Shop OBEX P1 Docs P2 Docs Learn Events
How to re-engineer fMRI? — Parallax Forums

How to re-engineer fMRI?

An fMRI machine is very expensive and deserves a cheap in-home model if possible. How can one be built in to a helmet? The idea as far as I've seen is a combination of radio waves and magnets that show what the blood vessels in the brain are doing. That feature plus a polygraph used at the same time is reportedly 100% effective at lie detection. There are serious needs for 100% correct lie detection, starting with courts which currently toss 25% average to 60% on some crimes innocent people into prison. Folks at home would like an accurate answer to family questions like, where were you last night?

So to start with, what are the physics involved that need to be simulated a cheaper way? What are the radio wave frequencies, detection techniques, magnet fields required? I've been in a CT scan machine and see the hula-hoop of electronics spinning around - can that motion be simulated like stationary radar antennas do? What else is needed to get the job done? Thank you.

Comments

  • Have you asked google? You have seriously underestimated what is required to accomplish this.
  • That was whee I got the info on radio waves and magnets. I have not found enough information to be useful so far.
  • There is no way you will be able to come close to this in a home setting and at a price less than 6 or more likely 7 figures not including the decimal places. The fMRI I have done every two years (as a lab rat for BNI) is performed in a normal MRI system with the appropriate software for the fMRI process. So if you have a (VERY) large RV gate, you may be able to get a trailer unit such as you would see at smaller regional hospitals. An alternative, which may be what you are thinking of, is MEG. This does not use blood flow as the indicator, rather it uses the magnetic fields caused by electrical currents caused by brain activity. EEG and other surface scan systems give an indication of what may be going on in the brain, but MEG can create a detailed image of where these fields are. It is a monster sized machine, but this link (https://spectrum.ieee.org/the-human-os/biomedical/imaging/a-new-wearable-brain-scanner) shows a smaller unit that may give you a better idea of the challenges of what you are looking at.. another link to current MEG technology https://ucsfbenioffchildrens.org/education/magnetoencephalography_meg_scan/
  • Dr Dee Wu is a good friend of mine who has pioneered the field of fMRI at the University of Oklahoma.
    Although I doubt he will offer to reverse engineer any of this for you, I do know that he is a big fan of Parallax.
    If your nice and he has some free time he might be willing to bounce an idea or two.

    https://www.oumedicine.com/dwu/c-clin/c01-neuro/c01_02/index.html
  • The M in MRI stands for magnetic. The whole thing is based on the fact that some atomic nuclei are magnetic and will align themselves with a suitably strong magnetic field; they will then precess, which causes them to absorb radio waves of a characteristic frequency. (MRI used to be called NMR for Nuclear Magnetic Resonance, but they changed it because of that scary word "nuclear.") MRI machines use some of the most powerful magnets ever built; most of them are superconducting, and all of them are dangerous if you make the mistake of approaching too closely while there is any magnetic material on your person. There simply is no way to build a small cheap MRI imager. The wonder is that it is possible to build such scanners at all.
  • Kirk FraserKirk Fraser Posts: 364
    edited 2018-11-29 01:51
    Beau and Frank, Thank you very much.

    RickB and localroger thank you too.
  • No response to this email yet. It sounds like the only way to clean up the Justice system.

    Subject: Small fMRI project?

    Dr. Wu,

    I am seeking a wearable fMRI after reading a combination of fMRI and Polygraph at the same time gives 100% accurate lie detection, which could free 25% of prisoners who are actually innocent. Is there a way you can make or help create a very low cost project to build such a device like projects sold on Parallax.com or other home hobby sites?

    Here is a websites that may provide ideas for this project.
    https://spectrum.ieee.org/the-human-os/biomedical/imaging/a-new-wearable-brain-scanner

    Thank you very much.
    Kirk W. Fraser
    www.theManChild.org

  • kwinnkwinn Posts: 8,697
    localroger wrote: »
    The M in MRI stands for magnetic. The whole thing is based on the fact that some atomic nuclei are magnetic and will align themselves with a suitably strong magnetic field; they will then precess, which causes them to absorb radio waves of a characteristic frequency. (MRI used to be called NMR for Nuclear Magnetic Resonance, but they changed it because of that scary word "nuclear.") MRI machines use some of the most powerful magnets ever built; most of them are superconducting, and all of them are dangerous if you make the mistake of approaching too closely while there is any magnetic material on your person. There simply is no way to build a small cheap MRI imager. The wonder is that it is possible to build such scanners at all.

    The magnetic field strength required for MRI makes head mounted or portable MRI imaging systems highly unlikely. Other imaging modes are more likely to be used for such scanners if they are at all possible.
  • potatoheadpotatohead Posts: 10,253
    edited 2018-11-30 04:39
    Nobody has shown we can identify intent via fMRI or any of these technologies yet.

    Non starter, currently. But, let's say we can do that. Could happen.

    Intent is a slippery thing. Intent is what we convict on too. (and that does bring up many arguments for handling crime very differently than we often do in the US too, but out of scope for this forum. Just saying.)

    A determined, informed subject can refactor intent, literally convince themselves of some reality. While this is not generally applicable to the problem at hand; namely, freeing innocent people, it does present a significant issue, given the pitch is 100 percent lie detection. It would take a trained group, investing very significant amounts of time to even consider untangling too. If it can even be untangled. If it's possible at all, the outcome would be expressed in terms of degrees of confidence, never 100 percent.

    Our current science has not even suggested doing any of that is plausible. And that is for ordinary people, who we would have to assume do not harbor intent to be set free, just because they can. Go ahead, filter for that. I'll wait.

    If they present any form of psychopathy, it's even worse.

    I submit these ideas alone would warrant a no reply to the email.





  • W9GFOW9GFO Posts: 4,010
    The other side of a "magic" device that can detect lies is that someone unwilling to submit to it may be judged to be guilty. Its very existence could send people to prison that won't allow it to be used on them due to moral convictions.

    Sounds like a good plot line for a Black Mirror episode.
  • Indeed!

  • kwinnkwinn Posts: 8,697
    W9GFO wrote: »
    The other side of a "magic" device that can detect lies is that someone unwilling to submit to it may be judged to be guilty. Its very existence could send people to prison that won't allow it to be used on them due to moral convictions.

    Sounds like a good plot line for a Black Mirror episode.

    Sounds like we're getting into "how many angels can dance on the head of a pin" territory with this "magical device" and "moral convictions". Assuming the device has been proven safe to use what possible moral conviction could be used for refusing to submit? Only thing I can come up with is that the person is guilty.
  • W9GFOW9GFO Posts: 4,010
    edited 2018-11-30 23:15
    kwinn wrote: »
    ...Only thing I can come up with is that the person is guilty.

    That is exactly my point. Can you honestly say that you have considered every possible consequence in allowing the state to compel someone to submit themselves to such an interrogation?


  • kwinnkwinn Posts: 8,697
    W9GFO wrote: »
    kwinn wrote: »
    ...Only thing I can come up with is that the person is guilty.

    That is exactly my point. Can you honestly say that you have considered every possible consequence in allowing the state to compel someone to submit themselves to such an interrogation?


    No, but could it be worse than what is happening now?
  • localroger wrote: »
    The M in MRI stands for magnetic. The whole thing is based on the fact that some atomic nuclei are magnetic and will align themselves with a suitably strong magnetic field; they will then precess, which causes them to absorb radio waves of a characteristic frequency. (MRI used to be called NMR for Nuclear Magnetic Resonance, but they changed it because of that scary word "nuclear.") MRI machines use some of the most powerful magnets ever built; most of them are superconducting, and all of them are dangerous if you make the mistake of approaching too closely while there is any magnetic material on your person. There simply is no way to build a small cheap MRI imager. The wonder is that it is possible to build such scanners at all.

    There are now research magnets in the 7T range (normal studies typically use 3T magnets and some particular studies require 5T magnets). There are sites with pictures of serious incidents with metal getting into the room and the aftermath. More uhm, senior members will note their Stent cards will show the maximum field strength allowable for that particular stent.

    To the OP, there are a number of good sites on MRI principles. And just like any other medical imaging system, there are trade offs. Power use, power output, field strength, processing algorithms to take k space data and turn it into image. But there may be noise in the signal, how would you be sure you are not convicting someone on an artifact?

    Goofy as our system is, it errs (or should by the original intent) on the side of defendants; of all the systems I have either read about or observed, ours is still the best system out there. Misuse and abuse discussions of "the system" is for sites other than this one.
  • Goofy as our system is, it errs (or should by the original intent) on the side of defendants; of all the systems I have either read about or observed, ours is still the best system out there. Misuse and abuse discussions of "the system" is for sites other than this one.

    It seems the courts agree that at present there are issues with this idea: http://stlr.org/2018/01/10/does-fmri-lie-detection-have-a-future-in-the-courtroom
  • This PET scanner is fully open source, I believe. See also the Terasic website

    https://openpet.lbl.gov/
  • I read in two sources that polygraph + fMRI does 100% accurate detection. The theory is both are less than 100% accurate alone but become 100% accurate used together because they detect blood flow changes to different areas in response to questions. A person telling the truth has a certain pattern of areas activated indicated by drawing blood flow and a liar will redirect the blood flow to other areas. Both devices together determine if a person is thinking normally or redirecting blood flow to say other body parts. I do not know if this system can be cheated, that would be answered by a study. I have seen a documentary showing Barbara Walters cheating a polygraph. jones court link only refers to fMRI alone, not in conjunction with polygraph. It claims only 70% accuracy for polygraphs.

    Courts typically abuse citizens who maintain their innocence by offering a polygraph test at the defendant's expense then if they pass the test, it is not even mentioned in court, costing the defendant thousands for nothing. Apparently an incentive to take a plea bargain. An fMRI exam would cost much more, which is a reason for trying to engineer a much cheaper machine so the whole process can be afforded by innocents who may be living in poverty. A combined device would be useful in many other cases where prison is not a problem, such as when person B lies denying running into the car of person A, blocking insurance payment. It could be used in Press Conferences to tell if a politician or a reporter is lying or telling the truth. In all cases better than 90-95% accuracy is required to get serious attention.

    The normal route is to wait for government funding to do a full study. All kinds of things may happen. The person who made the initial discovery may find some other research topic more interesting, he may have retired, or any such failure to follow up the discovery with a full study may happen. After one or more full studies, professionals would become more serious about making it cheaper.

    Special thanks to Tubular for the link to the PET scan project, and thanks to everyone else too.
  • Considering that the standard technique for beating a lie detector is to effectively convince yourself that you are telling the truth, it's hard to see how any method short of full-blown mind reading could be absolutely reliable. It is possible for us to construct inner realities that have nothing to do with external reality, and it is possible to do so deliberately and select the reality you are perceiving at any given time. This is the basis, for example, for "method acting." It seems to me that if someone has created an inner reality in which they are justified and innocent, their brain would function just like someone who really was justified and innocent when you ask them about it.
  • I read in two sources that polygraph + fMRI does 100% accurate detection.

    These people have good intent. Of that, I have no doubt. All I can say is, careful what you wish for.

    There are also people out there with nefarious intent. High claims = high rewards.

    Defeating something like this would happen in a very short time, and the same government that may fund the research, will fund doing that too.

    Ask why. It's good for you.

    And that is all I will put here on this. Good luck!



  • it is possible the whole idea is never going to be 100% accurate against people who have trained themselves to speak false constructs as true like Barbara Walters and all professional actors. In that case, full on mind reading wouldn't work either without a big artificial intelligence to compare real or imagined perceptions with actual reality. Perhaps with adequate mind input an AI could help focus on truth centers and discard lie centers in the brain. However people are different so some people may have developed their truth and lie centers in non-standard brain locations, which would make it more difficult to detect for each individual. On the other hand, it is not possible to be certain it's impossible without trying.

    If Barbara Walters could defeat simultaneous polygraph + fMRI, then one can assume she is speaking news reports as true without being sure they are true. Perhaps any lie detector can only tell what a person believes, without regard for actual reality. Yet if a machine can tell what a person believes v. what they say but don't believe it might be useful in court to a jury that has no other fact to base their decision on.
Sign In or Register to comment.