Brain scan implant that DT could (almost) build now

[anon3825968]

This may be of interest:

I just came across fNIRS, which stands for Functional Near InfraRed Spectroscopy.

In short (like, in ultra-short), it’s a technique to measure blood saturation in the upper layers of the brain by shining infrared light (810 - 900 nm) through the scalp and skull, detecting the returned light and measuring absorption in-between: the harder you think, the more blood floods to the bit of the brain involved in the thinking, the more the light dims. The wavelength of the IR light is chosen so absoption is minimum through skin and bone, and maximum through blood.

The hardware is very simple: one IR laser diode or IR LED, one photodiode. That’s it. Well, not quite, but close enough. The signal processing and math after that isn’t simple, but that’s another problem. Most importantly, it’s a problem that can be solved outside the body.

Best part of it? There are two open-source fNIRS projects: OpenNIRS and OpenfNIRS, and they’ve put out quite a lot rather practical design specs.

How could DT add value to this? Well, if you put the IR LED and photodiode onto a Flex-style implant and implant it subdermally, the light only has to cross the skull barrier. The LED could be lower-powered, and the S/N ratio could be higher than with an external device, because the light wouldn’t have to cross the scalp twice.

This means it could be powered by induction through the scalp with a low-power Qi-like thing. Possibly even only with NFC. And if it was possible to power it with NFC alone, then it would be possible to use the protocol to get the returned signal measurement from the photodiode.

Imagine that: a basic brain interface with a simple Flex device, a cellphone and an NFC patch kit. Wouldn’t that be nice?

Damn, we manufacture laser rangefinders at work, and one model we sell uses 905 nm lasers. I’m almost tempted to stick one onto my forehead and see if I can see something in the signal histogram. At 2000 euros a pop, I’m not sure my company would like that too much though - not to mention, I’m not paid to do that… Still, tempting.

[Pilgrimsmaster]

@anon3825968

I’m not sure my company would like that too much though - not to mention, I’m not paid to do that… Still, tempting.

A potential new revenue stream
A foot into the medical industry ( If they don’t have already )
A willing and able Guinea Pig

What do they have to loose?

[anon3825968]

1/ We don’t need another revenue stream: we hit 230% of our yearly target for 2020.
2/ We don’t need another ultra-regulated industry to deal with. We have enough red tape with the one we cater for right now.
3/ Yes

[ThePolishedTurd]

It could probably be done with a LF X-series. Just need an antenna like what the Xac access controller has. That would make the installation way less invasive.

[vampire_blue]

Jkjk I know these toys are EEG and not Fnirs, but it made me think of it.

This sounds pretty reasonable, seeing as we have already been able to use qi charging to run a raspberry pi inside the human body. It is mostly a matter of the tech and its form factor

[Eriequiet]

@anon3825968

1/ We don’t need another revenue stream: we hit 230% of our yearly target for 2020.

Also this… please make it stop
Can’t produce enough stuff for MIL and consumers during election years, plus covid

[anon3825968]

Yeah that’s the thing with that industry: the worse things get, the more they spend.

[Eriequiet]

It’s fun for a while… but please stop buying my products guys… I want to not be a work maybe

[amal]

my question is… is this measuring blood flow in the brain, through the skull, or blood flow in the capillaries on the outside of the skull? If it’s the former, through the skull, me thinks you’ll be needing an IR laser or something a wee bit more powerful than an LED.

[Satur9]

I saw a presentation on this by a team at IEEE SPMB 2019. Here’s their paper explaining the data processing they did on data from a NIRScout sensor array:
“Multi-class Classification of Motor Execution Task using fNIRS”[PDF]

[anon3825968]

The OpenNIRS paper mentions 100 to 200 mW - pulsed presumably. That’s through skin. I’d imagine a lot lower than that if you’re already against the skull.

Not outrageous I think.

[anon7067117]

Huh… I think I know who you work for!

[Satur9]

NIRScout

Source Illumination Types	Hybrid Choice: LED & Laser Sources
Source Wavelengths	LED: 760nm & 850nm; Laser: 785, 808, 830, & 850nm

[nil]

what kind of uses would an implant like this have? would we need to get a whole bunch of them installed for this to really do anything particularly interesting?

[anon3825968]

You can do a lot of interesting things with only one channel. For instance, if you measure the brain’s activity at the right location on the visual cortext (EEG, admittedly), you can detect flashing lights - and more importantly, the frequency of the flashing light. So you can have a screen with several zones flashing at different frequencies, and depending on where you look, the signal processor will pick up this or that frequency. That’s how early brain interface prototypes worked, to provide people with locked-in syndrome the ability to “type” by looking at flashing keyboard letters on a screen.

With the hypothetical DT fNIS implant, you could perhaps detect wakefulness, or simply detect that your brain is hard at work and do something when the state changes - like put on soft music, or put the kettle on if your brain has been working for too long. Possibly, depending on which area it’s implanted over exactly, you could train yourself to think certain thoughts to change the signal from the implant, thereby offering a way to control something directly.

It wouldn’t necessarily be practical, but it sure would be damn interesting.

[nil]

a little while after posting that, i was just thinking how something like this could be used in prosthetics. my thought was, take your standard bionic arm, which uses myoelectric sensors to open and close the hand, mayhaps you could combine that with something like this to add more function to the arm? first thought that came to mind was using the fnis chip to deploy a blade built into the arm, while leaving the myoelectrics to control the hand. or less destructively, use it to switch between different grip modes in the hand. this could be pretty fun to tinker with, if it goes anywhere. i’m sure there could be ‘augmentative’ prosthetic uses for this too. brain-controlled third thumb, maybe?

[fraggersparks]

An interesting concept.

@anon3825968

I’d imagine a lot lower than that if you’re already against the skull.

I disagree. I believe most of that IR loss is due to skull, not skin. Skin IR losses in this range are pretty minimal - it’s how pulse oximetry works.

[amal]

@Satur9

Look I just got my DT fNIR implants!

[Eriequiet]

I’ll take a neuralink thank you lol

[ODaily]

@Satur9

I just want the hat. You know, to troll people at work with.

Imagine I just walked into work with it on, sat down at a pc, plugged it into a USB port, and kicked back with my eyes closed and a look of concentration.

“Hey man, what’s tha…”
"Shush, working ! "