Watch a mouse eat a strand of dry pasta. It sits back on its haunches, grabs the piece with both front paws, rotates it, and feeds one end into its mouth while its hands shuffle the rest along. The whole thing looks effortless. It is not. Coordinating paws and jaw in real time is one of the harder tricks a mammal pulls off, and it happens dozens of times a meal without a thought.
A study published in Neuron on March 26 traces that coordination to a specific slab of cortex, and shows that the brain treats "moving the hands and mouth" and "syncing them together" as two separate jobs handled by two separate types of neuron.
Hunting for the conductor
Spinal and brainstem circuits already know how to twitch a forelimb or open a jaw. What was missing was the part of the brain that strings those elementary actions into feeding. Xu An and colleagues at Duke and Cold Spring Harbor went looking for it with optogenetics, flashing light onto patches of motor cortex in mice to see which ones triggered eating-like movements.
They landed on a region they call the rostral forelimb-orofacial area, or RFO. Stimulate it and a mouse produces coordinated paw-and-mouth movements that resemble natural eating, even with no food present. The RFO is wired in both directions to the sensorimotor areas that control the forelimb and the face, sitting in a position to bring those two streams together.
One cell type moves, another keeps time
The more surprising finding is how the RFO divides its labor. Two classes of projection neuron live there. Pyramidal tract neurons send their axons down to subcortical motor centers, the machinery that actually drives muscles. Intratelencephalic neurons instead talk to other cortical areas and to a part of the striatum, staying inside the brain's own coordination loops.
When the researchers recorded from freely eating mice, both cell types lit up during hand-mouth manipulation. But silencing them produced different failures. Switch off the pyramidal tract neurons and the dexterous hand-to-mouth movements themselves fall apart. Switch off the intratelencephalic neurons and the individual movements survive, but their timing goes off. The mouse can still move its paws and mouth. It just cannot line them up.
That split maps onto an intuitive distinction. One population is the muscle, the other is the metronome. The dexterity and the choreography are handled by different wires, which is why a brain can lose the rhythm of a skilled action while keeping the raw ability to perform its parts.
Why a pasta-eating mouse matters
Oromanual feeding, as the authors call it, is not unique to rodents. Primates do the same hand-and-mouth dance, and so do humans every time we bite an apple we are holding. Finding a dedicated cortical circuit for it in mice gives researchers a concrete, manipulable system to study how the brain builds any multi-limb skill out of simpler pieces. The logic here, drive the parts with one cell type and coordinate them with another, could well repeat elsewhere in motor cortex.
A few limits are worth keeping in view. This is mouse work, and the leap from a rodent nibbling pasta to a human's far more elaborate hand skills is not automatic. The stimulation experiments show that activating the RFO can produce feeding movements, which is not quite the same as proving the region is the sole author of natural eating. And the striatal side of the intratelencephalic circuit is sketched more than nailed down here. What the paper does establish cleanly is the division of labor: in this corner of cortex, executing a movement and timing it are genuinely separable operations, carried by neurons you can tell apart.
It reframes a mundane act. The next time you eat something you are holding, two different neural populations are quietly negotiating, one firing the movements, the other making sure your hand and mouth arrive at the same place at the same moment.
Comments