A BCI research lab for the moment after medicine
Brain-computer interfaces are medical devices today. They won’t stay that way. Between Labs studies how a touch or a feeling could pass from one person’s brain to another’s.
Writing, the telephone, film, the screen. All of it reaches us through the eyes and ears. A touch stays with the person who felt it. So does the mood of a room.
A brain-computer interface could carry more. The hard part is translation. No two brains encode the same experience the same way, so a pattern recorded in one person means little to another until it has been translated.
We call this the transmission of experience. It will take years of careful science to solve, and we intend to do it in the open. We will publish what we find, including what fails.
Our missionSolve the transmission of experience, then use it to close the distance between minds.
Implanted interfaces already let people with paralysis control a computer, and in recent trials, speak again through a synthesised voice. Restoring function is the right place for this technology to start.
Sooner or later these devices will matter to people who aren’t patients. We call that point the moment after medicine.
When it arrives, one of the most important things an interface could do is let two people share what something is like. Getting that science right, along with the rules around it, will take years. We have started.
The work has three parts. The first needs no surgery and starts with data that already exists.
Models that map one person’s brain activity onto another’s, so an experience recorded in one brain can be read in the terms of another. The same models could cut the calibration every new interface user sits through today.
Software that turns a decoded experience into stimulation a second person can feel, tuned to how their brain responds. It runs on implants and devices made by our partners.
Who can send and what gets kept are engineering decisions here. The four rules are below.
Each one has a pass bar, set in writing with independent reviewers before any data comes in. Null results get published too.
From one person’s brain activity alone, can a shared model tell which of several new experiences a second person had?
Where it standsShared models across people have been shown in fMRI for visual categories1 and for language.2 Nobody has shown it for touch or emotion yet. No stimulation is needed.
Can a touch felt by one implanted participant be rendered through a second participant’s own stimulation map, and identified by them in a blind test above chance?
Where it standsStimulating human somatosensory cortex produces touch sensations felt at specific points on the hand.3 Participants who tuned their own stimulation could match sensations to objects above chance without visual cues.4 Direct links between brains have so far carried single bits.5,6 As far as we know, no one has yet passed a percept from one person to another.
Can a calibrated state in one person, such as calm, be induced in another more reliably than a sham, with the receiver able to tell it came from someone else?
Where it standsMood can be decoded from intracranial recordings within a person.7 The effect of stimulation depends on the state a person is in at the time.8 Focused ultrasound can change activity in human sensory cortex without surgery.9 Passing a state between people, and knowing it was received, are still open.
Reading the brain will likely move at the speed of machine learning. Writing to it moves at the speed of clinical trials. That is why the first milestone stands on its own. There is also a serious argument that a brain can’t take in much more than language already gives it. We think that is worth testing.10
Everything we build has to meet them. A result that breaks one doesn’t count.
One sender, one receiver, both consenting. Feelings are never broadcast to a crowd.
The sender opens and closes the channel, on purpose, every time. A version of this already exists for speech interfaces: a silently imagined keyword that locks and unlocks the decoder.11
Transmission is live. Nothing is stored by default. Research data is kept only as long as an approved protocol requires.
The receiver has to know the experience came from someone else. If it feels like their own, milestone three has failed.
Every implanted participant in a partner study has ongoing care funded from the day they enrol. An independent ethics lead can veto any protocol.
Device makers build the hardware. We work on what passes between people, and license it to the companies that need it.
Cross-person models and stimulation software that run on your systems. We start with devices already in human studies and design the research alongside your clinical teams.
Sponsored research under local ethics approval. Publication is the default, after a short window for patent filings.
When transmission works in the lab, companies building for people will want it. The rules of transmission come with every licence.
Velco spent two decades building experiences for brands. He kept hitting the same limit: so little of an experience survives being passed on.
That limit led him to brain-computer interfaces and to Neural, his book on what the technology will mean for commerce and power. He founded Between Labs to start on the science now.
We’re looking for a co-founder to lead the science. You set the agenda and the standard of evidence.
Maybe you have built models of brain activity that hold up across people. Maybe you have produced sensation through stimulation and run the psychophysics to show what was felt. Either way, we want to hear from you.
Write to us →The lead scientist comes first. After that we hire scientists in decoding and stimulation, machine learning engineers, psychophysicists, and designers who help decide what is worth sending.
If you work on reading or writing the brain, we’d like an hour of your time. Device makers, research partners and press, write to us.