Fig. 5
From: A large-scale fMRI dataset for the visual processing of naturalistic scenes
The fMRI data from ImageNet and COCO experiments show higher response reliability in visual areas involved in object recognition. (a) The group mean maps of test-retest reliability of BOLD response across 1000 ImageNet categories. The test-retest reliability was first computed with each pair of sessions and then averaged across sessions and participants (N = 9). (b) The group mean maps of test-retest reliability of BOLD across 120 COCO images. The test-retest reliability was first calculated between odd runs and even runs within a session (i.e., participant) and averaged across participants (N = 9).
