The Brain Learns by Making Neurons More Alike express
A chef's knife and a peeler each handle different tasks, but when a cook learns a new recipe, they often grab the same knife for everything. The brain seems to do the same when mastering a new skill of distinguishing patterns.
In the experiment, monkeys trained for weeks to tell apart barely noticeable differences in complex patterns. During this time, scientists recorded the activity of hundreds of cells in the visual area V4 — a region responsible for recognizing shapes and colors. Early in training, neurons responded to patterns differently, like musicians each playing their own part. But by the end of training, they increasingly began to 'sing in unison': their responses to the same images became similar.
Moreover, each neuron individually began to carry more information about what was shown on the screen. It's like a choir: when everyone sings the same melody, each voice sounds louder and clearer, even if the parts are identical. The brain wasn't saving energy; instead, it was reinforcing confidence in the answer by repeating it many times.
This result supports the idea that the brain doesn't just filter incoming signals but builds an internal model of the world, like a chef who no longer looks at the recipe but 'feels' the dish. And when a decision is needed, it relies on this model rather than raw data.
Back to the kitchen: an experienced chef doesn't sort through all the knives but grabs one right away, because they know it will do the job. Similarly, after learning, the brain doesn't search for the most informative neuron but relies on a whole choir of cells that together give a confident answer.
Original abstract
as written by the authors arXiv:2603.07369
How does the brain optimize sensory information for decision-making in new tasks? One hypothesis suggests learning reduces redundancy in neural representations to improve efficiency, while another, based on Bayesian inference, predicts learning increases redundancy by distributing information across neurons. We tested these hypotheses by tracking population responses in macaque cortical area V4 as monkeys learned visual discrimination tasks. We found strong support for the Bayesian predictions: task learning increased redundancy in neural responses over weeks of training and within single trials. This redundancy did not reduce information but instead increased the information carried by individual neurons. These insights suggest sensory processing in the brain reflects a generative rather than discriminative inference process.
The V4 visual area in macaques is located in the occipital part of the brain and is involved in recognizing even such subtle differences as the distinction between very similar shades of color.