Research
Neural interfaces
Reading commands from brain signals. In 2012.
In 2012, Alexandr's research at the University of Southampton read commands from brain signals in three steps: capture, translate, recognise.
- Year
- 2012
- Where
- University of Southampton
- Headset
- Emotiv
- Form
- Research and a conference poster
What it does
- 01
Capture.
Brain signals, taken from the scalp with a headset rather than an implant.
- 02
Translate.
The analog signal turned into digital data a model can learn from.
- 03
Recognise.
Models trained to read commands in the data, such as go, stop and right, tuned by evolutionary algorithms.
- 04
Read the mood.
Beyond commands, the same signals carry emotional state, and the models were trained to recognise that too.
What it was for
The poster set out four uses for it.
- 01
Moving again.
Helping people who can't move to communicate and to steer a wheelchair.
- 02
Games that listen.
Games that change their world and their difficulty with the player's state.
- 03
Mood.
Reading mood to choose the music, or to add the right smiley to a message.
- 04
Safety.
Telling family when a child or an older person is unwell, and stopping a driver who's falling asleep.
Under the hood.What was hard, and what we did about it.
What's on record.
The poster, its three steps and its four uses. The papers that followed aren't online any more, so nothing here leans on them.
Why it's here.
The work since has the same shape: a messy signal in, a decision a person can trust out.
Where it is now
Research, not a product. It was 2012; the instinct stayed.
Get in touch
Still reading? Good sign. Write to us.
- 01A few lines is enough.
- 02A short call. Questions, not a pitch.
- 03A written proposal with a clear price, or a straight no.
Work with people who keep their word.
Next from the workshop · IteroZen
Brief in. Invoice out.