The macaque IT cortex but not current artificial vision networks encode object position in perceptually aligned coordinates.

Motion adaptation causes perceptually aligned shifts in object position coding in macaque IT cortex, but not in current artificial vision networks.

• Why it matters: Understanding how the brain encodes object position in a way that aligns with perception is crucial for unraveling visual processing and improving artificial systems, yet artificial networks lack this perceptual alignment.
• What they did: Researchers combined intracortical recordings in macaque IT cortex with human psychophysics, using a visual illusion to dissociate perceived object position from retinal location, and tested various neural network models.
• The result: Motion adaptation induced biases in IT population codes that mirrored human perception, while artificial networks did not show such shifts; applying IT-based transformations to models simulated these biases, highlighting a key missing element in artificial vision.