Hi, how do you compute the strength of EMI that can clip 1mV signal from typical EKG, EMG and any standard biopotential measuring equipment? ?



Supposed the green sine wave is the 1mV signal. And the red sine wave is the EMI. What strength of the EMI before it can clip the 1mV signal such that the output (at bottom) would become distorted?


Supposed the green sine wave (top) is the 1mV signal. And the red sine wave is the RFI (radiated). What strength of the RFI before it can clip the 1mV signal such that the output (at bottom) would become distorted?

I'm talking of radiated RFI interference, not from EMI from power supply lines because I used batteries, so it is disconnected from main.

Also I most often see EMI filters to filter power lines. Can these be used also to filter radiated interferences?

 

Hi, how do you compute the strength of EMI that can clip 1mV signal from typical EKG, EMG and any standard biopotential measuring equipment? ?

View attachment 316138

Supposed the green sine wave is the 1mV signal. And the red sine wave is the EMI. What strength of the EMI before it can clip the 1mV signal such that the output (at bottom) would become distorted?


Supposed the green sine wave (top) is the 1mV signal. And the red sine wave is the RFI (radiated). What strength of the RFI before it can clip the 1mV signal such that the output (at bottom) would become distorted?

I'm talking of radiated RFI interference, not from EMI from power supply lines because I used batteries, so it is disconnected from main.

Also I most often see EMI filters to filter power lines. Can these be used also to filter radiated interferences?

If you are worried about high frequency interference, that tends not to be a problem for a number of reasons: first is that for an electrode connected to a patient, the patient will shunt any rf energy to ground (remembering that your wires should be shielded and any rf that gets onto your wires will not have a lot of energy in it anyway) and second, your Instrumentation Amplifier (IA) doesn't have the bandwidth for anything beyond a few kHz or maybe a few 10s of kHz but nowhere near rf.
Your biggest interferes will be mains frequency and bio signals you are not wanting to see.
Sitting in a typical room, there are lights overhead operating at mains frequency and voltage and the floor, unless you are in a fully timber house, the floor is likely concrete and effectively an earth plane. So if the lights are at 240V and the ceiling is 2.4m then the electric field is around 100V/m (gross simplification to demonstrate the idea). A typical IA has a very high input impedance so this 100V/m field can impose itself very easily on the patient and end up on your IA inputs where it will very easily exceed the input common mode range of the IA and you get garbage waveforms.
There are a umber of ways of dealing with this situation. An ECG (which is way easier than an EEG) uses a "right leg drive" or RLD which is an average of the left leg and right and left arm signals which is then driven onto the right leg electrode. This does two things, first is it stops the patient 'floating' to whatever the prevailing electric field can get them to, and second it forces the patient down to a voltage within the IA common mode input range. There is a little more to this though. The average of the 3 electrodes I just described is known as the Central or Wilson terminal and notionally represents a point roughly in the centre of the chest of the patient. In the ECG the inputs to the IA's are all referenced to this terminal in a typical 3 lead system. So it is almost like having an electrode inside the patient. This has implications for the waveforms but that is not so relevant to your question at this stage.
In and ECG the very high CMRR of the IA's and the way that the RLD is derived and driven means the IA's can reject the mains frequency interference which can be several volts in amplitude while the ECG signal is only 1 or 2mV.
The other issue is muscle noise which manifests itself as a relatively large but very slow (for a stationary patient) drift of the ECG waveform called a baseline drift. The baseline drift can easily be 3 or more times the ECG signal amplitude and can also be very annoying as you try to track this ECG signal wandering up and down the screen.
I mention these ECG things because I think you will have similar issues for your EEG. I am no expert in EEG systems but suspect similar principles apply. Your version of the Central terminal you could arrange to be at any point within the brain box by averaging the signals from several electrodes. You will need to 'reference' the patient either with a driven electrode of an electrode from your local 0V to an electrode directly (since you are battery powered, but beware if you have a physical comms connection for example, because then you do have exposure to mains powered circuits and your battery power will mean nothing).
I think you will have to just give this a go to see what baseline drift you get. If you implement a reference of one type or another, a Central terminal or 0V reference, shielded cables and use a good IA you should get proper signals and good enough to then evaluate baseline drift, if it happens and any other nuances you find. But essentially, I think that will work ok for a first hit.
I really can't see any rf getting into your system and giving you grief, but if I am wrong about that, please let me know so I learn and maybe I can help you get rid of it too. It won't be there It will be fine!

 

If you are worried about high frequency interference, that tends not to be a problem for a number of reasons: first is that for an electrode connected to a patient, the patient will shunt any rf energy to ground (remembering that your wires should be shielded and any rf that gets onto your wires will not have a lot of energy in it anyway) and second, your Instrumentation Amplifier (IA) doesn't have the bandwidth for anything beyond a few kHz or maybe a few 10s of kHz but nowhere near rf.
Your biggest interferes will be mains frequency and bio signals you are not wanting to see.
Sitting in a typical room, there are lights overhead operating at mains frequency and voltage and the floor, unless you are in a fully timber house, the floor is likely concrete and effectively an earth plane. So if the lights are at 240V and the ceiling is 2.4m then the electric field is around 100V/m (gross simplification to demonstrate the idea). A typical IA has a very high input impedance so this 100V/m field can impose itself very easily on the patient and end up on your IA inputs where it will very easily exceed the input common mode range of the IA and you get garbage waveforms.
There are a umber of ways of dealing with this situation. An ECG (which is way easier than an EEG) uses a "right leg drive" or RLD which is an average of the left leg and right and left arm signals which is then driven onto the right leg electrode. This does two things, first is it stops the patient 'floating' to whatever the prevailing electric field can get them to, and second it forces the patient down to a voltage within the IA common mode input range. There is a little more to this though. The average of the 3 electrodes I just described is known as the Central or Wilson terminal and notionally represents a point roughly in the centre of the chest of the patient. In the ECG the inputs to the IA's are all referenced to this terminal in a typical 3 lead system. So it is almost like having an electrode inside the patient. This has implications for the waveforms but that is not so relevant to your question at this stage.
In and ECG the very high CMRR of the IA's and the way that the RLD is derived and driven means the IA's can reject the mains frequency interference which can be several volts in amplitude while the ECG signal is only 1 or 2mV.
The other issue is muscle noise which manifests itself as a relatively large but very slow (for a stationary patient) drift of the ECG waveform called a baseline drift. The baseline drift can easily be 3 or more times the ECG signal amplitude and can also be very annoying as you try to track this ECG signal wandering up and down the screen.
I mention these ECG things because I think you will have similar issues for your EEG. I am no expert in EEG systems but suspect similar principles apply. Your version of the Central terminal you could arrange to be at any point within the brain box by averaging the signals from several electrodes. You will need to 'reference' the patient either with a driven electrode of an electrode from your local 0V to an electrode directly (since you are battery powered, but beware if you have a physical comms connection for example, because then you do have exposure to mains powered circuits and your battery power will mean nothing).
I think you will have to just give this a go to see what baseline drift you get. If you implement a reference of one type or another, a Central terminal or 0V reference, shielded cables and use a good IA you should get proper signals and good enough to then evaluate baseline drift, if it happens and any other nuances you find. But essentially, I think that will work ok for a first hit.
I really can't see any rf getting into your system and giving you grief, but if I am wrong about that, please let me know so I learn and maybe I can help you get rid of it too. It won't be there It will be fine!

Thanks. But my main application is EMG with 1mV to 5mV input, not ECG nor EEG. In EMG, there is no RLD (Right Left Drive). Instead, you put the ground lead at the wrist or nearby, then the Active and Reference (or + and - of the differential inputs) to the muscle surface in question. Can it get the electrical field of the person without the RLD used in ECG? I'm just beginning to use it so don't have much experience in the waveforms or noises. Please describe the above noise production using EMG use only. RF still won't disturb it? The amplifier is only powered by battery so no problem about direct line intereference from 60Hz AC. Thanks.

 

Thanks. But my main application is EMG with 1mV to 5mV input, not ECG nor EEG. In EMG, there is no RLD (Right Left Drive). Instead, you put the ground lead at the wrist or nearby, then the Active and Reference (or + and - of the differential inputs) to the muscle surface in question. Can it get the electrical field of the person without the RLD used in ECG? I'm just beginning to use it so don't have much experience in the waveforms or noises. Please describe the above noise production using EMG use only. RF still won't disturb it? The amplifier is only powered by battery so no problem about direct line intereference from 60Hz AC. Thanks.

Sorry, I read EMG / EKG and thought EEG. Ok, well your signals should be (in relative terms) big bold and beautiful. The muscle noise problem with an ECG in your case is not noise, it is the signal!
Sounds like you already have a reference node in place so stick with that is my advice. It will help to keep the signals into the IA within the IA common mode input range. You still need an IA but with both inputs connected to electrodes that span the muscle or part of the muscle, you are trying to detect signals in.
Don't worry about EMG if you are not looking to detect signals in the brain box. Your 1mV to 5mV signal levels sound right to me for an average sized muscle.
Do worry some about electrical safety and remember that this device is connected to a patient so if something goes wrong there is zero chance they can disconnect from the danger. I know you are running o batteries, my point before was only that if any part of your circuit is connected to other equipment (a serial port is an obvious possibility) and that equipment is mains powered then unless that equipment is rated for medical use then there is that slight risk of electric shock. Mains interference was not my concern from that angle.
I do not believe you will have rf interference issues and if you keep your IA inputs clean and balanced and your IA has sufficient CMRR you should be good to go. Using two electrodes that are essentially the same and some gel (salty glycerine / water based lube) for the signals and the biggest electrode you can for the reference, also with gel.
RF really should not be a problem unless there is a transmitter in desperately close proximity to the patient. Even then, I'd be surprised. If you do get a lot of noise, try reading up on IA and maintaining the CMRR. It is a big topic and worth while for you to know about given the project you are working on. You might get wat looks like noise when it is actually your electrodes picking up the signals from lots of muscles. Upper arms and legs the muscles are larger and more discrete but in the core they are many and packed in tight. Forearms might need careful electrode placement to detect a single muscle in isolation.
Let me know how you go.

 

Sorry, I read EMG / EKG and thought EEG. Ok, well your signals should be (in relative terms) big bold and beautiful. The muscle noise problem with an ECG in your case is not noise, it is the signal!
Sounds like you already have a reference node in place so stick with that is my advice. It will help to keep the signals into the IA within the IA common mode input range. You still need an IA but with both inputs connected to electrodes that span the muscle or part of the muscle, you are trying to detect signals in.
Don't worry about EMG if you are not looking to detect signals in the brain box. Your 1mV to 5mV signal levels sound right to me for an average sized muscle.
Do worry some about electrical safety and remember that this device is connected to a patient so if something goes wrong there is zero chance they can disconnect from the danger. I know you are running o batteries, my point before was only that if any part of your circuit is connected to other equipment (a serial port is an obvious possibility) and that equipment is mains powered then unless that equipment is rated for medical use then there is that slight risk of electric shock. Mains interference was not my concern from that angle.
I do not believe you will have rf interference issues and if you keep your IA inputs clean and balanced and your IA has sufficient CMRR you should be good to go. Using two electrodes that are essentially the same and some gel (salty glycerine / water based lube) for the signals and the biggest electrode you can for the reference, also with gel.
RF really should not be a problem unless there is a transmitter in desperately close proximity to the patient. Even then, I'd be surprised. If you do get a lot of noise, try reading up on IA and maintaining the CMRR. It is a big topic and worth while for you to know about given the project you are working on. You might get wat looks like noise when it is actually your electrodes picking up the signals from lots of muscles. Upper arms and legs the muscles are larger and more discrete but in the core they are many and packed in tight. Forearms might need careful electrode placement to detect a single muscle in isolation.
Let me know how you go.

EMG and ECG has both similar 1mV to 10mV signal. How come ECG needs the RLD (Right Leg Drive) even though it has reference lead too? Aren't there ECG where the ground is used instead of the RLD?

Why doesn't EMG use the concept of RLD?

For IA with singular output. Can it work with RLD? If yes. I'll try to add RLD to my amplifier. Or does RLD only compatible with IA that has differential output too?

I read some articles like :

" Connecting the electrode directly to the common is undesirable for two reasons. 1) If the circuit is not isolated, dangerous currents could flow through the third electrode. 2) A poor electrode contact may present up to 100 kΩΩ of resistance between the patient and the common. The most common and effective use of the third electrode is to connect it to a driven-right-leg circuit [2], [3]. This circuit overcomes both of the problems listed above. It reduces the effective electrode resistance by several orders of magnitude, and it allows only a safe amount of current to flow through the third electrode. "?

Btw.. are you familiar with the gtec USBAMP?

 

EMG and ECG has both similar 1mV to 10mV signal. How come ECG needs the RLD (Right Leg Drive) even though it has reference lead too? Aren't there ECG where the ground is used instead of the RLD?

Why doesn't EMG use the concept of RLD?

For IA with singular output. Can it work with RLD? If yes. I'll try to add RLD to my amplifier. Or does RLD only compatible with IA that has differential output too?

I read some articles like :

" Connecting the electrode directly to the common is undesirable for two reasons. 1) If the circuit is not isolated, dangerous currents could flow through the third electrode. 2) A poor electrode contact may present up to 100 kΩΩ of resistance between the patient and the common. The most common and effective use of the third electrode is to connect it to a driven-right-leg circuit [2], [3]. This circuit overcomes both of the problems listed above. It reduces the effective electrode resistance by several orders of magnitude, and it allows only a safe amount of current to flow through the third electrode. "?

Btw.. are you familiar with the gtec USBAMP?

You have many questions
Not familiar with the Gtec bioamp.
The RLD and a 'reference' electrode can be similar in some respects but the RLD takes the concept to the next level. A reference electrode (and as far as I know you can have a reference electrode connected OR an RLD electrode, but both simultaneously would be weird and I don't know why it would be done) will serve to stop the patient potential from moving beyond the IA CMRR range. It is effectively trying to hold the big sack of salty water at a comfortable level for the IA inputs.
The RLD is a derived signal based on the average of the three sensed electrodes which is otherwise known as the Central terminal (or Wilson terminal in some texts) and this represents the potential at a specific point in the body, usually the centre of the chest. By driving the same signal onto the right leg the idea is to put the whole body at roughly the same potential but also to take any noise pickup in the three sensed electrodes and cancel that noise out (to a degree). Your quoted article points out that the RLD, doing what it does, also compensates for the resistance between the electrodes and the body (internally). I think their claim is a bit wild to be honest but their claim is the ultimate goal of the RLD. Also, while I think of it, their concern about dangerous currents in reference electrodes is, I think, a concern about isolation, which is not a concern for you based your circuit being battery powered (and having no connection to a mains powered device that does not have medical grade isolation from the mains.) What you are working on is classified as a BF type patient connection (Body connection, Floating) with maximum leakage currents that are really quite small. I don't remember those numbers off the top of my head as what I do is classed differently and there is only so much stuff I can keep in my head. If you want to know those numbers or other regulatory details and if you are not already aware of this, then, 60601 is your friend and one stop shop (once you include the relevant collateral and particular standards in the 60601-1 and 60601-2 series.)
In a 3 lead ECG the three electrode signals are derived and amplified with respect to the Central terminal before they are combined to produce the lead I, II and III waveforms.
In your case, if you have several electrodes all focused on the same area and you needed to eliminate a type of noise that can be reduced or eliminated in this fashion, then a driven electrode may be what you need. But if you are sensing a single muscle then I think if you had three electrodes arranged linearly with the centre electrode being your common or 0V node you should get good results.
A trick I have used before, to eliminate the 3rd electrode is to terminate your two sense inputs to your local 0V via 10Meg resistors. I know I have made this work as an interim measure for an ambulatory ECG (or Holter monitor) just beware the impact on the CMRR and the electrode contact needs to be pretty damn good but the technique is otherwise sound enough especially since your signals are far bigger than typical ECG signals.
If you want to know more about RLD, I recommend reading data sheets and app notes for TI bio front end chips. You will find a far more complete explanation of the RLD and how the IA's are arranged and how signals are derived for an ECG with RLD.
Further to the basic principles of the RLD, the concept of a Central terminal is a powerful one. To have a virtual electrode at any arbitrary point within the body without first needing a surgical procedure to put it there is really really useful. Accurate placement of this virtual electrode is really really difficult though depending of course on exactly where you want it and how accurately you need it placed. There are many factors that may cause this virtual electrode to wander around a bit too. Twelve lead ECG use the principle as well and 3D mapping systems (tomography) as used in cardiology use it as well as ultrasound and manual feedback from fluoroscopy as well. I'm just saying the principle is out there and well documented and described in tons of literature if you wanted to deep dive into it.
If you were doing EEG (brain wave stuff) I think it would be an incredibly useful technique to use to isolate signals at arbitrary points inside the brain box. I have not done that kind of work and it is probably either common place or 'what they used to do years ago, just after the dinosaurs died out'. I don't follow that particular field closely at all.
I think that covers your questions.

 

Thanks for all the information. My last questions

My application is neither ECG nor EEG but combinations of both so my ultimation resolution is 10 microVolt. So I must think of interferences that can affect microvolts.

My question is connected to the following.

https://www.sciencedirect.com/science/article/abs/pii/S0165027014001666

"Active electrodes are often billed as enabling this mode of data collection, and this feature is taken as an additional justification of their relatively high cost. The follow example illustrates why active electrodes should have this property. In passive amplification systems, interference currents that come from the main power and “wirelessly” couple to the participant and to the electrode wires (capacitive coupling), multiply by the interelectrode impedance that gives as a result the interference voltage that corrupts the EEG signal before it gets to the amplifier (Metting Van Rijn et al., 1990). A typical interference current is of the order of 20 nA, which, given an interelectrode impedance of even 10 kΩ, yields an interference voltage of 200 μV, a magnitude capable of drowning the EEG signal being measured, which is normally between 10 and 100 μV (Aurlien et al., 2004)."

Is it true? When I tried EEG and using unshieded wire and slippers. I dont recall seeing the capacitive coupling. In the article,
200uV interference voltage can totally down any 10uV signal! What you make of it? My application deal with 10 microVolt.

When you mentioned shielded wire. Can it totally eliminate any interference current of 20nA or so? How much interference current is retained using shielded wire? Do you have any data? I got the Metting Van Rijn reference and saw this figure:



When you wear slippers. The interference current of 20nA can still course thru the wire? if the wire is shielded. What would be the usual value of the interference current. Zero?

They are last questions. Thanks!

 

I just had a thought which may or may not be relevant.
If you have never done this before and want to get an idea of what to expect and how things work, I would strongly advocate you get a good quality IA, the AD8227 for example but not a bio front end chip to start with as they can take a lot of setting up to get any signal out and to start with you will want to limit the possibilities for the thing not to work.
Hook up your IA, start with a gain of say 100 to 200 or 300 at the most, add a couple of 10meg resistors to ground on your IA, a CRO on the output (50 to 100ms/div) and stick a couple of electrodes (I used pieces of copper foil in my first experiments) one under your left arm and the other about centre of your chest and behold the magic of biomedical electronics!
If you are not beholding any magic then at least the circuit is very very simple to modify and understand etc etc and should give something quite quickly and then you can start testing your ideas.
I think this is the best way to start on your journey into biomed. It can't hurt but it may well help.

 

I once tried the Brainmaster EEG version 1 and could see the effects of capacitive coupling. It can respond to my hands waving even when a meter away unconnected. But im not sure when connected that it can do a 200uV override of the 10uV signal.

Im asking bec im deciding whether to buy a $2000 active electrode system. So the article I quoted in last message is really true? Pls confirm if the interference current can really be 20nA and after interacting with 10kOhm resistor can produce 200uV noise. and it cant be eliminated by any shielded wire. Bec if its true then ill loan money to get the $2000. Thanks.

 

I once tried the Brainmaster EEG version 1 and could see the effects of capacitive coupling. It can respond to my hands waving even when a meter away unconnected. But im not sure when connected that it can do a 200uV override of the 10uV signal.

Im asking bec im deciding whether to buy a $2000 active electrode system. So the article I quoted in last message is really true? Pls confirm if the interference current can really be 20nA and after interacting with 10kOhm resistor can produce 200uV noise. and it cant be eliminated by any shielded wire. Bec if its true then ill loan money to get the $2000. Thanks.

In a previous posing in the thread you said: " my main application is EMG with 1mV to 5mV input, not ECG nor EEG " and now you are saying EEG and 10uV signals and quote a circuit diagram that makes allowance for patient connections to be earthed.
Ok, I will be frank, please do not be offended, since you mislead me about the application a lot of my responses are irrelevant in many ways so it was a waste of my time.
First point, NEVER earth patient connections. Even if you have a perfectly reliable connection to a protective earth, if the mains circuit has, at any point in the circuit, a short circuit active to earth, then for a brief moment the protective earth circuit will rise to approximately 50% of the mains voltage unless the fault is a bit further upstream where the protective earth conductor can be smaller than a phase conductor so the protective earth circuit will go even higher until the circuit breaker interrupts the current. It is bad form to hard wire a patient to such a circuit.
Next point is simply this: dealing with a bio signal of 10uV is no easy thing. This signal is much smaller than a typical IA offset voltage and as you have seen in the literature you posted, even a small electrode resistance under the influence of the IA bias and offset currents will typically produce an even bigger offset. So you will need to deal with the fact that a good proportion of the signal you will amplify will have an offset bigger than the signal and will need to be removed at some point before your circuits run out of headroom.
Next comes the interferers. These can dwarf the offsets and everything else if your signal chain and signal processing is not handled very carefully. One of the biggest complicating factors here is that you cannot use the protective earth to shunt any of this noise away. You will need to either use concepts such as the RLD reimagined for your application or use an amplifier circuit embedded in the electrode itself or one of the thousands of other ways to deal with 10uV bio signals. You will need to find the solution that you need for your application. Most of the standard techniques will probably not apply or work for you. There will be a solution though. It just may take a while for you to find it.
Based on your questions and your responses to my responses, I can see you are bright and learn quickly but this job is better handled when you have a bit more experience and some grasp of the regulatory requirements. With experience you will avoid the vast number of traps waiting in a design of this type and knowledge of the regulatory requirements is not just something you need to get approval or to tick a box. You need it as a formal assessment of the safety of this device that will be used on humans. Medical devices with direct connections to the patient must be designed with that fact always in your mind and those parts of the standard that relate to that fact and the methods of keeping the patient and the operator protected. Being battery operated does not instantly render your device safe. It only changes the risk factors and how they are assessed. Safety must not just be an afterthought. It should be part of your thinking and design from the start.
Ok, I have said my piece, you have wasted my time and I have given you as much direction as I can anyway. Where you go from here is up to you. If you continue, find a mentor you can access easily. You will need more than the help you can get on a forum.
Good luck.

 

In a previous posing in the thread you said: " my main application is EMG with 1mV to 5mV input, not ECG nor EEG " and now you are saying EEG and 10uV signals and quote a circuit diagram that makes allowance for patient connections to be earthed.
Ok, I will be frank, please do not be offended, since you mislead me about the application a lot of my responses are irrelevant in many ways so it was a waste of my time.
First point, NEVER earth patient connections. Even if you have a perfectly reliable connection to a protective earth, if the mains circuit has, at any point in the circuit, a short circuit active to earth, then for a brief moment the protective earth circuit will rise to approximately 50% of the mains voltage unless the fault is a bit further upstream where the protective earth conductor can be smaller than a phase conductor so the protective earth circuit will go even higher until the circuit breaker interrupts the current. It is bad form to hard wire a patient to such a circuit.
Next point is simply this: dealing with a bio signal of 10uV is no easy thing. This signal is much smaller than a typical IA offset voltage and as you have seen in the literature you posted, even a small electrode resistance under the influence of the IA bias and offset currents will typically produce an even bigger offset. So you will need to deal with the fact that a good proportion of the signal you will amplify will have an offset bigger than the signal and will need to be removed at some point before your circuits run out of headroom.
Next comes the interferers. These can dwarf the offsets and everything else if your signal chain and signal processing is not handled very carefully. One of the biggest complicating factors here is that you cannot use the protective earth to shunt any of this noise away. You will need to either use concepts such as the RLD reimagined for your application or use an amplifier circuit embedded in the electrode itself or one of the thousands of other ways to deal with 10uV bio signals. You will need to find the solution that you need for your application. Most of the standard techniques will probably not apply or work for you. There will be a solution though. It just may take a while for you to find it.
Based on your questions and your responses to my responses, I can see you are bright and learn quickly but this job is better handled when you have a bit more experience and some grasp of the regulatory requirements. With experience you will avoid the vast number of traps waiting in a design of this type and knowledge of the regulatory requirements is not just something you need to get approval or to tick a box. You need it as a formal assessment of the safety of this device that will be used on humans. Medical devices with direct connections to the patient must be designed with that fact always in your mind and those parts of the standard that relate to that fact and the methods of keeping the patient and the operator protected. Being battery operated does not instantly render your device safe. It only changes the risk factors and how they are assessed. Safety must not just be an afterthought. It should be part of your thinking and design from the start.
Ok, I have said my piece, you have wasted my time and I have given you as much direction as I can anyway. Where you go from here is up to you. If you continue, find a mentor you can access easily. You will need more than the help you can get on a forum.
Good luck.

Actually all your advices help immensely because we have 2 people in our BCI and robotic interface project. One of us is supposed to focus on the EMG part in controlling the augmented cytoskeleton. One of us will focus on the EEG part in controlling it via the brain. My part earlier was supposed to be the EMG part. Then the last few days. We switched our sides. And since we both will work together with EMG and EEG, then all your advices help full time.

I was just left with a puzzle. Because EEG is very common now. And it uses 10uV. Other people pls tell me how most EEG is able to handle this thing about "dealing with a bio signal of 10uV is no easy thing. This signal is much smaller than a typical IA offset voltage and as you have seen in the literature you posted, even a small electrode resistance under the influence of the IA bias and offset currents will typically produce an even bigger offset. So you will need to deal with the fact that a good proportion of the signal you will amplify will have an offset bigger than the signal and will need to be removed at some point before your circuits run out of headroom.".

How do the hundreds of EEG products deal with the above? Does it mean the Instrumentation Amplifier used in EMG and EEG are not the same? Thanks a lot guys!

 

Actually all your advices help immensely because we have 2 people in our BCI and robotic interface project. One of us is supposed to focus on the EMG part in controlling the augmented cytoskeleton. One of us will focus on the EEG part in controlling it via the brain. My part earlier was supposed to be the EMG part. Then the last few days. We switched our sides. And since we both will work together with EMG and EEG, then all your advices help full time.

I was just left with a puzzle. Because EEG is very common now. And it uses 10uV. Other people pls tell me how most EEG is able to handle this thing about "dealing with a bio signal of 10uV is no easy thing. This signal is much smaller than a typical IA offset voltage and as you have seen in the literature you posted, even a small electrode resistance under the influence of the IA bias and offset currents will typically produce an even bigger offset. So you will need to deal with the fact that a good proportion of the signal you will amplify will have an offset bigger than the signal and will need to be removed at some point before your circuits run out of headroom.".

How do the hundreds of EEG products deal with the above? Does it mean the Instrumentation Amplifier used in EMG and EEG are not the same? Thanks a lot guys!

The hundreds of EEG products out there were not designed by a newbie. They were designed by clever and experienced engineers who also knew how to design the products in compliance with the regulatory standards. As I said, you may well succeed but you will also likely fall into many traps along the way.
You need to be clearer about what you are doing when you ask for advice. Jumping around as you did puts my advice in a confused context which will not help you. And now that I know you are a team working on a project, if you want more advice, you need a consultant. My rates are reasonable. I am not going to coach you through to a finished system for free.