We propose to combine three technologies to predict what makes an image memorable or forgettable: neuro-imaging technologies recording where encoding happens in the human brain (spatial scale), when it happens (temporal scale), and what types of computation are performed at the different stages of storage (computational scale.