Action Potentials and Circuits

In 2004, a paralyzed man named Matthew Nagle became the first human to move a computer cursor using only his thoughts. Surgeons had implanted a device called BrainGate — a tiny grid of 96 silicon electrodes — into his motor cortex. Every time he imagined moving his hand, neurons underneath the array fired electrical spikes called action potentials, and a computer learned to read those spikes as commands.

How a neuron generates an electrical signal called an action potential. Why the sodium-potassium pump is essential to keep neurons ready to fire again. How brain-computer interface (BCI) electrodes detect and translate neural spikes. How a real BCI system, BrainGate, turned Matthew Nagle's thoughts into cursor movement.

A resting neuron sits at about -70 millivolts (mV) inside compared to outside its membrane — this is its resting potential. When enough input pushes that voltage up past a threshold near -55 mV, voltage-gated sodium channels snap open. Positive sodium ions (Na+) rush in, spiking the voltage up to about +30 mV in a process called depolarization. Almost immediately, potassium channels open and potassium ions (K+) rush out, bringing the voltage back down — repolarization. The whole spike takes roughly 1 to 2 milliseconds, and it is all-or-nothing: either the threshold is crossed and a full spike fires, or nothing happens at all.

After a spike, sodium is now inside the cell and potassium is outside — the opposite of where they started. The sodium-potassium pump fixes this by using energy (ATP) to push 3 sodium ions out for every 2 potassium ions it pulls back in. This resets the neuron's resting potential so it is ready to fire another action potential. Without this pump constantly running, a neuron could only fire a limited number of times before running out of the right ion balance.

A brain-computer interface needs to detect these tiny electrical spikes from outside the cell. The Utah Array, one of the most common BCI implants, is a 4mm by 4mm grid holding 96 hair-thin silicon needles, each one recording voltage changes from nearby neurons. Software called a decoder watches the pattern of spikes across all 96 channels and matches that pattern to an intended movement, like moving a cursor left or opening a robotic hand.

Matthew Nagle, paralyzed from the neck down, received the BrainGate implant in 2004 as part of a clinical trial. After calibration sessions where he imagined specific movements while researchers matched his neural spike patterns to a cursor, he learned to move a computer cursor, open email, and play a simple video game using thought alone. His trial proved that a working circuit — from single neuron, to electrode, to decoder, to computer — could restore a form of movement.