Author Topic: How to solve the “TI ADS1299 EEG demo kit is not save for humans” problem?  (Read 15856 times)

0 Members and 1 Guest are viewing this topic.

Offline Saskia

  • Super Contributor
  • ***
  • Posts: 2893
  • Country: de
  • you unlock this door with the key of imagination
Working in pharma I can tell that we do spend at least a decade to develop a substance into a medication and do the proper non clinical and clinical safety tests before we get the approval to bring the medication to the market.

At our company there is an army of safety experts, pathologists, toxicologists, etc (far more than 1000 employees) involved just to get approval to prepare the submission for requesting approval to start a clinical study. We have our QA department and extremely regular audits by the EMA, FDA, etc. In these audits we have to explain all of those measures we take to ensure drug safety. Including our SOPs, processes to deal with adverse events, incident management, etc. up to sourcing of components and validation of test equipment. If e.g. a scale does not meet validation criteria, it will get thrown out.

Non-compliance with the rules would get us shut down. Obviously our management is liable for ensuring compliance with the rules, and non-compliance will get employees fired on the spot. No one here wants to catch the blame for any patient mishap.

Medical devices that pose a risk to the patient will undergo similarly rigorous inspections and regulatory approvals.

Can you prove, that

- you have the proper certification and training to be conducting this kind of experiments ?
- you have the proper guidelines to be able to do a validation of the device that you are trying to build ?
- that the manufacturer of your evaluation board certified it for use in this kind of experiments ?
- that you have a setup which does not negate the board manufacturer's certification and specifications for use in medical experiments ?
- did you validate your setup according to good laboratory practice and good clinical practice PRIOR to any experiment taking place ?
- is this validation documented according to good documentation practice ?
- that your lab is actually fit for this kind of experiments ? A validation for labs is also required, see above ....
- that your lab equipment is certified and validated ?
- did you validate the software you are using to take those measurements ? if not, study Computer System Validation & Compliance first.

These are just a few superficial remarks scratching on the surface of this topic.

If the manufacturer does explicitly exclude his evaluation board for use in medical experiments, you should not even think about using it in such manner.



 
The following users thanked this post: SilverSolder, chris_leyson, newbrain

Offline Tarloth

  • Regular Contributor
  • *
  • Posts: 147
  • Country: ar
In the paper, a Rasperri PI is used to aquire data from the TI EEG board and then stored on a SD card on the Rasperri PI. and then further analyzed by a MATLAB based GUI. i did not find any hints in the paper that the Rasperri PI was used for safety reasons. But does the Rasperri PI not solve the problem to prevent current to run from a PC to the TI EEG board to the EEG headseat?

No, RPI not improve the device or turn it "safer", probably turn it more riskie. I did read superficially the paper and scientist didn't test in humans, only it's a proof of concept, they didn't affirm that this configuration it's SAFE, only affirm that it's possible to meassure the signals.

You need to learn basic concepts, first at all signal isolation.

I suppouse that you are not thinking to start a crowd founding campain with another dangerous product,  isn't it?
 

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
It seems like the authors of the papers did not adress the topic safety at all. I saw that safety recommendations for instance for OpenBCI EEG are to use a Wifi setup. The Rasperri PI has wifi transmission, therefore the risk of unwanted current from a PC to the EEG headset could be solved by a Wifi bridge from the PC to the Rasperri PI.
« Last Edit: September 25, 2020, 02:32:27 am by Peter Gamma »
 

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
In the paper, a Rasperri PI is used to aquire data from the TI EEG board and then stored on a SD card on the Rasperri PI. and then further analyzed by a MATLAB based GUI. i did not find any hints in the paper that the Rasperri PI was used for safety reasons. But does the Rasperri PI not solve the problem to prevent current to run from a PC to the TI EEG board to the EEG headseat?

I suppouse that you are not thinking to start a crowd founding campain with another dangerous product,  isn't it?

No, since finding a croud for a research project is hopeless.
 

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
I saw someone else is looking for a crowd for his HackEEG (by Starcat, LLC)

8-32 channel TI ADS1299 Arduino shield for EEG, EMG, and EKG bio-signals acquisition

https://www.crowdsupply.com/starcat/hackeeg

but is this device safe for humans?
 

Offline aristarchus

  • Regular Contributor
  • *
  • Posts: 107
  • Country: 00
I saw someone else is looking for a crowd for his HackEEG (by Starcat, LLC)

8-32 channel TI ADS1299 Arduino shield for EEG, EMG, and EKG bio-signals acquisition

https://www.crowdsupply.com/starcat/hackeeg

but is this device safe for humans?


Their disclaimer on the page you specify tell it all..

"Important Warning

Although HackEEG has some of the features of a medical device, it does NOT have any certifications (FDA, CE, IEC, etc.) and is NOT officially approved for medical or diagnostic use. It is your responsibility to ensure your safety when using the device. Furthermore, you should never power the device from a non-isolated power source."


Don't you read the page that you, yourself, show to us?

 

Offline DrG

  • Super Contributor
  • ***
  • !
  • Posts: 1199
  • Country: us

Don't you read the page that you, yourself, show to us?

It does not appear so to me. He also does not appear to answer questions. There is also very little evidence that any of the considerable advice that has been given has had any real effect. It is a *very* important topic and yet he seems to be stuck on "is it safe", unwilling to pursue any study of even the fundamentals - as has been noted.

It is possible that he was the dentist in Marathon man. Now, I don't know that for a fact, I am just saying it is possible :)


- Invest in science - it pays big dividends. -
 
The following users thanked this post: george.b

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
I've read the pages and saw that the HackEEG can be connected to a Raspberry Pi 4 sending Lab Streaming Layer data over WiFi, which is a similar setup as was used for the TI EEG board in the papers. There is nothing to answer or not to answer anymore.
 

Offline Saskia

  • Super Contributor
  • ***
  • Posts: 2893
  • Country: de
  • you unlock this door with the key of imagination
well, if he tries this on humans, report it to the authorities and watch in awe when they throw the book at him.

 

Offline aristarchus

  • Regular Contributor
  • *
  • Posts: 107
  • Country: 00
This guy is just a shill.
Never really discussed nor listened what others said and finally indirectly advertised a crowdfunding campaign trying to convince us that its just fine to use it.
Clever way indeed!

Suppose that some moderator if not ban him, then at least move this thread to Crowd Funded Projects......
 

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
This is the EEVblog with the slogan «no script, no fear, all opinion». But it seems that some people have a problem with this openness. If so, they are in the wrong blog.

OpenBCI locked some of my forum threads and I can t login to my forum account anymore. For what reason? The same problem in the Garmin forum, and also Matlab deleted all of my posts in the Matlab forum, for reasons I cannot follow.

Companies who are not open to critisism and change, and companies who want to control forum posts in an arbitrary way  will loose their costumers. OpenBCI would be one of the most promising companies for products for meditation training based on scientific findings. But will this happen, if OpenBCI is not open to change, critisism and developement?
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 6126
  • Country: 00
[...]
OpenBCI locked some of my forum threads and I can t login to my forum account anymore. For what reason? The same problem in the Garmin forum, and also Matlab deleted all of my posts in the Matlab forum, for reasons I cannot follow.
[...]

Well, you could be a new Galileo, who is being unfairly punished by the established Church for heresy.   Or, you could just be a heretic!  :D
 
The following users thanked this post: george.b

Offline jogri

  • Frequent Contributor
  • **
  • Posts: 398
  • Country: de
There is a BIG difference between being openminded and being absolutely reckless...

If your posts got deleted in three different forums, maybe the forums aren't at fault here? Maybe, just maybe your mindset has crossed the fine red line from being openminded to being reckless with the physical health and lifes of other people?

Being openminded also means ACCEPTING other world-views (for example the fact that your experiments might hurt other humans and that it violates quite a few ethic guidelines). But no, disregard the probability that fifty people on four different forums telling you that your plan is seriously flawed and dangerous might have a point, they just fear your open mindedness...

At this point, should you go through with your original plan of strapping this thing to other humans, it is nothing more than aggravated battery.

I still had hope that enough people telling you the exact reasons WHY it's a dangerous idea (ffs, we had professionals here who do such things for a living explain it to you in great detail) would change your mind, but oh boy have i been sorely disappointed.

It feels like you want to discuss the advantages of speeding while being absolutely hammered... Sure, you could say that it's just "being openminded towards different world views", but i'm pretty sure that 99% of other forum users would call you reckless beyond belief.
 
The following users thanked this post: george.b

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5216
  • Country: us
Aggravated battery?  That is over the top.  Maybe if strapped to people who have not been informed of possible risks and if there was demonstrated intent to harm.  While the OP is overly dismissive of concerns this kind of response only encourages his opinion that those voicing concerns are overstating the problems.
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13164
  • Country: gb
I have read this entire thread with interest. I totally agree that TI are taking the very sensible position of stating that the evaluation board is not safe for use on humans. They would be foolish to not protect themselves from unsafe projects made by people who mistakenly thought the evaluation board provides all the required safety measure as in some sort of approved ‘building block’ design. As has also been stated, the approval of medical devices for use on humans is understandably heavily controlled to avoid unfortunate incidents as a result of unsafe designs or practices.

I worked with X-Ray equipment and used approved models that were known to be safe. There were relatively inexpensive dental X-Ray heads coming out of China that looked OK on paper but were found to be very unsafe when examined in detail. Simple NE555 exposure timers using cheap components and inadequate X-Ray shielding made them unlicenceable in the UK. They can be , and are, intercepted on arrival by UK Customs. A case of the regulations protecting users which is good. Anyone can build what they claim is a medical device so the public does need protection.

If you look at well known brand sealed lead acid battery data sheets, like those from Yuasa, there is often a warning that they are not intended for medical use and must not be used in medical equipment. This is a standard warning as the batteries have not been licenced for that potentially life supporting role. The fact that the same manufacturer produces an identical looking Medical use battery that behaves identically on my medical battery analyser is of no consequence. We do not know what additional quality control is applied for these “medical grade” batteries but should take the manufacturers advice on what to use in medical scenarios to avoid terrible mistakes where people use the wrong equipment or part in a medical application. We may consider the risk minimal but a human life damaged or lost due to such a decision is unconscionable.

So I sit here thinking what can I say to actually help the OP rather than just tell him “do not use that evaluation board for human research, no matter what you think about it’s safety” ...... well when I want a specialist and very expensive piece of equipment, like my thermal cameras..... I head to either an electronics salvage reseller or eBay.

Why? Well I can buy a £36K piece of equipment for less than £1K and it will likely perform better than any new affordable hobby device of similar cost. In the case of medical equipment, I bought Faxitron X-Ray machines that cost the NHS £55K for around £1K from specialist medical equipment surplus resellers. I tested them and found them to have low hours and perfect safety shielding still in place. They were an amazing buy for me but then I knew what I was doing whenI bought them. The NHS often disposes of equipment that is perfectly functional but has reached 10 years old and is out of support or subject to scheduled replacement.

In my opinion, a viable solution for the OP is to buy a professional medical use licenced EEG from a medical surplus dealer on eBay. Buy the required accessories from the manufacturer and ask for a full service from them. The manufacturers service agents can test the equipment at reasonable cost and declare the unit fit for service. The is is a regular occurrence with surplus medical equipment that is sold by Surplus seller when such equipment is sent to countries that cannot afford to buy new equipment. This is the most sensible and safe path to follow. The equipment will be well designed and safe plus it is licenced. Do not be tempted to buy a used unit and not have it safety tested though as the history of the unit will not be known.

I wish the OP success in his research but ‘playing’ with experimental electronics designs on humans, even volunteers, should be avoided. Any research resulting from such practices will be poisoned by the apparent cavalier attitude of the author towards correct procedures. Even valid research would likely be ignored as a result.

Fraser
« Last Edit: December 19, 2020, 05:53:09 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 
The following users thanked this post: nctnico

Offline george.b

  • Frequent Contributor
  • **
  • Posts: 383
  • Country: br
There is, after all, such a thing as being so open-minded that your brains fall out.
 

Offline Saskia

  • Super Contributor
  • ***
  • Posts: 2893
  • Country: de
  • you unlock this door with the key of imagination
Furthermore our company e.g. would blacklist him as quite obviously the nonconformance to ethic rules might be a floodgate for liability lawsuits. No, thank you.
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13164
  • Country: gb
Just to balance my earlier comments on a medical EEG.... there are EEG ‘toys’ with limited channels available and as they are not for medical use they likely avoid some of the more demanding tests. They are likely considered consumer grade toys or appliances.

https://www.ebay.co.uk/itm/Emotiv-Insight-5-Channel-EEG-Headset-White-/193605334503?_trksid=p2349624.m46890.l49292

Whether such a device is what the OP is researching we do not know.

Fraser
« Last Edit: December 19, 2020, 06:07:14 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13164
  • Country: gb
https://www.emotiv.com/insight/

Specifications state that the RAW EEG output is available as well as the interpreted controls

There are Academic and Student software licence options  :-+

Note this disclaimer though.......

Disclaimer – EMOTIV products are intended to be used for research applications and personal use only. Our products are not sold as medical devices — as defined in EU directive 93/42/EEC. Our products are not designed or intended to be used for diagnosis or treatment of disease.

Fraser
« Last Edit: February 06, 2021, 10:39:30 am by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Peter GammaTopic starter

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ch
  • Meditation Research Institute (CH), Physiologist
    • www.petergamma.org
Do posters here post to answer questions?

Or do posters here post to promote their products, which is spamming and illegal?

Or do posters post here to make other posters stop posting, which is a violation of personality
and illegal?
« Last Edit: February 06, 2021, 04:23:19 am by Peter Gamma »
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5216
  • Country: us
Is the EEVBlog a forum for sellers to promote their products?

Do posters here post to answer questions?

Or do poster posts here to make other posters stop posting, which is a violation of personality
and illegal?

If you expect people to pay attention to your opinions you must pay attention to theirs.  You have received a variety of answers, ranging from people who question your sanity and understanding to those who feel you don't fully understand some of the issues you are dealing with.  And you have summarily rejected them all.  You are expecting people to reject the lessons of their years of experience, and their knowledge of the reactions of others to imposing even slight risks on them or the people around them. 

Perhaps you are wiser than everyone here.  But if so, you need no help from anyone else and shouldn't expect others to sign on merely to cheer on your decisions.  AFAIK no one has made you stop posting.  All anyone has done is repeat their answers when you ask the same question again.  So there is no violation of personality.  No one is forcing you to do anything. 

But think of it this way.  Let us assume something bad happens to someone you experiment on.   It doesn't need to have anything to do with what you are doing.  All that needs to be true is that the person who has the bad thing happen believes that your work was the cause.  They develop a heart arrhythmia, or essential tremor or something else which sounds like it might be related and doesn't have clear causes in most cases.  So you are defending yourself against a legal action.  Do you feel that you could bring this thread (or any of the other threads that you feel have unfairly limited you) into court as evidence of a good faith effort to prevent harm?
« Last Edit: February 06, 2021, 04:24:39 am by CatalinaWOW »
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 16618
  • Country: 00
Companies who are not open to critisism and change, and companies who want to control forum posts in an arbitrary way  will loose their costumers.

Pot, meet kettle....

 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13164
  • Country: gb
Just to be very clear on my post detailing a EEG machine..... I have no connection with the product and provided details in an effort to help the OP and others. If the OP considers my responses self serving in any way, he is mistaken.

You cannot help some people  :-//
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline rsjsouza

  • Super Contributor
  • ***
  • Posts: 5984
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
Interesting; I read most of the thread therefore hopefully I caught it all.
The issue at hand with a complete board is that they cannot (nor want) to put their rear ends on the line for the various medical regulations - which also vary wildly across the world.
You can surely use the ADS1299 device and the suggested circuit as a base to create your own equipment but, as others have said, the hurdles are exposed here.
If your research goal is to capture and validate data for diagnostic purposes,   using anything other than a certified equipment will most probably create challenges to your results in a peer review.
If the goal of your research is the equipment on itself (a low cost equipment for specific markets, for example), then looking at the target countries' specific laws will have a big impact on the direction and design decisions of your research.
As others have said, these trials periods are very long for some reasons; one of them is to create enough track record to increase the reliability and allow for changes during the trial period.
Although I am not an expert, In the past I worked with some medical companies manufacturing equipment and the level of scrutiny is very high.
Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 

Offline Zlotnik

  • Regular Contributor
  • *
  • Posts: 59
  • Country: nl
Disclaimer: I'm not in the medical device industry but another failure critical industry, space. So while some principles I describe below apply, the details I'm sure are quite different and the regulations are -rightly!- much more stringent in the medical world.

I think all this talk fixating on finding mitigations for individual failure modes ("Use a laptop, battery power is safe!" - "Add high-Z electrodes!" - "Healthy humans in benign circumstances are robust" - "Use a RasPi, more safe!" - "Use WiFi, airgapped!") show one fundamental misunderstanding, which makes the medical equipment pros on the forum so queasy to give more help than saying "please don't":

Fixing some issue you found is not the point.
Chances are there's another issue that you didn't fix, and your gizmo is only as safe as the unsafest link.

Instead, you need to do proper Reliability Engineering!
Eg, start with a FMEA, a Failure Mode and Effects Analysis: Basically you go through all components, all functions, all interactions etc in a formal way, and identify all failure modes at all steps. Then you analyse the effects each failure mode has, and (at least in space, where we like to do FMECA instead of FMEA to better direct resources), the criticality of each failure. Most will be trivial. Some will be critical but known in advance. Some will be critical and new.
Only THEN can you argue that fixing failure mode X makes the device more safe, as you KNOW there's no more serious failure mode you didn't fix.

What happens next depends on the requirements you have to comply with. No idea about medical, I'd expect them to be pretty stringent. In our world, typically you require no single point failure shall propagate to a system failure, and call it a day once the system is proven to be SPF free and achieves the required reliability. Sometimes we even waive individual SPFs. Yes, we're cheap. But we also don't mass-produce devices that are applied to warm bodies in uncontrolled conditions, for us it's just a billion or so going up in smoke. Cheap.

I imagine the medical device pros will (rightly) criticise me for over-simplifying, as I'm sure LOADS of critical analysis comes on top, which is what makes all the paperwork so expensive (paper itself being cheap)
In particular, I would not be surprised if in the medical world you also are required to consider multiple failures and do a Fault Tree Analysis (FTA). Much more complex, you think of any undesirable outcome and, using formal logic, work backwards through all ways this can come to pass.

Anyway, by immediately thinking about applying band-aids to individual failure modes, instead of analysing the overall system in a way that allows to capture all failure modes in a traceable way, you demonstrate you have not understood the point the medical pros in this thread were trying to make: you need to make your gizmo certified reliable to plug in humans.

Hence no help in finding better band-aids.

And I'd like to chime in with the medical pros: apply to your own body what you will (I know I have), but DON'T EVER work on test subjects this way.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf