In my opinion [0] a much more interesting track to pursue than super high-T scans that only work on mice (dead mice at that) is low field MRI scanning. MRI is amazing, but incredibly bulky and slow. In Canada you have to live for 6 months with a knee injury before surgery just because of the wait to access an MRI. Companies like Hyperfine (not affiliated) and Deepspin (just learned of them) make low power portable machines (I believe NFL teams were early costumers).
A single MRI image is not super helpful, still more art than science, the real benefit comes from having periodic scans against a baseline. Boutique health clinics offer yearly scans to high net-worth individuals, rolling this out to a larger population could have a huge impact.
[0] educated opinion, I also collaborated on a PNAS MRI paper
> A single MRI image is not super helpful, still more art than science, the real benefit comes from having periodic scans against a baseline.
An MR when someone has a problem is pretty helpful.
Screening MR is not really a good thing except in a very narrow range of conditions - you just generate unnecessary procedures, more imaging and stress.
In regard to low field MR, the likes of Siemens Deep Resolve and their low field magnet (0.55T) have taken this a long way, and really improves image quality. It looks useful for imaging around implants and metal.
If you have a major issue, say a tumor, MRI is great, but for smaller things I agree with you that a single image is not great for a quantitative assessment. What does have a lot of value is when you have a baseline image and you perform yearly scans that are compared against each other (similar to how carotid ultrasounds are used for cholesterol assessment)
> In asia you can walk into a clinic and see a doctor in 30m
... If you're rich relative to the general population, then you can go to an expensive private clinic and be seen quickly, yes.
I doubt that's the case for public hospitals and in any case, I've been inside public hospitals in Indonesia, Cambodia, and Vietnam (all for other people's needs, nothing major thankfully). Trust me, you do not want to get treated in one of those and as a westerner you'll never even see them unless you friend gets bitten by a monkey in the Cambodian jungle (for example) and there's no private clinics nearby. I always assumed the casual attitude towards hygiene that Asian people have wouldn't extend to hospitals. I was wrong. There were literally blood stains on the wall in the corridor. The toilet was a hole in the ground with a bucket of water and a scoop to wash your hands and ass with. There was no soap. This wasn't a small hospital, it was the main hospital for a city of about 200k people.
This is the result of single payer or nationalised systems where private healthcare is not allowed.
So in systems like the UK, you can walk into a clinic for an MRI, you just have to pay privately for that access, usually covered if you have additional medical insurance. If you want the procedure though nationalised healthcare (NHS), there is a wait.
Health insurance is relatively cheap in the UK compared to the US. A lot of frontline care is done via the NHS contracts with the GPs and the A&E service at the NHS hospitals.
This is also how it works in US, you just go to imaging place and they do it immediately. You just have to pay for it. In countries with state controlled healthcare, service is rationed.
> The healthcare system in the west is just plain stupid.
What the hell is "the west" in this context? Do you have experience with every health care system in Europe? Why is "NZ" considered "the west"?
> In asia you can walk into a clinic and see a doctor in 30m.
You can in many places in Europe too.
The flip side of the asian systems can be incredibly overworked and stressed health care workers. I know the service in Taiwan is fantastic (though a bit stressful, was way busier than in Norway), but I also know a nurse working there saying the working conditions are quite awful. They also encourage doing way more testing than necessary. That makes it so testing is pipelined and efficient. You get economies of scale. But doing this is generally considered a very bad idea, because you can get false/misleading diagnosis and undergo unnecessary treatments with potentially bad side effects. Along with all the wasted time and resources that could have been spent working on actually sick people. Which in asian countries can mean that poor people don't get the resources they need to get treatment.
I walked into an ER in Tucson, Arizona last winter, and was out within an hour with a CT scan and clear diagnosis. Took me another 2 days to get the raw data. It was awesome.
My coworker fell in Costa Rica, hurt her knee and came back and had an MRI the same week. I saw my doctor in mid-March and she ordered an MRI on my neck, it's scheduled for next week. Do some people wait 6 months... yes, those that can wait 6 months are asked to wait 6 months. Those that need it quickly, get it quickly. It's the triage method.
I also had the option to book a private MRI and get it next day. It would have been $650CAD or $483USD. I happen to be in LA when my issue came up and I looked at private MRI. It was way more expensive than the Canadian alternative (however, it could have been LA/Hollywood pricing).
No it's not. I tweaked my knee pretty badly to the point where i would wake up at night from deep sleep due to pain (and I generally have a higher pain tolerance) and the best the doctor could do was prescribe ultra sound cause getting an MRI scan would take 6-7 months.
Even then the earliest appointment I could get for a knee specialist to look at my knee is 6 months.
It really depends on where you're. If you're in the city of Toronto, good luck getting an MRI appointment.
> those that can wait 6 months are asked to wait 6 months
You do know that you're body doesn't decide to wait 6 months for an MRI and stop developing scar tissue right?
In India, you can book an appointment for MRI scans at private clinics the very same day or at worst within 1-2 days. Costs about USD 100 or lesser.
There is also this startup - https://qure.ai (Not my employer, I know someone who works there) which has already automated the report generation part for several classes of X Ray / CT / Ultrasound scans. This takes away the requirement of having a trained Radiologist to go through the scan and then create a report, which certainly helps in scaling the availability of such scans - you just need the machine and trained staff to operate it, and are not dependent upon having a doctor with a postgrad in radiology to write the reports.
MRI scans in our budget MRI lab (yes that’s correct) a km away is 2500 Rs or like 60 USD and less than an hour wait time. How MRIs got commoditized in india through competition should be studied hard.
I just checked and MRI scans in Ireland are about 300eur, while the average monthly wage is about 3300eur.
I know average monthly wage isn't the most robust of measures, but still, looks to me like MRIs are pretty expensive in India relative to people's ability to pay for them.
Healthcare in India is rather amazing. When my wife was pregnant, we could get a sonography scan at a top private hospital for $30-40 without any insurance. The scan was done by an experienced doctor, not a technician.
Off topic: "high net-worth" is word "rich" not politically correct anymore or what is the thinking behind using three words instead of one? This reminded me of "George Carlin on soft language"
There are plenty of "old money" rich that are no longer "high net-worth" individually but their family name and other resources available to them more than make up for it.
In Uruguay I got (multiple times) a 3T MRI for $300 in under 24h (https://cudim.org/). Or 1.5T for free, no more than a week after the doctor asks. Anyone in the world can go there and pay for that and get it.
It's insane to belive I have a 3T MRI 2 blocks away from my home.
Now while staying in the US I just pray I don't need have health issue.
>in Berlin I can get a knee MRI appointment within two days
Are you sure about this? Luckily I have not been in need of an MRI for a few years, but I remember having to wait a bit despite having private insurance.
Hamburg here, same. Got 2 MRIS in the last 5 years and both could have been the next business day if it would have fit my schedule. and all that as a "kassenpatient"
Yes that does sound quite interesting and valuable v line of research but don’t you think one-off high-res MRI scans might advance humanities understanding of the brain in a meaningful way more so than low-field MRIs for all? I think this is more about the long game (scientific development) than how an average person right now could benefit from the new advance perhaps.
My family has had same-day MRIs in the USA. I’ve never heard of anyone waiting, ever, outside maybe a few days or longer if they’re trying to line up their own schedule with availability.
In Mexico you can buy a (private) head
MRI scan with contrast for around $300 USD. It's really cheap. So theoretically you could do one every couple of months.
Not an either-or situation at all. There is no one “more interesting track”. Always tradeoffs as I am sure you know.
The key advances in this paper are four-fold:
1: Much higher spatial resolution and contrast that is essential to define boundaries of brain regions
2. Much faster scan time using multiple angle acquisitions
3. Accurate segmentation of brain regions of interest even for highly diverse genomes of mice.
4. Accurate alignment and multimodal merging of MRI with 3D lightsheet microscopy from the same animal with fluorescent labelling of multiple CNS cell types.
fMRI in mice is limited in its resolution. To generate time series data we use large sets of isogenic (identical twin) mice. See figure 5 in the paper for a comparison of young and old mice that are identical twins—one a young adult and one quite old adult.
Canada triages MRI access to high-urgency cases. My partner got an immediate MRI and 3 follow-ups within a month when she presented at an ER with stroke symptoms.
> Boutique health clinics offer yearly scans to high net-worth individuals, rolling this out to a larger population could have a huge impact.
And are the outcomes of these high net-worth individuals any better than the overall population, in ways that cannot as easily be explained by other factors? I'm not really sold on the idea, and a lot of scientists and doctors aren't either.
> In Canada you have to live for 6 months with a knee injury before surgery just because of the wait to access an MRI.
Really? I had a lower back MRI appointment at the hospital inside a month or so about 2 years ago, in the middle of the pandemic. Is it the surgery wait itself that takes that long, or did waiting times generally become that worse in that 2 years?
All appointments are triaged. Some people get them 2 minutes after seeing a doctor, some people wait 6 months. It all depends on need. Source - work in a hospital, have an MRI scheduled next week.
Sure it would be nice that everyone gets an MRI when they need it, but they are expensive pieces of equipment, requiring a very specialized room and staff.
damn that’s fucked up. at my medical facility when my doc ordered an MRI i literally just walked over to radiology and got it done on the spot for like a $40 copay.
I didn't expect this one would be one of those "IN MICE" things...
In a decades-long technical tour de force led by Duke’s Center for In Vivo Microscopy with colleagues at the University of Tennessee Health Science Center, University of Pennsylvania, University of Pittsburgh and Indiana University, researchers took up the gauntlet and improved the resolution of MRI leading to the sharpest images ever captured of a mouse brain.
It's not just "in mice", it's "in dead mice" AFAICT, as they must open the mouse to use LSM. Mice tend to not work after you open them. It also is difficult to get high resolution images of samples disturbed by sources of noise such as blood flow. I had an uninformed perspective that for a given field strength the resolution is limited by how long the subject can stay still. This is obviously a lot longer for one that isn't alive and is screwed to the platform.
Of course, lots of important research will benefit from improved scanning resolution on dead subjects -- one might get mislead by the headline, especially if one isn't a research scientist mostly concerned with scanning things you've poisoned or genetically programmed to die in interesting ways.
Yes, several hours per image. This resolution is still an achievement. But in vivo images (alive) are much more useful as they enable longitudinal studies, but obviously far more challenging to get anything close to good resolution in a mouse. More like 150 micron in alive, sedated mouse brains, even from 9.4T.
Co-author here: No this is actually becoming a less expense acquisition protocol at the cost per voxel per unit of scan time. Of course not intended for clinical application but rather to understand the basic biology and genetics of Alzheimer’s, Huntington’s And Parkinson’s diseases, addiction, and aging.
Does anyone think we understand AD and how best to intervene? This technology is already proving useful. Stay tuned—-this is the tech paper but the application papers are in the wirks.
Strange that a lot of the comments here basically say "who needs this?"
But I remember when TVs were 640x480-ish and 1920x1080 was amazing. And 4k.
Maybe this won't be for knees in live patients. Maybe it won't be for MRIs in the far-flung corners of the world.
But what if it was for researchers to find elusive parts of the brain? What if the high quantity of data led to better algorithms for current MRIs to make them faster (which probably means cheaper).
I think of pure science and how it doesn't immediately teach us something, but later it answers (or provokes) questions and moves the needle.
Just saying.
That doesn't mean I don't groan at some advances - like any advance that "adds value" to some product by collecting more data without your consent. :)
Yes, unfortunate PR title that we the co-authors do not control. This generates some adverse gut response—-even for us. Headline writing is tricky in terms of the balance between reality and README.
This is a very clickbaity title. Over 6 years ago I worked as a student on reconstruction of 50 micron human brain (part, occipital lobe) MRI images, also captured in 9.4T.
Obviously a 10x higher resolution is still impressive, but it’s a mouse brain, and it’s certainly nowhere close to 64M times the resolution of comparable prior work…
Co-author here. Recommend reading the paper itself. The resolution comparison is to clinical MRI, not previous state of the art. Improvement compared to previous state of the art is substantial in terms of both resolution AND throughput. This work is a genuine game-changer for our neurogenetics team. We can generate 100s of very high resolution MRI dataset and get great global estimates of volumes and connectivity. We are using these methods in combination with “humanized” Alzheimer’s disease mice.
Side question: are there any interesting brand new medical imaging modalities being worked on these days or is all work focused on improving existing ones?
Yes, lots of innovative new algorithms to process the diffusion tensor metrics and images. That is how we got down to nominal resolution of 5 microns for the tractography. The effective resolution is closer to 10 microns.
I have a heart stent that is only rated to 3T, so I'm kind of sad that I can't make use of the newer MRI technologies (granted, this particular MRI imaging doesn't seem to be a thing for humans yet).
I'm wondering if, at some point, I'll be able to have it replaced with a newer material that isn't ferromagnetic (or wouldn't be affected like mine is)
Your stent is only rated to 3T because that's about the absolute maximum field strength used by human clinical MRI systems, with the vast majority being half to a quarter that.
As you noted, this imaging was done at much higher field strengths, ~10T. That sort of field strength is practical at the scale of a system sized for a mouse. For a human, it's likely to never be practical for a whole slew of reasons related to the energy needed, forces on the machine, the increase field size and strength affecting room construction, safety thresholds, etc.
1.5T and 3T are very common, and 4-7T are definitely used in a number of hospitals. UMN has a 10.5T magnet they've shoved humans into. Their stent is rated to 3T for a variety of reasons including:
1. That's all they've tested (as you alluded to).
2. It might actually fail at higher fields.
3. Stents act as antennas and can focus RF into unwanted places like cooking the middle of the brain, and it's a lot more likely that the stent is rated for exactly 3T than up to 3T since the relevant frequencies are proportional to field strength, and the associated wavelengths can coincide more or less with the length of the stent (kind of like how the Tacoma Narrows bridge could have experienced a lot more wind, but that particular set of factors was problematic). It _might_ be certified for 0-3T, but that takes additional research to actually do.
4. Any number of other reasons. High fields have a variety of counter-intuitive properties.
We have a complete map of C. elegans, a small worm with 302(ish) neurons. We have a complete wiring diagram of the adult worm’s neurons as well as a “fate map” showing how those connections form during development.
Nevertheless, it’s fairly hard to simulate. Neuromodulators cause all sorts of “spooky action at (synaptic) distance” that the wiring diagram can’t explain, for example.
to pile on, most impressive "complete mapping" so far is without a doubt the hemibrain connectome which took months and cost ~$50M, for a not quite good enough resolution image of a microscopic brain that was quite dead.
The only interesting question in here so far and also the only one that got a downvote from someone. 4 microns is about the size of the cell body of the smallest neurons so this technique could potentially give you the position of every neuron in the brain but isn't going to pick up the dendrites and synapses. You'd need resolution closer to 10 nanometers for that. 1 micrometer (micron) = 1000 nanometers
Readit News
A single MRI image is not super helpful, still more art than science, the real benefit comes from having periodic scans against a baseline. Boutique health clinics offer yearly scans to high net-worth individuals, rolling this out to a larger population could have a huge impact.
[0] educated opinion, I also collaborated on a PNAS MRI paper
An MR when someone has a problem is pretty helpful.
Screening MR is not really a good thing except in a very narrow range of conditions - you just generate unnecessary procedures, more imaging and stress.
In regard to low field MR, the likes of Siemens Deep Resolve and their low field magnet (0.55T) have taken this a long way, and really improves image quality. It looks useful for imaging around implants and metal.
https://pubmed.ncbi.nlm.nih.gov/30932247/
https://bmjopen.bmj.com/content/12/1/e056572
https://www.siemens-healthineers.com/magnetic-resonance-imag...
https://www.siemens-healthineers.com/magnetic-resonance-imag...
The healthcare system in the west is just plain stupid.
In asia you can walk into a clinic and see a doctor in 30m. Want to get a X-ray or mri etc at hospital? An hour?
Yet in NZ a friends sister has been waiting 4 months now to have her shoulder checked by a specialist.
It’s easier to fly to asia and see a doctor than it is to try see a doctor is a western country.
... If you're rich relative to the general population, then you can go to an expensive private clinic and be seen quickly, yes.
I doubt that's the case for public hospitals and in any case, I've been inside public hospitals in Indonesia, Cambodia, and Vietnam (all for other people's needs, nothing major thankfully). Trust me, you do not want to get treated in one of those and as a westerner you'll never even see them unless you friend gets bitten by a monkey in the Cambodian jungle (for example) and there's no private clinics nearby. I always assumed the casual attitude towards hygiene that Asian people have wouldn't extend to hospitals. I was wrong. There were literally blood stains on the wall in the corridor. The toilet was a hole in the ground with a bucket of water and a scoop to wash your hands and ass with. There was no soap. This wasn't a small hospital, it was the main hospital for a city of about 200k people.
So in systems like the UK, you can walk into a clinic for an MRI, you just have to pay privately for that access, usually covered if you have additional medical insurance. If you want the procedure though nationalised healthcare (NHS), there is a wait.
Health insurance is relatively cheap in the UK compared to the US. A lot of frontline care is done via the NHS contracts with the GPs and the A&E service at the NHS hospitals.
What the hell is "the west" in this context? Do you have experience with every health care system in Europe? Why is "NZ" considered "the west"?
> In asia you can walk into a clinic and see a doctor in 30m.
You can in many places in Europe too.
The flip side of the asian systems can be incredibly overworked and stressed health care workers. I know the service in Taiwan is fantastic (though a bit stressful, was way busier than in Norway), but I also know a nurse working there saying the working conditions are quite awful. They also encourage doing way more testing than necessary. That makes it so testing is pipelined and efficient. You get economies of scale. But doing this is generally considered a very bad idea, because you can get false/misleading diagnosis and undergo unnecessary treatments with potentially bad side effects. Along with all the wasted time and resources that could have been spent working on actually sick people. Which in asian countries can mean that poor people don't get the resources they need to get treatment.
I’m not sure what going on here, as an MR could be done for that tomorrow. However I’d guess that the hold up is a free healthcare system specialist.
If it was a private referral or after an accident, you’ll get seen much sooner. Unfortunately the public system is under pressure.
Deleted Comment
My coworker fell in Costa Rica, hurt her knee and came back and had an MRI the same week. I saw my doctor in mid-March and she ordered an MRI on my neck, it's scheduled for next week. Do some people wait 6 months... yes, those that can wait 6 months are asked to wait 6 months. Those that need it quickly, get it quickly. It's the triage method.
I also had the option to book a private MRI and get it next day. It would have been $650CAD or $483USD. I happen to be in LA when my issue came up and I looked at private MRI. It was way more expensive than the Canadian alternative (however, it could have been LA/Hollywood pricing).
No it's not. I tweaked my knee pretty badly to the point where i would wake up at night from deep sleep due to pain (and I generally have a higher pain tolerance) and the best the doctor could do was prescribe ultra sound cause getting an MRI scan would take 6-7 months.
Even then the earliest appointment I could get for a knee specialist to look at my knee is 6 months.
It really depends on where you're. If you're in the city of Toronto, good luck getting an MRI appointment.
> those that can wait 6 months are asked to wait 6 months
You do know that you're body doesn't decide to wait 6 months for an MRI and stop developing scar tissue right?
There is also this startup - https://qure.ai (Not my employer, I know someone who works there) which has already automated the report generation part for several classes of X Ray / CT / Ultrasound scans. This takes away the requirement of having a trained Radiologist to go through the scan and then create a report, which certainly helps in scaling the availability of such scans - you just need the machine and trained staff to operate it, and are not dependent upon having a doctor with a postgrad in radiology to write the reports.
I just checked and MRI scans in Ireland are about 300eur, while the average monthly wage is about 3300eur.
I know average monthly wage isn't the most robust of measures, but still, looks to me like MRIs are pretty expensive in India relative to people's ability to pay for them.
https://www.youtube.com/watch?v=o25I2fzFGoY
It's insane to belive I have a 3T MRI 2 blocks away from my home.
Now while staying in the US I just pray I don't need have health issue.
That is insane. I dunno about the rest of Germany but in Berlin I can get a knee MRI appointment within two days.
I know this from personal experience, twice, in the last five years.
Because from my personal experience (in Cologne), the interaction goes like this:
Anecdotally, when I dislocated my shoulder in Sweden in 2006, the obly way to even get an MRI was to pay out of pocket.Are you sure about this? Luckily I have not been in need of an MRI for a few years, but I remember having to wait a bit despite having private insurance.
6 months is crazy.
The key advances in this paper are four-fold:
1: Much higher spatial resolution and contrast that is essential to define boundaries of brain regions
2. Much faster scan time using multiple angle acquisitions
3. Accurate segmentation of brain regions of interest even for highly diverse genomes of mice.
4. Accurate alignment and multimodal merging of MRI with 3D lightsheet microscopy from the same animal with fluorescent labelling of multiple CNS cell types.
fMRI in mice is limited in its resolution. To generate time series data we use large sets of isogenic (identical twin) mice. See figure 5 in the paper for a comparison of young and old mice that are identical twins—one a young adult and one quite old adult.
The wait time for government subsidized scans is longer but still not 6 months, and Delhi has far more people than any Canadian city.
This is over-testing, and it leads to over-diagnosis and over-treatment and it is a well recognised form of harm in healthcare.
And are the outcomes of these high net-worth individuals any better than the overall population, in ways that cannot as easily be explained by other factors? I'm not really sold on the idea, and a lot of scientists and doctors aren't either.
Really? I had a lower back MRI appointment at the hospital inside a month or so about 2 years ago, in the middle of the pandemic. Is it the surgery wait itself that takes that long, or did waiting times generally become that worse in that 2 years?
Sure it would be nice that everyone gets an MRI when they need it, but they are expensive pieces of equipment, requiring a very specialized room and staff.
I assume the 9.4T magnet is a very small bore and couldn't take a human. (E.g. https://www.monash.edu/researchinfrastructure/mbi/facilities...).
Of course, lots of important research will benefit from improved scanning resolution on dead subjects -- one might get mislead by the headline, especially if one isn't a research scientist mostly concerned with scanning things you've poisoned or genetically programmed to die in interesting ways.
You start heating up your subjects, so you can’t scan for long.
Does anyone think we understand AD and how best to intervene? This technology is already proving useful. Stay tuned—-this is the tech paper but the application papers are in the wirks.
But I remember when TVs were 640x480-ish and 1920x1080 was amazing. And 4k.
Maybe this won't be for knees in live patients. Maybe it won't be for MRIs in the far-flung corners of the world.
But what if it was for researchers to find elusive parts of the brain? What if the high quantity of data led to better algorithms for current MRIs to make them faster (which probably means cheaper).
I think of pure science and how it doesn't immediately teach us something, but later it answers (or provokes) questions and moves the needle.
Just saying.
That doesn't mean I don't groan at some advances - like any advance that "adds value" to some product by collecting more data without your consent. :)
https://blogs.scientificamerican.com/scicurious-brain/ignobe...
I'd call that 2.5 orders-of-magnitude sharper given how 4 is close to the middle of the log10 range between 1 and 10.
Deleted Comment
I'm wondering if, at some point, I'll be able to have it replaced with a newer material that isn't ferromagnetic (or wouldn't be affected like mine is)
As you noted, this imaging was done at much higher field strengths, ~10T. That sort of field strength is practical at the scale of a system sized for a mouse. For a human, it's likely to never be practical for a whole slew of reasons related to the energy needed, forces on the machine, the increase field size and strength affecting room construction, safety thresholds, etc.
1. That's all they've tested (as you alluded to).
2. It might actually fail at higher fields.
3. Stents act as antennas and can focus RF into unwanted places like cooking the middle of the brain, and it's a lot more likely that the stent is rated for exactly 3T than up to 3T since the relevant frequencies are proportional to field strength, and the associated wavelengths can coincide more or less with the length of the stent (kind of like how the Tacoma Narrows bridge could have experienced a lot more wind, but that particular set of factors was problematic). It _might_ be certified for 0-3T, but that takes additional research to actually do.
4. Any number of other reasons. High fields have a variety of counter-intuitive properties.
Heart stents are rated to 3T as that’s what was tested - it’ll almost certainly be fine at a higher field, but no one has tested it.
I’m an MR tech.
https://www.siemens-healthineers.com/magnetic-resonance-imag...
https://www.ismrm.org/smrt/safety_page/2021.Shellock.Coronar...
We have a complete map of C. elegans, a small worm with 302(ish) neurons. We have a complete wiring diagram of the adult worm’s neurons as well as a “fate map” showing how those connections form during development.
Nevertheless, it’s fairly hard to simulate. Neuromodulators cause all sorts of “spooky action at (synaptic) distance” that the wiring diagram can’t explain, for example.