We live in the age of neuroquackery. Almost everything you hear about the brain in the popular media is false.
For example, serotonin is not the “happy chemical”. It’s actually a metabolic depressant and vasoconstrictor, associated with dozens of the most debilitating diseases known to humankind. But that’s a longer topic for another article.
Which brings us to one of the current hot topics in the media: brain-computer interfaces and Facebook’s efforts to read our minds.
With the announcement of Elon’s musk’s “neural lace” and Facebook’s “mind reading technology”, there are few more talked about topics in brain science.
In today’s newsletter, I’m going to give you a brief primer on the most popular brain measuring & imaging technologies and tell you why each of them is ill-suited for the aims Elon & Zuck have laid out.
With that, let’s get started.
These are the primary brain measuring & imaging technologies:
EEG (Electroencephalogram): Measures electrical current changes in the brain. When neurons “fire” (release chemicals at the synapse) they undergo a robust electrical potential change. This can be picked up via electrodes placed on the head (usually with a conductive jelly—which amplifies the signal).
When I did a sleep test down at Stanford Hospital, I had these glued all over my head. It took about 30 minutes for a trained nurse to do this properly, and it made me look insane.
I did purchase one of the early “at home”, gel-free EEGs, but it was an inaccurate mess.
Electrodes (implanted): If you want to get really crazy, you can open up the skull and place electrodes IN the brain.
I was lucky enough to witness an operation to do this (on a human) while in college. Stanford is one of the world leaders in “deep brain stimulation” treatment for Parkinson’s. In that disease, dopamine-producing neurons in the substantia nigra start to degenerate, causing them to produce and release less dopamine (which is one of the primary movement-related chemicals in the brain). To combat this, a surgeon can open up the brain and slide an electrode down where these dopaminergic neurons cluster. Pulsed electrical signals are sent down the electrode to stimulate the neurons in the desired manner.
It’s an intense, and LONG, operation. I think I was there for the first 4 hours of a much longer surgery.
NIRS (Near Infrared Spectroscopy): If you shine a really, really powerful infrared light against the skull, you can measure blood-flow in the brain. Since active neurons require more oxygen and glucose, they need more blood compared with their less active neighbors. So the blood flow rate is a good proxy for neural activity. Since every protein has a specific electromagnetic absorption signature, you can emit and detect the returning signal to determine which proteins or chemicals are in the area (and in what concentrations). This device measures hemoglobin concentrations and helps the researcher see the change in hemoglobin concentration (and thus neural activity) over time.
I actually used to use an at-home NIRS device, but it didn’t seem very accurate… I gave up on the at-home version of the technology.
MRI / fMRI: There are two major flavors of MRI technology – MRI and fMRI.
MRI stands for Magnetic Resonance Imaging.
fMRI stands for Functional Magnetic Resonance Imaging.
MRI produces extremely detailed snapshots of the brain, and is, therefore, the tool of choice for cancer and structural abnormality detection.
fMRI is a video-camera for the brain. But, like a video-camera, produces much less detailed snapshots. Like the NIRS, it measures blood flow but using a different method. It measures the ratio of oxygenated to deoxygenated hemoglobin—the ratio of hemoglobin WITH oxygen to hemoglobin WITHOUT oxygen.
So what should Facebook use?
Since thoughts are not static entities, but contained within neural activity, to do what Facebook wants we need to use one of the video-camera technologies listed above. That boils down the possibilities to:
- fMRI
- NIRS
- EEG
- Implanted electrodes
fMRI:
As a mind-reading device for at-home / popular use, I can’t think of a worse technology than fMRI. It’s huge (see picture below), requires professional technicians, and is insanely expensive. There’s no way that this technology will be used for Facebook’s mind-reading initiative.
NIRS:
These devices are fairly cheap, but (as I said before) the accuracy isn’t great. I don’t think that they’d have enough time (or location) sensitivity to pick up on something as subtle and short-lived as a thought. In addition, NIRS is good for looking at cortical activity (the outside layer of the brain), but can’t go deeper. It also can’t measure any activity through hair… so everyone using one of these devices would have to be bald.
EEG:
EEG suffers from the same problems as NIRS. It can only measure surface (cortical) activity and requires skin-to-skin contact (and/or a healthy dose of conductive gel). It has good time resolution, but the output you get can be a bit fuzzy—the spatial resolution isn’t always great from what I’ve seen.
Implanted Electrodes:
Unless we require everyone to go through dozens of hours of invasive (and high dangerous) surgery, I don’t think that this will work.
So I have no idea how Facebook and Musk are going to create this mind-melding tech using the current versions of the technologies that exist today.
If you have any ideas (or thoughts) please send them over.
