How to solve the “TI ADS1299 EEG demo kit is not save for humans” problem?

[ralphrmartin]

@ ralphrmartin

Thank you for your answers, this is a good point. Since there is a high agreement that the risk with completely battery powered EEG or ECG devices connected to a battery powered PC is low to damage humans or animals, how has someone to proceed to get a device which is validated for research on humans and animals?

Assuming you are based in Switzerland, this would seem to offer advice on the subject:

https://www.swissmedic.ch/swissmedic/en/home/medical-devices/regulation-of-medical-devices/medical-device-regulation_online-guide.html

You really need to study this carefully, but I note this guidance includes:
----------
"Research with medical devices carried out prospectively on humans is regarded in Switzerland as a clinical trial with medical devices. ... The recruitment of trial participants may commence only when the following conditions are met:

The investigational device must have reached a stage of development that allows it to be used in humans...
All organisational measures must have been taken....
The approval of the cantonal Ethics Committee must have been obtained. This approval is required for all clinical trials without exception. Decisions of other ethics committees (e.g. own committees of hospitals, foreign or private ethics committees) are not sufficient...
The approval of Swissmedic must have been obtained...
----------

Basically, there is a standardised set of rules and procedure you must go through.

[2N3055]

I found this chart that looks kind of interesting in this context:




Source: https://www.asc.ohio-state.edu/physics/p616/safety/fatal_current.html

That is across body volume. If you attach electrodes directly to your chest, together with electrolyte gel, impedances get really low and current flows directly over danger zone...
Not to mention some people have medical implants...
If you don't know exactly what you're doing it's playing with fire..

[Someone]

If you keep everything battery powered, including the laptop that is processing the data,  I would like to hear anyone explain how that could possibly pose a risk from a technical / engineering perspective.

Hello Dunning-Kruger?

This has come up before
https://www.eevblog.com/forum/projects/leakage-current-problem-in-emg-circuit/
Explain how hypothetical device under single fault condition remains safe.

It is the opposite of Dunning-Kruger to ask questions! 

Because you so confidently state as "fact" all the nonsense in your posts, which people who work in this field immediately see as misleading and untrue. Pretending like you are asking a question amongst all that, when its a leading question, is not helping. This is not an area for guessing or relying on uniformed estimates/interpretations.

The standards encompass decades of accumulated knowledge on the subject, they are the quickest/cheapest way to achieve safety and some level of defensible position if something goes wrong. Trying to work it out from first principles or published data, you are almost certain to miss some of the important details or critical pieces of evidence that underlies the safety standards.

You can start pulling up research and trying to piece together a full explanation, but thats usually reserved for when its a new area that isn't fully covered by the standards. At that point its still unproven and needs to head down the pathways that the medical field require, trials and statistically valid proof. Not just picking some unrelated figures you think might be applicable.

[electrolust]

I make the real deal here in the EEVblog, and publish my knowledge about these devices here for free. Maybe someone else gives here something for free, too.

Do i need security tests at all for battery powered devices to use them for research? Is it possible to negotiate a special disclaimer for this purpose to decline liability, for instance when someone uses for instance a TI EEG board and connects it to AC during research experiments?

wow! your judgement is progressively worse! now you just want your subject to accept the risk?

in the US, one cannot waive gross negligence via contract or waiver. if asked, it's ok to sign because the waiver itself is void. (so-called ordinary negligence can be waived away, just not gross negligence)

you can get liability insurance to cover gross negligence in some US states. it's gonna be expensive.

i've no idea about any other country.

the fact that you would use a non-certified device, to me, is gross negligence.

further poor judgement is demonstrated by your request for legal advice from an anonymous internet forum! never take legal advice from non-lawyers! above advice included.

[KaneTW]

How to solve the "not safe for humans" problem? Don't use it on humans. Medical equipment needs to be certified to very strict standards because the allowable probability for failure is extremely small, and for good reason. Money is cheap, human lives aren't.

If you don't have the monetary capabilities to buy a proper certified EEG, don't do research using it.

[Tarloth]

Two years ago an acquaintance, who was a medical student at a university in the interior of my country, contacted me to help him on exactly this matter. He wanted to use the exact same or a similar module sold by adafruit.

His knowledge of electronics was so, so low that he did not even understand why he could not use it in research with living beings. His stubborn mind already imagined conspiracy reasons protecting "the big pharmaceuticals". He never assumed that he was simply not qualified to do what he was doing. He even sent me pictures with the module connected to a patient and powered by a cell phone source of dubious (poor) quality. He told me he was doing it to "save" batteries. What risk could there be if when connecting the phone it did not burned? He possessed a whole series of absolutely wrong thoughts, fallacies in every sense that he nevertheless considered correct, so correct that he perceived himself as ahead of his time.

The reasoning was exactly your reasoning, he had no idea what he had to do, not even what he needed to learn, but nevertheless he supposed to know a lot about the subject and did gave his opinion on everything and believed blindly in his genius, despite being absolutely lacking in knowledge of physics, electronics, biology or sensors. Actually one of those people who have no idea, but by reading something on a blog, they believe they are capable of doing anything and anyone who contradicts it is part of a conspiracy that tries to stop him.

Luckily, and to my peace of mind, he no longer experimented with those modules and changed his career. Hopefully over time he learns that he did not know what he was doing, and if he is really lucky, one day he will learn what he should have done.

Sorry Peter if this is too hard or offensive for you, but it is fantastic and desirable that you play with your inventions, until the moment comes when your ignorance threatening others life. In that case it is the obligation of the rest to tell you that if you have to ask "why is it dangerous" it indicates that you do not have the knowledge to realize it on your own and therefore it is a danger for others that you continues to experience.

I really wish that no one ever gets hurt by your stubbornness.

[george.b]

TI's statement that it's not safe for use on humans has one very simple reason: they did not get the relevant certification for that piece of equipment to be used on humans. Nor should they, since it's intended for evaluating the performance of their ADS1298, not for medical usage. By disallowing its usage on humans, they're waiving liability for any problems that might occur therefrom, and others have already pointed out things that could go wrong. Guess with whom does this liability sit if you decide to go ahead and use it on humans anyway?

[Ice-Tea]

in the US, one cannot waive gross negligence via contract or waiver. if asked, it's ok to sign because the waiver itself is void. (so-called ordinary negligence can be waived away, just not gross negligence)

Probably true anywhere. Have you seen those signs at playgrounds stating "managment can not be held accountable for accidents"?

This is bollocks. If it turns out the playground devices were in ill repair or did not comply to regulations, they are still liable. Even if you manage to find a moron that signs a waiver for your tests, if something happens and they figure out you've been using a device that *explicitly* states it should not be used on humans the burden of proof will be squarely placed on your shoulders.

[SilverSolder]

If you keep everything battery powered, including the laptop that is processing the data,  I would like to hear anyone explain how that could possibly pose a risk from a technical / engineering perspective.

Hello Dunning-Kruger?

This has come up before
https://www.eevblog.com/forum/projects/leakage-current-problem-in-emg-circuit/
Explain how hypothetical device under single fault condition remains safe.

It is the opposite of Dunning-Kruger to ask questions! 

Because you so confidently state as "fact" all the nonsense in your posts, which people who work in this field immediately see as misleading and untrue. Pretending like you are asking a question amongst all that, when its a leading question, is not helping. This is not an area for guessing or relying on uniformed estimates/interpretations.

The standards encompass decades of accumulated knowledge on the subject, they are the quickest/cheapest way to achieve safety and some level of defensible position if something goes wrong. Trying to work it out from first principles or published data, you are almost certain to miss some of the important details or critical pieces of evidence that underlies the safety standards.

You can start pulling up research and trying to piece together a full explanation, but thats usually reserved for when its a new area that isn't fully covered by the standards. At that point its still unproven and needs to head down the pathways that the medical field require, trials and statistically valid proof. Not just picking some unrelated figures you think might be applicable.

I do see where you are coming from,  but it is still a little harsh.  After all, we are still dealing with science / technology, and not religion.   So if an unintuitively low voltage is not safe, it seems reasonable enough, and in the eevblog spirit, to approach the answer in a scientific / technological way instead of "strike down the heathen, that hath entered the holy inner sanctum!"

[2N3055]

If you keep everything battery powered, including the laptop that is processing the data,  I would like to hear anyone explain how that could possibly pose a risk from a technical / engineering perspective.

Hello Dunning-Kruger?

This has come up before
https://www.eevblog.com/forum/projects/leakage-current-problem-in-emg-circuit/
Explain how hypothetical device under single fault condition remains safe.

It is the opposite of Dunning-Kruger to ask questions! 

Because you so confidently state as "fact" all the nonsense in your posts, which people who work in this field immediately see as misleading and untrue. Pretending like you are asking a question amongst all that, when its a leading question, is not helping. This is not an area for guessing or relying on uniformed estimates/interpretations.

The standards encompass decades of accumulated knowledge on the subject, they are the quickest/cheapest way to achieve safety and some level of defensible position if something goes wrong. Trying to work it out from first principles or published data, you are almost certain to miss some of the important details or critical pieces of evidence that underlies the safety standards.

You can start pulling up research and trying to piece together a full explanation, but thats usually reserved for when its a new area that isn't fully covered by the standards. At that point its still unproven and needs to head down the pathways that the medical field require, trials and statistically valid proof. Not just picking some unrelated figures you think might be applicable.

I do see where you are coming from,  but it is still a little harsh.  After all, we are still dealing with science / technology, and not religion.   So if an unintuitively low voltage is not safe, it seems reasonable enough, and in the eevblog spirit, to approach the answer in a scientific / technological way instead of "strike down the heathen, that hath entered the holy inner sanctum!"

You are 100%  correct. It is science and technology. And precisely because of that, what is proven is already proven and documented.
It is not open to opinions, and beliefs. We KNOW some things. And even IF something is wrong, you have to enter the field and first get to know the knowledge and how it came to be, before you decide you can do better.

You cannot just ignore 40 years of research and say, "I don't have a clue about this, but this, here, is inconvenient so I will just ignore it, and will try to find and idiot who will sign it is safe so I can do what I need, and if something goes wrong, it's not my fault"

That is wrong on so many levels..

[SilverSolder]

If you keep everything battery powered, including the laptop that is processing the data,  I would like to hear anyone explain how that could possibly pose a risk from a technical / engineering perspective.

Hello Dunning-Kruger?

This has come up before
https://www.eevblog.com/forum/projects/leakage-current-problem-in-emg-circuit/
Explain how hypothetical device under single fault condition remains safe.

It is the opposite of Dunning-Kruger to ask questions! 

Because you so confidently state as "fact" all the nonsense in your posts, which people who work in this field immediately see as misleading and untrue. Pretending like you are asking a question amongst all that, when its a leading question, is not helping. This is not an area for guessing or relying on uniformed estimates/interpretations.

The standards encompass decades of accumulated knowledge on the subject, they are the quickest/cheapest way to achieve safety and some level of defensible position if something goes wrong. Trying to work it out from first principles or published data, you are almost certain to miss some of the important details or critical pieces of evidence that underlies the safety standards.

You can start pulling up research and trying to piece together a full explanation, but thats usually reserved for when its a new area that isn't fully covered by the standards. At that point its still unproven and needs to head down the pathways that the medical field require, trials and statistically valid proof. Not just picking some unrelated figures you think might be applicable.

I do see where you are coming from,  but it is still a little harsh.  After all, we are still dealing with science / technology, and not religion.   So if an unintuitively low voltage is not safe, it seems reasonable enough, and in the eevblog spirit, to approach the answer in a scientific / technological way instead of "strike down the heathen, that hath entered the holy inner sanctum!"

You are 100%  correct. It is science and technology. And precisely because of that, what is proven is already proven and documented.
It is not open to opinions, and beliefs. We KNOW some things. And even IF something is wrong, you have to enter the field and first get to know the knowledge and how it came to be, before you decide you can do better.

You cannot just ignore 40 years of research and say, "I don't have a clue about this, but this, here, is inconvenient so I will just ignore it, and will try to find and idiot who will sign it is safe so I can do what I need, and if something goes wrong, it's not my fault"

That is wrong on so many levels..

Another observation is that when beginning with basic electronics, the human body may not be the first "breadboard" that you should be playing with! 

Better to begin by blowing up some less valuable things first.

[Peter Gamma]

Several researchers have done research with the TI ADS1299 EEG board. I found one work where it was used to measure EEG currents on a treadmill. Indonesian scientists evaluate the EEG board since several years:

https://www.researchgate.net/publication/337896178_EEG_data_acquisition_system_32_channels_based_on_Raspberry_Pi_with_relative_power_ratio_and_brain_symmetry_index_features

Their works does not make any sense when it cannot be used with humans. How will the Indonesian scientists solve this problem? It does not makes sense to do research on a board which cannot be used with humans. Why does Texas Instruments not make an update so that it can be used with humans? If Texas Instruments does not do it, the Chinese will to the job.

[KaneTW]

If you actually read the paper, they did NOT use it on humans. Only on simulators. Their research showed that the design is potentially valid so that a company can potentially turn it into a product.

In any case, here's how it is: Texas Instruments provides evaluation kits. Medical equipment manufacturers turn it into a safe and certified device. An eval kit is not an actual product.

[Peter Gamma]

The Indonesian scientist used a simulator. But the evaluation was for Acute Ischemic Stroke Identification, Electroencephalography and relative power ratio and brain symmetry index features. As long as there is no device available which is safe for humans, the research results of the Indonesian research scientists cannot be applied, not even when it is not for a medical purpose, but only to use the TI board as a reference device for accuracy tests.

I was successful with another Aliexpress seller. He agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops.

https://www.aliexpress.com/item/4000244594585.html?spm=5261.ProductManageOnline.0.0.1f1a4edfSeZqwP

[TheUnnamedNewbie]

Several researchers have done research with the TI ADS1299 EEG board. I found one work where it was used to measure EEG currents on a treadmill. Indonesian scientists evaluate the EEG board since several years:  [link omitted in quote]
Their works does not make any sense when it cannot be used with humans. How will the Indonesian scientists solve this problem?

Maybe they did not solve the problem, used it anyways. The product works, it just shouldn't be used that way. Just because others did something wrong and unethical, and used it for tests anyways (which I'm not sure they did - I did not read the paper and am not familiar with the field), does not make it right to do so. There are also people who use lead in their paint, or who don't wear seat belts, or do any number of other stupid and dangerous things. That does not make it okay for you to do the same.

Even by powering it with batteries, you could through a fault endanger a person. Next to the more obvious 'shock leading to death' (which is really not even negated by using batteries, as you really don't need a whole load of voltage to pose a risk when attaching stuff with conductive gells to the body!), things like leaking charge into the body (through DC offset) can have dangerous results.

Why does Texas Instruments not make an update so that it can be used with humans? If Texas Instruments does not do it, the Chinese will to the job.

Because of very obvious reasons:
1) They don't want to. They are not a medical devices company, but a chipvendor trying to sell their devices. Investing the research and effort into making this a medically qualified product is just not what they are interested in.
2) They don't want to deal with the legal headaches that come from having something qualified
3) Their goal with this board is to prove that their chip can do cool stuff so they can sell it. That is what evaluation boards are for. Making it safe for humans does not help them in any way - it makes the board more expensive, makes it more complicated obscuring their device performance, etc...

[Fungus]

Why does Texas Instruments not make an update so that it can be used with humans? If Texas Instruments does not do it, the Chinese will to the job.

4) Because Texas Instruments doesn't know exactly how people will use it, how they'll power it, or any modifications or "repairs" somebody might make before attaching it to other people.

If you use this on humans, you're entirely on your own. You can't go online and get other people to approve or support your decision.

(and if you do go online and ask, you should expect a thread exactly like this one...)

[SilverSolder]

[...]  another Aliexpress seller [...]  agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops. [...]

Removing mains power from the equation is a good first step, but it appears you could still get in trouble from the 19V battery in a typical laptop.  There should probably be resistors in the leads going to the pads, to limit the current to something safe in case of the full battery voltage occurring on it while skin resistance is low. 

Disclaimer:  I'm not condoning unsafe practices, but if someone is reading this and absolutely insists on doing crazy stuff, these are among the kinds of things you should be thinking about!

[electrolust]

I was successful with another Aliexpress seller. He agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops.

that won't protect you [from liability] in the slightest

[Someone]

[...]  another Aliexpress seller [...]  agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops. [...]

Removing mains power from the equation is a good first step, but it appears you could still get in trouble from the 19V battery in a typical laptop.  There should probably be resistors in the leads going to the pads, to limit the current to something safe in case of the full battery voltage occurring on it while skin resistance is low. 

Disclaimer:  I'm not condoning unsafe practices, but if someone is reading this and absolutely insists on doing crazy stuff, these are among the kinds of things you should be thinking about!

This is NOTHING to do with 19V (or some other voltage) in a laptop power supply or battery. You really need to stop posting on this topic encouraging the half-arsed dodging of real knowledge.

This cannot be distilled down to a couple of simple dot points to follow, many of the requirements are interdependent, there is no simple answer as to how to make a safe electromedical device.

[SilverSolder]

[...]  another Aliexpress seller [...]  agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops. [...]

Removing mains power from the equation is a good first step, but it appears you could still get in trouble from the 19V battery in a typical laptop.  There should probably be resistors in the leads going to the pads, to limit the current to something safe in case of the full battery voltage occurring on it while skin resistance is low. 

Disclaimer:  I'm not condoning unsafe practices, but if someone is reading this and absolutely insists on doing crazy stuff, these are among the kinds of things you should be thinking about!

This is NOTHING to do with 19V (or some other voltage) in a laptop power supply or battery. You really need to stop posting on this topic encouraging the half-arsed dodging of real knowledge.

This cannot be distilled down to a couple of simple dot points to follow, many of the requirements are interdependent, there is no simple answer as to how to make a safe electromedical device.

It seems to me that even without being an expert on the subject, a moderately bright high school science student should be able to tell you that 19V on the pads of a medical device is a bad idea, given the body resistances quoted in this thread (down to a couple of hundred Ohm) and the max limit of 500uA through the body.

Unless that conclusion is directly wrong...   what's your point?

[Tarloth]

Indonesian Scientists that publish in Researchgate? Researchgate it is not a publication with reference or controls, it is a good place to get something a little better than the average internet garbage, but it does not guarantee any seriousness or certainly. At least use responsible and traditional sources if you're going to make any kind of claim about "scientist" work.

What are you actually expecting from this thread? Someone doing your job because you have no idea what to do?

[Someone]

[...]  another Aliexpress seller [...]  agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops. [...]

Removing mains power from the equation is a good first step, but it appears you could still get in trouble from the 19V battery in a typical laptop.  There should probably be resistors in the leads going to the pads, to limit the current to something safe in case of the full battery voltage occurring on it while skin resistance is low. 

Disclaimer:  I'm not condoning unsafe practices, but if someone is reading this and absolutely insists on doing crazy stuff, these are among the kinds of things you should be thinking about!

This is NOTHING to do with 19V (or some other voltage) in a laptop power supply or battery. You really need to stop posting on this topic encouraging the half-arsed dodging of real knowledge.

This cannot be distilled down to a couple of simple dot points to follow, many of the requirements are interdependent, there is no simple answer as to how to make a safe electromedical device.

It seems to me that even without being an expert on the subject, a moderately bright high school science student should be able to tell you that 19V on the pads of a medical device is a bad idea, given the body resistances quoted in this thread (down to a couple of hundred Ohm) and the max limit of 500uA through the body.

Unless that conclusion is directly wrong...   what's your point?

Because you dumb it down to these simple points. What does the laptop battery have to do with the (hypothetical) electrodes? How many parts/steps/components/interactions are there between those points?

Its distraction, and not relevant to the larger issue. You can keep pushing "clever" points and we'll keep shouting you down for being an idiot.

[SilverSolder]

[...]  another Aliexpress seller [...]  agreed to update the website on Aliexpress with his offer, to write in the documentation to use the EEG board only with battery powered laptops. [...]

Removing mains power from the equation is a good first step, but it appears you could still get in trouble from the 19V battery in a typical laptop.  There should probably be resistors in the leads going to the pads, to limit the current to something safe in case of the full battery voltage occurring on it while skin resistance is low. 

Disclaimer:  I'm not condoning unsafe practices, but if someone is reading this and absolutely insists on doing crazy stuff, these are among the kinds of things you should be thinking about!

This is NOTHING to do with 19V (or some other voltage) in a laptop power supply or battery. You really need to stop posting on this topic encouraging the half-arsed dodging of real knowledge.

This cannot be distilled down to a couple of simple dot points to follow, many of the requirements are interdependent, there is no simple answer as to how to make a safe electromedical device.

It seems to me that even without being an expert on the subject, a moderately bright high school science student should be able to tell you that 19V on the pads of a medical device is a bad idea, given the body resistances quoted in this thread (down to a couple of hundred Ohm) and the max limit of 500uA through the body.

Unless that conclusion is directly wrong...   what's your point?

Because you dumb it down to these simple points. What does the laptop battery have to do with the (hypothetical) electrodes? How many parts/steps/components/interactions are there between those points?

Its distraction, and not relevant to the larger issue. You can keep pushing "clever" points and we'll keep shouting you down for being an idiot.

The power supply voltage (in this case, the laptop battery at 19V) seems a reasonable worst case voltage that is likely to appear on the pads in a worst case  fault condition, assuming you don't do anything crazy that can step that voltage up even higher.

Assuming you actually do have a technical/scientific foundation for your argument, you are failing miserably to present it in an understandable way.

[exmadscientist]

The power supply voltage (in this case, the laptop battery at 19V) seems a reasonable worst case voltage that is likely to appear on the pads in a worst case  fault condition, assuming you don't do anything crazy that can step that voltage up even higher.

Assuming you actually do have a technical/scientific foundation for your argument, you are failing miserably to present it in an understandable way.

And the problem is, 19V is lethal under certain fault conditions for a Type CF Applied Part (IEC 60601), which is what you have.

Let me repeat that in more blunt terms, with emphasis: YOU CAN STILL KILL SOMEONE IF YOU SCREW THIS UP BADLY ENOUGH.

Look. I do this for a living. You can not and should not under any circumstances put that eval kit on a patient. You might put it on yourself -- we have done that before. And in that case, we were literally in the same building as the nearest hospital, knew where the two nearest defibrillators were, had skilled staff monitoring the whole time (both engineers and first responders), nothing was anywhere near the heart, proper isolation was observed at all times, the person with the electrodes on volunteered themself, and we still thought it was crazy dangerous and there was no way we would ever, ever, ever ask one of our research patients to do it.

Face it. Eval kits don't go on patients. Period. End of bloody discussion.

Best to stop this nonsense now before your IRB sees any of it.

[Fungus]

The power supply voltage (in this case, the laptop battery at 19V) seems a reasonable worst case voltage that is likely to appear on the pads in a worst case  fault condition, assuming you don't do anything crazy that can step that voltage up even higher.

Assuming you actually do have a technical/scientific foundation for your argument, you are failing miserably to present it in an understandable way.

And the problem is, 19V is lethal under certain fault conditions for a Type CF Applied Part (IEC 60601), which is what you have.

Let me repeat that in more blunt terms, with emphasis: YOU CAN STILL KILL SOMEONE IF YOU SCREW THIS UP BADLY ENOUGH.

Look. I do this for a living. You can not and should not under any circumstances put that eval kit on a patient. You might put it on yourself -- we have done that before. And in that case, we were literally in the same building as the nearest hospital, knew where the two nearest defibrillators were, had skilled staff monitoring the whole time (both engineers and first responders), nothing was anywhere near the heart, proper isolation was observed at all times, the person with the electrodes on volunteered themself, and we still thought it was crazy dangerous and there was no way we would ever, ever, ever ask one of our research patients to do it.

Face it. Eval kits don't go on patients. Period. End of bloody discussion.

Best to stop this nonsense now before your IRB sees any of it.

I wonder how anybody survives a single session with one of these attached to them:

https://en.wikipedia.org/wiki/Electrical_muscle_stimulation