You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have been working on a dedicated open-source hardware, to be usable with the PupilEXT software in the future (see this discussion for details and updates: #53).
Once the first version of it is finished (this fall or next spring), I am planning to publish all the technical documentation and a markdown tutorial for it: It will be open for anyone to modify, rework, extend, make and use, but I would also like to officially certify a basic version (1 camera + 1 IR illuminator version) of the pupillometer/eye tracker device, which I would sell with a minimal profit in a made-to-order fashion. Selling these is expected to cover the expenses of all necessary TÜV certificates, and also to motivate me to stay in the game of developing the open and free hardware and software in the future as well. Besides any personal interest, I think certification would be a very important leap, but it can only be done through an organization (i.e. company), which is the source of the devices produced.
The need of certification
All universities anywhere in the world that can conduct psychophysics experiments have some form of an ethical and safety review process for every experiment planned. Some universities have an institutional review board (IRB), but in many places (for example in my country Hungary) there are no specific IRBs for this, but all universities delegate to a common review body that can give ethical approval to any psychological/psychophysiological (explicitly non-invasive) experiment in the country. This body does not have any engineer members, as their purpose is solely related to reviewing psyhological experiments using e.g. questionnaires, or psychophysiological experiments that are performed with a certified and safety-proven device, e.g. an out-of-the-box EEG device. (At least in my country, there is also a more serious unified review board that can accept invasive and rarely practised methods too (e.g. tDCS, deep brain stimulation), and they consist of medical professionals.)
There is a grey area in law concerning self-assembled (DIY) hardware for experimental use on humans:
In theory (at least in my country, and possibly many others as well, especially EU states), the review board could accept any non-invasive research method, given that "evidence" and known "current state of science" by no question supports the notion that the device is safe for use on humans. But there is uncertainity regarding what level of scientific knowledge the review board actually has, and what they accept as proper evidence. They can mostly decide easily, as marketed devices explicitly state that the devices function in accordance with safety directives, proven by standard procedures carried out by an authorised certifying body. Home-made DIY devices are always in the grey zone: in practice they are usually very basic (e.g. a lever that must be pushed by the participant, and registers the reaction time or strength of the lever push), but can be a matter of debate when there is any harmful effect or emission concerned: e.g. strong strobe light that can cause dizzyness, laser that can harm the eyes, or invisible or barely visible light that can stress the retina without the participant noticing, possibly leading to over-exposure.
The question:
Is a self-made/DIY light source (such as the one in the planned PupilEXT Remote Hardware device project) safe for human experiments?
From the engineer's point of view:
NIR emission from a few little LEDs (as found in today's remote eye-tracker and pupillometer devices), reaching the user's eyeballs from half a meter viewing distance is not a serious concern. Exposure is nothing compared to e.g. looking into a metal furnace (which and similar cases were historically the motivation behind declaring exposure safety limits), or even looking out of the window on a sunny day, or staring into a bonfire when camping in the woods. The only thing that can happen is that the eye retinas warm up over time, but that would need a quite strong light source anyway.
From the layman's point of view:
No matter what an engineer (or a group of them) beleives, solid evidence must be obtained on scientific grounds that prove by some standard way of hypothesis testing in accordance to widely-accepted regulations, that a certain device does not cause harm to the user or to surrounding devices. E.g. causes no electric shock, no over-exposure of light, emits no electromagnetic radiation interfering with pacemakers, does not introduce harmonic noise to the power grid, etc. Of particular importance here: IEC EN 62471 standard declares the procedure for measuring photobiological safety of lamps, and by no surprise, all commercial eye-tracker and pupillometer devices are tested using it, and need to qualify as low-risk.
I inspected the documentation of several regular, branded eye-trackers/pupillometers, and I have found that they comply with the following standards and directives (or their ancestors), regarding how they work and what they are made of:
Basic standards for at least EU:
Eye safety standard (EN 62471)
Electromagnetic compatibility (emission: EN 55022; immunity: EN 61000-4-3, EN 55022, EN 55024)
Additional standards for other regions:
US - FCC part 15
Canada - Interference-Causing Equipment Standard (ICES-003)
Radio and Telecommunications Terminal Equipment Directive (R&TTE) 1999/5/EU
Low Voltage Directive 2014/35/EU
Restriction of Hazardous Substances Directive (RoHS 2.0) 2011/65/EU
Safety for non-invasive research on human subjects: Bearer of responsibility
The ways I suggest to deal with certification:
Level 1,) Leaving it up to the researcher, different for each case:
The online documentation and assembly tutorial of the hardware remote PupilEXT device would contain a „Declaration of Responsibility” form, which may be filled out and signed by the person assembling the device, stating that to their best knowledge the device is by all aspects safe for human use (under the prescribed storage and operating conditions). This document could be attached to the application for ethical approval for every experiment using the device. It would not be a „real” certification by an independent third party, but at least everyone would be responsible for their work.
Level 2,) Harmonized using a company (matter of question now):
I would create an organisation which is stated a manufacturer of Hardware Remote PupilEXT devices: a „for-profit” company. This makes the certification process for a specific hardware model (e.g. the simplest model, containing 1 camera and 1 IR light source, and necessary circuitry) by an internationally recognised certifier third party. These processes are costly (~20-25k EUR for all certifications together) but a minimal profit from selling the devices would cover the expenses (e.g. ~3-4k EUR for each device, leaving ~2k EUR profit per unit). Still „manufacturing” would not be exclusive, and all materials would be open-source: anyone could make them, but they would not have the certification on their device (unless they also make thorugh the certification process). Note that these certification processes must be bound to an organization, even if it is a one-man-band like the company that I could make.
Note that:
It is important that we are only talking about HARDWARE here, that would be compatible with the present and future versions of PupilEXT software. The software is developed using my university resources, but hardware is my „home” hobby project and the starting version would be my intellectual property. However I will publish editable 3D model files and an extensive tutorial on how to make the devices. Also, anyone would be free to make them, edit them, extend them, etc. But the certified version would be the starter one - at least until someone (me or anybody else) is willing to make through another certification process for a newer version in future.
Once all certificates are obtained, devices would be made-to-order within a few months.
I was also thinking about a non-profit organization or foundation, but it seems to me it would be nearly impossible to create and maintain efficiently. Also pupillometry may stay too niche for such an effort, and it would be hard to fulfil criteria of non-profit organizations (organizing public events etc.), so a keep-it-simple company would be necessary to maintain quality of work in the long run.
This would need lots of paperwork, also including certification, taxation, supply chain certification – which would be definitely more than a hobby for me, so some miminal profit would be made once the price of all certifications is covered.
It is very important that I do not have any "startup" idea in mind, or making a considerable profit from all this. This company would be mainly about making a certified version the device and providing a basic monetary motivation for me to stay in the game also in the future, even if I opt to work in industry after finishing my PhD and leaving the academic field.
Key arguments and considerations:
These would be certified devices, just like any other lab equipment. CE, FCC and usage safety proven.
It would provide an extra level of replicability for studies conducted with these devices, as they would be made with the very same components, tools, methods, so each device would be closely identical, making it a proper scientific instrument.
I would make through the company administration and certification of the device (it is a luck that I studied some management too), though tax regulations are a bit unfriendly in my country, giving this solution a hard-start.
~3-4k EUR is very friendly price for such a device. However I am feeling a little guilt that this could have a negative effect on current scientific eye-tracking/pupillometry company scene due to the low price of this device. I would not like to cause too harsh competition with the price, but the goal would be to keep it as accessible to people as possible, while covering expenses for all certification procedures. In my opinion, the field of pupillometry in middle and eastern Europe is very much lagging behind western Europe due to the very high prices of devices, but such a cheaper device would probably help the problem - while also sparing the additional resources of researchers from assembling it for themselves, and providing a certified version for legal use in experiments.
At least a minimal monetary compensation for the work done already and planned in future - It would also indirectly motivate me to keep taking part in developing the (free and non-profit) software also in the future, in my free time.
What do you think?
If you are reading this as a part of the community currently developing the PupilEXT software:
What are your thoughts?
Would you agree with my propositions?
What do you think would need futrher consideration?
Additionally, if you are a researcher who would like to use the PupilEXT software with dedicated hardware:
Would you rather build the device (ordering parts from several suppliers, obtaining tools, assembling the device, and filling out the declaration of responsibility form, for a BOM price of ~1-1,5k EUR) or buy it ready-made (~3-4k EUR)?
Do you think the price would be fair?
How many of these devices would you be interested in building or buying?
(I deliberately did not start a "Polls" discussion, so that anyone can share their thoughts freely in detail.)
Also, I would like to specifically ask @BZandi as the maintainer of the software project: What are your thoughts on this matter?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
In a nutshell
Hi Everyone!
I have been working on a dedicated open-source hardware, to be usable with the PupilEXT software in the future (see this discussion for details and updates: #53).
Once the first version of it is finished (this fall or next spring), I am planning to publish all the technical documentation and a markdown tutorial for it: It will be open for anyone to modify, rework, extend, make and use, but I would also like to officially certify a basic version (1 camera + 1 IR illuminator version) of the pupillometer/eye tracker device, which I would sell with a minimal profit in a made-to-order fashion. Selling these is expected to cover the expenses of all necessary TÜV certificates, and also to motivate me to stay in the game of developing the open and free hardware and software in the future as well. Besides any personal interest, I think certification would be a very important leap, but it can only be done through an organization (i.e. company), which is the source of the devices produced.
The need of certification
All universities anywhere in the world that can conduct psychophysics experiments have some form of an ethical and safety review process for every experiment planned. Some universities have an institutional review board (IRB), but in many places (for example in my country Hungary) there are no specific IRBs for this, but all universities delegate to a common review body that can give ethical approval to any psychological/psychophysiological (explicitly non-invasive) experiment in the country. This body does not have any engineer members, as their purpose is solely related to reviewing psyhological experiments using e.g. questionnaires, or psychophysiological experiments that are performed with a certified and safety-proven device, e.g. an out-of-the-box EEG device. (At least in my country, there is also a more serious unified review board that can accept invasive and rarely practised methods too (e.g. tDCS, deep brain stimulation), and they consist of medical professionals.)
There is a grey area in law concerning self-assembled (DIY) hardware for experimental use on humans:
In theory (at least in my country, and possibly many others as well, especially EU states), the review board could accept any non-invasive research method, given that "evidence" and known "current state of science" by no question supports the notion that the device is safe for use on humans. But there is uncertainity regarding what level of scientific knowledge the review board actually has, and what they accept as proper evidence. They can mostly decide easily, as marketed devices explicitly state that the devices function in accordance with safety directives, proven by standard procedures carried out by an authorised certifying body. Home-made DIY devices are always in the grey zone: in practice they are usually very basic (e.g. a lever that must be pushed by the participant, and registers the reaction time or strength of the lever push), but can be a matter of debate when there is any harmful effect or emission concerned: e.g. strong strobe light that can cause dizzyness, laser that can harm the eyes, or invisible or barely visible light that can stress the retina without the participant noticing, possibly leading to over-exposure.
The question:
Is a self-made/DIY light source (such as the one in the planned PupilEXT Remote Hardware device project) safe for human experiments?
From the engineer's point of view:
NIR emission from a few little LEDs (as found in today's remote eye-tracker and pupillometer devices), reaching the user's eyeballs from half a meter viewing distance is not a serious concern. Exposure is nothing compared to e.g. looking into a metal furnace (which and similar cases were historically the motivation behind declaring exposure safety limits), or even looking out of the window on a sunny day, or staring into a bonfire when camping in the woods. The only thing that can happen is that the eye retinas warm up over time, but that would need a quite strong light source anyway.
From the layman's point of view:
No matter what an engineer (or a group of them) beleives, solid evidence must be obtained on scientific grounds that prove by some standard way of hypothesis testing in accordance to widely-accepted regulations, that a certain device does not cause harm to the user or to surrounding devices. E.g. causes no electric shock, no over-exposure of light, emits no electromagnetic radiation interfering with pacemakers, does not introduce harmonic noise to the power grid, etc. Of particular importance here: IEC EN 62471 standard declares the procedure for measuring photobiological safety of lamps, and by no surprise, all commercial eye-tracker and pupillometer devices are tested using it, and need to qualify as low-risk.
I inspected the documentation of several regular, branded eye-trackers/pupillometers, and I have found that they comply with the following standards and directives (or their ancestors), regarding how they work and what they are made of:
Basic standards for at least EU:
Additional standards for other regions:
Proving accordance to directives for EU:
Safety for non-invasive research on human subjects: Bearer of responsibility
The ways I suggest to deal with certification:
Level 1,) Leaving it up to the researcher, different for each case:
The online documentation and assembly tutorial of the hardware remote PupilEXT device would contain a „Declaration of Responsibility” form, which may be filled out and signed by the person assembling the device, stating that to their best knowledge the device is by all aspects safe for human use (under the prescribed storage and operating conditions). This document could be attached to the application for ethical approval for every experiment using the device. It would not be a „real” certification by an independent third party, but at least everyone would be responsible for their work.
Level 2,) Harmonized using a company (matter of question now):
I would create an organisation which is stated a manufacturer of Hardware Remote PupilEXT devices: a „for-profit” company. This makes the certification process for a specific hardware model (e.g. the simplest model, containing 1 camera and 1 IR light source, and necessary circuitry) by an internationally recognised certifier third party. These processes are costly (~20-25k EUR for all certifications together) but a minimal profit from selling the devices would cover the expenses (e.g. ~3-4k EUR for each device, leaving ~2k EUR profit per unit). Still „manufacturing” would not be exclusive, and all materials would be open-source: anyone could make them, but they would not have the certification on their device (unless they also make thorugh the certification process). Note that these certification processes must be bound to an organization, even if it is a one-man-band like the company that I could make.
Note that:
It is important that we are only talking about HARDWARE here, that would be compatible with the present and future versions of PupilEXT software. The software is developed using my university resources, but hardware is my „home” hobby project and the starting version would be my intellectual property. However I will publish editable 3D model files and an extensive tutorial on how to make the devices. Also, anyone would be free to make them, edit them, extend them, etc. But the certified version would be the starter one - at least until someone (me or anybody else) is willing to make through another certification process for a newer version in future.
Once all certificates are obtained, devices would be made-to-order within a few months.
I was also thinking about a non-profit organization or foundation, but it seems to me it would be nearly impossible to create and maintain efficiently. Also pupillometry may stay too niche for such an effort, and it would be hard to fulfil criteria of non-profit organizations (organizing public events etc.), so a keep-it-simple company would be necessary to maintain quality of work in the long run.
This would need lots of paperwork, also including certification, taxation, supply chain certification – which would be definitely more than a hobby for me, so some miminal profit would be made once the price of all certifications is covered.
It is very important that I do not have any "startup" idea in mind, or making a considerable profit from all this. This company would be mainly about making a certified version the device and providing a basic monetary motivation for me to stay in the game also in the future, even if I opt to work in industry after finishing my PhD and leaving the academic field.
Key arguments and considerations:
What do you think?
If you are reading this as a part of the community currently developing the PupilEXT software:
Additionally, if you are a researcher who would like to use the PupilEXT software with dedicated hardware:
(I deliberately did not start a "Polls" discussion, so that anyone can share their thoughts freely in detail.)
Also, I would like to specifically ask @BZandi as the maintainer of the software project: What are your thoughts on this matter?
Thank you for your time!
Beta Was this translation helpful? Give feedback.
All reactions