured and predicted responses during the testing video. (b) The predicted accuracy of encoding models for different ROIs using different layers of CNNs. The accuracy of predictions for each ROI was calculated by averaging across the top 100 vertices within the ROI and across participants. The error bars in the graph represent the standard error, while the curves stand for the average performance.
