Summary
Highlights
The video begins by contrasting two hypothetical paintings with identical characteristics but vastly different valuations: one a masterpiece, the other forgotten. It questions why some art becomes highly valuable and eternal while other similar art is discarded, suggesting a deeper reason than pure luck or patronage. It proposes that if art is subjective, the widespread agreement on masterpieces implies a 'formula' for memorability.
The speaker introduces the idea that the secret to memorable art lies within the brain, not the art itself. This concept is explored through the Mandela Effect, where many people share identical false memories of cultural icons like the Monopoly Man, C-3PO, and Waldo. Psychologist Wilma Bainbridge suggests these aren't parallel universes but predictable errors in human memory, coining the term 'visual Mandela effect'.
Wilma Bainbridge's research involved testing false memories of altered cultural icons, like the Fruit of the Loom logo (which many falsely remember with a cornucopia). Her findings indicated that the brain has a propensity to remember certain images, with the 'wrong' versions often being more 'sticky' not due to misinformation or association, but an inherent quality. Further studies with thousands of ordinary faces showed that some faces are universally remembered while others are universally forgotten, irrespective of attractiveness or emotion, suggesting an 'X factor' or hidden visual code for memorability.
Bainbridge extended her research to art by conducting an experiment at the Art Institute of Chicago. Volunteers explored a gallery and were later quizzed on the paintings they saw. Surprisingly, despite individual preferences, some artworks were consistently remembered by everyone, and others forgotten. Factors like painting size and context (modern, spacious vs. cluttered, traditional galleries) influenced memorability, but beauty, emotion, or colorfulness did not predict stickiness. This indicated a 'mysterious arrangement of pieces' that appealed to the brain.
The video then introduces the ResMem AI model, which was trained to predict image memorability by learning human patterns of remembering and forgetting. This AI, unfamiliar with art history, was fed paintings from the Chicago Art Institute study. Remarkably, ResMem predicted that paintings humans consider masterpieces were more memorable than obscure works. This suggests that masterpieces like 'American Gothic' hack the human brain's visual processing, creating a 'cognitively sticky' image. This challenges romantic notions of art, implying a formula for visual memorability.
The video concludes by highlighting a profound application of this research beyond art: its use in detecting early cognitive decline. People with early Alzheimer's disease diverge from healthy individuals in remembering specific images, which could serve as a warning sign. This work underscores that our memories are not mere recordings but are curated by our brains based on rules we are only beginning to understand, whether it's a false memory or a famous painting. The ResMem tool is made public for users to test their own creations' memorability.