We used to think the brain worked like a precision machine.
One cell for one job. A “place cell” for your location. A “Jennifer Aniston cell” for her face. It made sense. It was clean. It was elegant.
It was also mostly wrong.
Stefano Fusi, a researcher at Columbia University, has been looking at the data. What he sees isn’t a collection of hyper-specialized experts. It’s a crowd of generalists. “These [earlier] neurons are really the.exception, not the rule,” he says.
The abundant cells? The ones we ignored because they weren’t flashy enough? They might actually be doing the heavy lifting.
The myth of the single-purpose neuron
Back in 2013, Fusi spotted something odd in monkey brains. He found neurons with “mixed selectivity.” These cells didn’t just care about one thing. They fired when multiple inputs hit them at once. Task-related? Yes. Location-specific? No. They cared about a mix.
But one monkey study doesn’t settle the debate. You need scale. You need to know if this is a quirk or the norm.
So Fusi turned to the International Brain Laboratory. They have recordings from over 100 mice. These weren’t just sitting there. The mice were working. They saw an image on the left or right. They had to move it to the center. They did it by turning a wheel.
Simple task? Maybe. But the neural activity told a different story.
How mixed selectivity dominates the cortex
The researchers mapped the activity. They looked for the specialists.
They found them. But only in specific pockets. The primary visual cortex, for instance, held onto its special status. Those cells cared about vision. That was their job.
Everywhere else? Chaos. Or rather, complexity.
Neurons outside the visual cortex didn’t just process the image. They cared about where it was on the screen. They cared about the decision to move left or right. They cared about the physical act of turning the wheel. One cell. Many jobs.
The important thing to stress is that there are specialized neurons [such as place cells]. But these are not the norm.
It wasn’t just about the task. It was about how different neurons responded to the same stimulus. Take the somatomotor areas. These regions plan movement. Some cells fired wildly when the mice licked or moved their whiskers. Others? They barely budged.
Individual neurons aren’t myopically focused. They aren’t tunneling onto one bit of info. They’re casting a wide net.
Why generalists might be better
Ben Hayden from Baylor College of Medicine calls this a “giant smack in the face” to the old theory.
The problem with searching for specialists is the endless search. You miss a cell? You say, “Maybe I just need to record more.” Hayden points out the scale of this new study. They recorded 14,000 neurons.
Still, mostly generalists.
This changes how we view brain capacity. If every neuron only did one thing, you’d need infinite cells to handle every possible combination of inputs. Generalists solve that. By responding to multiple inputs in slightly different ways, they maximize the information the brain can process.
This likely covers memory, cognition, decision-making, and perception. Not separate departments. A mesh.
Where do we go from here?
The mouse task was simple. A real world problem is multidimensional. Fusi admits this. He wants to see when neurons decide to focus versus when they spread out.
The next step isn’t mice. It’s primates. It’s humans.
His team is collaborating with researchers studying epilepsy patients. These patients are conscious during open-brain surgery. It’s a rare window. They can record human neurons in real-time.
Fusi wants to shift the lens entirely. Stop looking at one cell at a time.
It’s hard. It’s messy. But it’s necessary. One neuron doesn’t talk to one other neuron. Each neuron connects to roughly 10,000 others on its dendritic tree. It sees the whole network.
So we should too.
