New neurotech eschews electricity for ultrasound(spectrum.ieee.org)
spectrum.ieee.org
New neurotech eschews electricity for ultrasound
https://spectrum.ieee.org/bci-ultrasound
24 comments
According to the article, ultrasound can measure brain activity and it can influence brain activity in deeper regions of the brain without invasive probes.
This sounds very promising vs electrical techniques.
This sounds very promising vs electrical techniques.
Stimulation but not recording.
“Recording deep neural activity via ultrasound can’t be done through the skull, but it can be done by removing a piece of skull and laying the device on the surface of the brain.”
“Recording deep neural activity via ultrasound can’t be done through the skull, but it can be done by removing a piece of skull and laying the device on the surface of the brain.”
There's an obvious ick factor, but when I think about it, it's not really so bad compared to some of the other things we do to our bodies. I can easily imagine getting a chip of skull replaced with a subdermal device, like you might get a tooth replaced or a pacemaker installed.
To me it's scarier because of the blood-brain barrier and restricted immune response--keeping disease out in the first place is probably more important than usual.
> Ultrasound can also be used to estimate neural activity within brain regions by measuring local changes in blood flow via the doppler effect
I don't know much about how their stimulation is supposed to work, but I do know quite a lot about ultrasonic imaging and I can tell you this is not going to get you the neural interface that people really want. It's going to get you slightly better "think really hard and the drone flies!" toys.
I don't know much about how their stimulation is supposed to work, but I do know quite a lot about ultrasonic imaging and I can tell you this is not going to get you the neural interface that people really want. It's going to get you slightly better "think really hard and the drone flies!" toys.
From the article:
> For stimulation, focused ultrasound waves can flexibly target specific regions throughout the brain from outside the skull.
> For stimulation, focused ultrasound waves can flexibly target specific regions throughout the brain from outside the skull.
this is used for treatment of tremors now, fairly successfully.
If of interest to anyone here, my company has built an open source software library for transcranial focused ultrasound stimulation: https://github.com/agencyenterprise/neurotechdevkit
Its goal is to accelerate the development of agency-increasing neurotechnology and lower the barrier of entry for any developers to be able to solve open problems in neurotech without having to have their own hardware or human subjects. It's starting with ultrasound, which we find quite promising, and we hope to expand in the future to other areas as well.
Its goal is to accelerate the development of agency-increasing neurotechnology and lower the barrier of entry for any developers to be able to solve open problems in neurotech without having to have their own hardware or human subjects. It's starting with ultrasound, which we find quite promising, and we hope to expand in the future to other areas as well.
>The fUSI technique estimates changes in neural activity by measuring changes in blood flow. Neurons, like all cells, need blood to function. Increases in neural activity require increases in blood flow, which fUSI techniques measure by projecting ultrasound at a brain region of interest and recording the waves that bounce back. When sound bounces off of a mass of flowing blood, the returning sound waves wiggle at a different frequency than those emitted. fUSI uses this Doppler shift phenomenon to estimate changes in blood flow and, by proxy, the electrochemical chatter of neurons.
So, it seems this works similar to fNIRS which measure changes in the blood oxygen level via near-infrared spectroscopy. The downside of this is, that changes of the blood oxygen level happen significantly slower than the generation of action potentials thus making this technology not really feasible for real-time applications.
So, it seems this works similar to fNIRS which measure changes in the blood oxygen level via near-infrared spectroscopy. The downside of this is, that changes of the blood oxygen level happen significantly slower than the generation of action potentials thus making this technology not really feasible for real-time applications.
First one to stimulating the pleasure centers wins the 'wire head' prize :-)
The links from Sarah Constantin are awesome, thanks!
The links from Sarah Constantin are awesome, thanks!
not pleasure centers but getting there:
- https://consciousness.arizona.edu/sema-lab
Can it be done remotely? :-)
There is a brain washing machine operating in the University of Oulu.
>The operation of FUS LiFu is based on low-intensity, focused ultrasound. The device allows for the temporary opening of the blood-brain barrier and the delivery of water-soluble drugs to the precise location in the brain. Combining FUS technology with the specialized expertise in Oulu, the ultra-fast functional magnetic resonance imaging (fMRI), opens up entirely new possibilities in brain disease research.
https://www.oulu.fi/en/news/brain-cleaner-accelerates-brain-...
>The operation of FUS LiFu is based on low-intensity, focused ultrasound. The device allows for the temporary opening of the blood-brain barrier and the delivery of water-soluble drugs to the precise location in the brain. Combining FUS technology with the specialized expertise in Oulu, the ultra-fast functional magnetic resonance imaging (fMRI), opens up entirely new possibilities in brain disease research.
https://www.oulu.fi/en/news/brain-cleaner-accelerates-brain-...
That's super cool - and I don't know why we don't look into this as a replacement to provide a 3D mapping of parts of the body, for scans such as x-rays.
Another thing I'd like to see - why not both?
There're some very useful AI applications associated with being able to map between different forms of the same data;
With accurate enough processing, a low-end device using the cheaper of any two mechanisms could be used to emulate a more expensive high-end device.
(The way things are, this will most likely be implemented with a subscription model, a proprietary cloud processing engine, and a vendor locked operating system.)
Another thing I'd like to see - why not both?
There're some very useful AI applications associated with being able to map between different forms of the same data;
With accurate enough processing, a low-end device using the cheaper of any two mechanisms could be used to emulate a more expensive high-end device.
(The way things are, this will most likely be implemented with a subscription model, a proprietary cloud processing engine, and a vendor locked operating system.)
Magnetic resonance spectroscopy is capable of highly detailed 3D imaging of anatomical tissue. However it's time consuming and expensive.
3D ultrasound has been a thing for quite a while.
The problem is that ultrasound just isn't very good for general imaging. The physics of ultrasound means that you must have direct skin contact with a gel mediator to couple the ultrasound into your skin. The signal cannot pass through open air.
Any type of ultrasound imaging would require removing your clothes, having the technician's hands on you, and you get covered in goop.
Compared to other imaging technologies, there's no contest. For x-ray, you simply stand in front of the plate, or put your body part on top of it. No disrobing, no goo, no one even has to touch you. With CT and MRI, you just put the body part (or whole body) in the scanner. Clothes and all.
Ultrasound has it's place, but it really is an inferior technology.
The problem is that ultrasound just isn't very good for general imaging. The physics of ultrasound means that you must have direct skin contact with a gel mediator to couple the ultrasound into your skin. The signal cannot pass through open air.
Any type of ultrasound imaging would require removing your clothes, having the technician's hands on you, and you get covered in goop.
Compared to other imaging technologies, there's no contest. For x-ray, you simply stand in front of the plate, or put your body part on top of it. No disrobing, no goo, no one even has to touch you. With CT and MRI, you just put the body part (or whole body) in the scanner. Clothes and all.
Ultrasound has it's place, but it really is an inferior technology.
"inferior" technology? Let's not make value judgements on this. X-ray and MRI both have significant and different problems of their own that ultrasound doesn't have. Horses for courses.
This is not a value judgment and the concept of ultrasound does not have feelings you need to white knight for.
I'm stating the objective fact that ultrasound is worse at producing images than other imaging techniques. There is no "both sides" argument to be made here.
I'm stating the objective fact that ultrasound is worse at producing images than other imaging techniques. There is no "both sides" argument to be made here.
No, you said "but it really is an inferior technology.". It's not inferior if you need something completely safe, or if you're imaging something with good acoustic contrast, or if you need something that doesn't need its own room, or any other reason why you might prefer ultrasound over MRI or CT.
The fact that it's used at all would suggest a priori your comment is wrong.
Even if you want to restrict the scope of ultrasound being an inferior technology to only ultrasound being worse at producing images, that's not true either. Show me a CT or MRI showing what a Doppler ultrasound image shows, or that shows a real-time live 4-chamber view of the heart (to pick a couple of obvious examples - there are many more). So now you might say ultrasound is objectively worse at at producing images that CT and MRI excel at, at which point, what are you trying to say?
The fact that it's used at all would suggest a priori your comment is wrong.
Even if you want to restrict the scope of ultrasound being an inferior technology to only ultrasound being worse at producing images, that's not true either. Show me a CT or MRI showing what a Doppler ultrasound image shows, or that shows a real-time live 4-chamber view of the heart (to pick a couple of obvious examples - there are many more). So now you might say ultrasound is objectively worse at at producing images that CT and MRI excel at, at which point, what are you trying to say?
Optimistic about ultrasound for sleep and aging applications, the waste clearance processes are just being understood. Turning brain areas off has also been shown to help different types of performance.
The article mention other techniques also stimulating the brain, but aside form transcranial magnetic or DC stimulation I never of this working.
Has anyone usefully stimulated a brain (ie input information) using electrodes?
Has anyone usefully stimulated a brain (ie input information) using electrodes?
This seems interesting for medical diagnostic applications. I can not envision this as a way to a working bci.
* https://sarahconstantin.substack.com/p/ultrasound-neuromodul...
* https://sarahconstantin.substack.com/p/how-ultrasound-neurom...
* https://sarahconstantin.substack.com/p/transcranial-ultrasou...
* https://sarahconstantin.substack.com/p/why-should-neuroenhan...
* https://sarahconstantin.substack.com/p/whos-working-on-ultra...