At 9 a.m. on a Thursday in July 2021, Martin Picard sat on a twin bed and watched his own blood flow from an IV in his arm through a hole in the wall.
Every hour for twenty-four hours, including while he slept, a nurse channeled that blood to a research team next door. When awake, he handed over saliva and filled out a survey about his mood.
The room was a metabolic research chamber, one of only fifty of its kind in the world. It stood less than a shoulder's width from a stainless steel sink and a porcelain toilet, and it was deliberately that small: to keep Picard from burning any energy beyond the bare minimum needed to keep him alive.
Napping was forbidden. Meals were tailored to his caloric needs and served on a schedule. Bedtime was 11 p.m. sharp.
He was the first volunteer in an experiment run by his own lab, the Mitochondrial Psychobiology Lab at Columbia. The question was almost embarrassingly basic: what does it cost, in energy, simply to exist?
Picard suspects the answer to a much larger question has been hiding inside the organelle everyone was taught to ignore. "If the energy stops flowing," he says, "there's no more you. There's no more consciousness, there's no more emotion, there's no more life."
For a century, this was the wrong place to look. Neurons were the stars of neuroscience — the spidery, electrically excitable cells that fire and wire and think. Mitochondria were the dull supporting cast, introduced in middle-school biology as the "powerhouses of the cell" and promptly forgotten.
The neuroscientist Hannah Critchlow remembers being taught, two decades ago, that mitochondria were "static 'powerhouses' of the cell, dutiful but conceptually dull." That picture is now coming apart.
The reason it is coming apart is a set of numbers that refuse to sit quietly.
The brain is 2 percent of body mass and consumes roughly 20 percent of the body's resting energy. Every perception, memory and idea is metabolically expensive; thought is, quite literally, the most costly thing the body does.
Zoom in and the bill gets stranger. Within neurons, maintaining the resting membrane potential takes about 15 percent of the energy, firing action potentials about 16 percent — and synaptic transmission alone accounts for roughly 44 percent, split across both sides of the synapse.
In other words, the majority of a neuron's energy is spent not on existing but on communicating. The synapse — the very junction the wiring paradigm treats as the seat of cognition — is where the fuel disappears fastest.
And it is mitochondria that supply it. They provide about 95 percent of the energy the nervous system needs, delivered precisely when and where the demand spikes.
Mitochondria provide about 95 percent of the energy the nervous system needs, delivered precisely when and where the demand spikes.
This is why the reversal is more than a rebranding. Critchlow puts the shift plainly: intelligence depends "not simply on the brain's wiring, but on the uninterrupted flow of energy through that wiring."
Weight for weight, she notes, our brains are more mitochondrial than neural. Each neuron carries thousands of these organelles, filling up to 40 percent of its volume.
Map the connectome all you like. You will have charted the roads without knowing whether there is any fuel in the tanks.
The strongest evidence that this matters comes from watching the machinery in living brains. At Imperial College London, Oliver Howes and colleagues used PET scanning to measure the availability of mitochondrial complex I — the largest enzyme at the very start of the chain that produces ATP — in the brains of healthy volunteers.
The finding was blunt: higher IQ scores correlated with greater complex I availability. More capacity to generate energy, more capacity for complex thought.
The animal work converges. At Mount Sinai, Yuko Hara's team examined synapses in the prefrontal cortex of rhesus monkeys and found that working memory tracked the shape and density of mitochondria there.
Strong performers had more structurally healthy mitochondria. Poor performers showed malformed ones — the signature of oxidative stress and reduced ATP output.
Notice what this implies. If intelligence is throttled by the flow of ATP through the synapse, then cognition has a metabolic ceiling — a hard upper bound set not by how the wiring is arranged but by how well it is powered.
No amount of connectome-mapping explains that ceiling away, because the ceiling is not in the wiring diagram. It is in the power supply.
Here is where the idea does something a purely neural theory cannot. The same organelle that grades your working memory also grades your grip strength, your vision, your resilience to stress — because mitochondria do this work in every cell that has them, and we carry an estimated ten million billion of them, roughly a tenth of our body mass.
This is why a metric called intrinsic capacity can fuse body and mind into a single number without cheating. Its developer, the aging researcher John Beard, defines it as everything "from the skin in" — how strong you are, how clear your vision is, how sharply you think.
A wiring-first model has no honest way to put memory and grip strength on the same axis. An energy-first model does, because underneath both sits the same currency.
David Geary, a psychologist who has argued this case for years, calls cellular energy "the lowest common currency driving the development and expression of all biological systems." Deficits at the bottom ripple upward through every level above.
That is a strong claim, and it has been contested. Critics rightly note that mitochondria are only one part of a larger system of ATP production; glycolysis contributes too, and substrate supply matters. Geary concedes the point — the focus, he agrees, should be the whole system.
But the caveats do not rescue the old picture. They relocate the action downward, from the network to the metabolism.
And there is now mechanism, not just correlation. In aged mice, the transcription of mitochondria's own DNA falls out of step with the very neuronal activity that once triggered it — and restoring that coupling improved memory. The energy supply is not a fixed backdrop. It responds to use, and it fades with age.
Which brings us to the humbling part.
If the mind runs on an energy budget, the levers that raise the ceiling are not neural exotica. They are the oldest, least glamorous instructions in the book.
Exercise, Critchlow points out, does not merely bulk up muscle; it bulks up the hippocampus, and it pushes mitochondria to become more numerous and more efficient. Sleep, movement, fuel — the tired advice turns out to be a maintenance manual for the mind's power supply.
She warns that a "quiet energy crisis" may be unfolding inside modern bodies: sedentary lives, chronic stress and nutritional excess straining the very systems that generate cellular power.
We have been treating the substrate of thought as something we can outsmart. It is not. It is something we feed.
The final inversion is the one worth sitting with. The physical basis of human cognition is a two-billion-year-old accident — a larger cell that engulfed a bacterium and, instead of digesting it, kept it. Our capacity for reflection, imagination and doubt rests on that ancient microbe.
Before we spend our worry on machines that might out-think us, we might notice what we do to the substrate we already have. We routinely starve it, exhaust it, and let it decay.
If the energy stops flowing, there's no more you. The most sophisticated cognition the planet has produced runs on a merger we forgot to maintain — and the instructions for maintaining it were never exotic. They were only ignored.