We often call it the mind’s eye: the ability to picture a face, navigate using an internal map, or conjure up familiar objects with your eyes closed. Varun Wadia, a neuroscientist at Cedars‑Sinai Medical Center and Caltech, and colleagues report in Science that the neurons that respond when we see objects also respond when we imagine them, and they use the same neural code to do so.
The team recorded from more than 700 individual neurons in the ventral temporal cortex—a brain region central to object recognition—by working with 16 people with epilepsy who already had electrodes implanted for clinical reasons. Participants first viewed a range of images on a screen (faces, animals, plants, words and small objects such as sunglasses or water bottles) while researchers measured which neurons fired and how strongly. That pattern of responses allowed the researchers to map the neural ‘‘code’’ that represents different visual items.
In a second phase, the same participants closed their eyes and imagined objects they had just seen while the researchers continued to monitor the same neurons. About 40% of the neurons that responded to an object during perception reactivated during imagination, and they did so with roughly the same intensity. The overlap was consistent enough that the team could decode which specific object a person was imagining—identifying, for example, an airplane—and even infer attributes such as size, viewpoint and whether the object was indoors or outdoors.
Kalanit Grill‑Spector, a psychology professor at Stanford who was not involved in the work, says single‑neuron evidence of shared coding is new and could help improve computational models of vision and inform treatments for visual disorders like macular degeneration. Thomas Naselaris, a neuroscientist at the University of Minnesota, notes the finding clarifies how imagination fills in missing views or combines familiar parts into novel mental images (think of mentally attaching a unicorn’s horn to a horse).
This cellular‑level result complements earlier brain‑imaging studies that showed overlapping circuits for perception and imagery but could not resolve individual neurons. It also builds on prior research into how visual systems represent faces and objects.
The study does not solve all puzzles about imagery. Conditions such as aphantasia—where people report little or no voluntary visual imagery—remain unexplained. Lead experimenter Itzhak Rutishauser recounts a scientist telling him, “I don’t see anything when I close my eyes.” People with aphantasia may rely more on verbal or conceptual representations than on visual ones, and understanding their strategies will require direct neuronal investigations similar to this study.
Overall, the results provide some of the strongest evidence to date that imagining and seeing travel the same neural path, reactivating the same populations of neurons in the service of internal visualization.