Back to the journal

What is a Brain Computer Interface, explained simply

Neurons, electrodes and algorithms: how a thought becomes a command, no neuroscience degree required.

Explainer by Antonio Blescia 2 min read

Leggi in italiano →

A Brain Computer Interface (BCI) is a system that lets you control a device without using your muscles, starting from the activity of the brain.

1. The brain is electric

When neurons talk to each other they produce tiny electric currents. When millions of neurons work in sync, their signal becomes strong enough to be measured from the outside, on the scalp. This technique is called electroencephalography (EEG).

2. Listening to the signal

The electrodes of an EEG headset measure voltage differences of a few microvolts, that is, millionths of a volt. The signal is extremely faint and full of noise: eye blinks, jaw movements, the mains power in your home.

3. Recognising an intention

The heart of a BCI is the software that recognises recurring patterns in the signal. Some examples:

Paradigm What the user does What the software looks for
Motor imagery Imagines moving a hand Rhythm changes over the motor cortex
SSVEP Looks at a flickering light An oscillation at the same frequency
P300 Waits for a rare stimulus A peak about 300 ms after the stimulus

4. From command to movement

Once the intention is recognised, the system sends a command to the device: a cursor, a virtual keyboard or, in SmartChair’s case, a powered wheelchair.

It is not mind reading

A BCI does not read your thoughts. It only recognises a small number of patterns that the user produces voluntarily after some training. It is a tool you learn to use, a bit like learning to drive.

In upcoming journal entries we will look at each of these steps in detail.

  • #bci
  • #eeg

More journal entries

  1. Who is Bones, the journal's scribe

    Bones is the artificial intelligence assistant that helps me write this journal. Here is where the name comes from, what it does and why it is useful to the whole project.