Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Medical Brain analysis through neural oscillation

  1. Sep 29, 2012 #1
    Is neural oscillation the only method in which we can analyze and communicate with the brain?
  2. jcsd
  3. Sep 29, 2012 #2


    Staff: Mentor

    Welcome to the Physics Forum!

    No we talk to the brain and do MRI scans to see how things were processed.
  4. Sep 29, 2012 #3

    I'm specifically interested in bi-directional methods for communicating and analyzing the brain.

    For instance, the ability to analyze brain patterns and understand that the subject is looking at the color blue or experiencing sadness. And in the other direction, the ability to stimulate or activate the brain in such a way that the subject perceives the color blue or feels sadness.

    I understand that an MRI can analyze brain patterns and, to a certain extent, decipher their meaning. But it's not bi-directional, correct?
    Last edited: Sep 29, 2012
  5. Sep 29, 2012 #4


    Staff: Mentor

    A functional scan shows where blood is flowing and the assumption is made that thats where the brain is processing the info but its not been proven conclusively.

  6. Sep 29, 2012 #5
    To a large extent things like this are a function of location rather than "oscillations". Electrically stimulating a precise spot can, in fact, give someone things like the experience of blue or sadness.

    With depth electrodes you can also receive signals, and activity at a certain location can be correlated with the subjects report of what he is experiencing.

    That last procedure is completely invasive and is never performed on healthy subjects for mapping purposes. It's limited to people who are going to have brain surgery anyway due to intractable seizures or tumors.
  7. Sep 30, 2012 #6


    User Avatar
    Science Advisor

    Do you mean something like this? I can smell burnt toast (just a dramatization)

    (interview, see around 2:11)

    A similar experiment, much simpler but more controlled: http://www.ncbi.nlm.nih.gov/pubmed/1607944

    In deep brain stimulation (DBS) for Parkinson's, there is some evidence that the stimulus is best when it is periodic. http://jn.physiology.org/content/104/2/911.long

    However, some might say that DBS is not "communicating" with the brain in the sense that it may work not by restoring normal function, but by preventing abnormal signals from reaching the cortex. More generally, I believe the mechanism by which DBS works is not known. http://www.ncbi.nlm.nih.gov/pubmed/20850966

    I haven't watched this, but it should be at least entertaining, since it's a TED talk

    There seems to be experimental investigation into using TMS for depression
    Last edited by a moderator: Sep 25, 2014
  8. Sep 30, 2012 #7
    This is what I'm talking about except that after Penfield they began implanting depth electrodes so they could both send and receive signals from points in the limbic system. The electrodes are very thin and they are stripped of insulation at regular intervals so that each electrode can either send or receive at different points along its length. I think they might put as many as ten in the suspect hemisphere to locate the seizure focus as precisely as possible. Penfield was limited to triggering seizures from the cortex. The focus might be far removed from the point where the seizure activity reaches the cortex, so surface stimulation isn't all that reliable.

    This is also the method Penfield used to map the sensory homunculus, IIRC: surface stimulation.

    The Parkinson's thing is somewhat different. There they are actually treating symptoms, not simply trying to trigger an event, so the frequency is important. There is a similar-but-different seizure treatment: vagus nerve stimulation. A device in connected through the skin of the neck to the vagus nerve. When it detects seizure activity it sends a kind of damping signal up the nerve into the brain. These have had mixed results. They have to be fine tuned for each patient and that tuning process takes a long time.

    Anyway, this is only "communicating" is the crudest sense, such that I can see why some wouldn't call it communicating at all.
    Last edited by a moderator: Sep 25, 2014
  9. Sep 30, 2012 #8
    These methods seem very imprecise. Is there a method that may one day be adopted for use in mass consumer goods? Or at least a method that's headed in that direction?

    My limited knowledge of this field suggests that neural oscillation is currently the most precise and unobtrusive method for analyzing neural activity -- even if it’s limited to clusters of thousands or millions. I was curious if there were other methods -- equally unobtrusive and ideally much more precise -- under development?

    Here are a few more questions:

    1. EEG and MEG detect the activity of thousands or millions of neurons. I imagine this is a limitation of both the technology and approach. Would it be useful to measure the activity of a single neuron?
    2. Do neurons only really matter in large clusters? (e.g., does a single neuron control anything?)

    These last two may be a bit far fetched but interesting nonetheless.

    1. Consider the possibility of mapping or recording the neural activity required to learn a single word. Would this process be similar amongst a large portion of the population or is it highly specific to each individual?
    2. Could it be possible to stimulate the brain so that it mimics the previously recorded neural activity? And if so, would that result in the subject learning the word?
  10. Sep 30, 2012 #9


    User Avatar
    Science Advisor


    However, for practical use I would imagine that you'd want many neurons to be involved in the effect. Otherwise, just jiggling the implant a little in the brain might kill the single important neuron.

    I don't know of any methods that try to be more precise, without being more invasive. Most of this is still technology being developed.

    Maybe stuff like:
  11. Oct 9, 2012 #10
    For humans it's very difficult as doing anything that records with higher resolution than current methods requires getting rid of the skull. For electrical signals, the skull acts as a spatial filter, smearing together signals that are separated spatially. The effects are greater for higher frequency oscillations, as higher frequency oscillations tend to be less coherent across space than lower frequency oscillations. Other methods like fMRI give better spatial resolution, but much worse temporal resolution (and fMRI uses blood oxygen level as a proxy for neural activity. And needs a lot of liquid helium.)

    This is done in brain slice preparations and in animal experiments using sharp electrodes implanted into the brain (or with more fancy imaging techniques). It gives you a different type of information; whether it is useful depends on the question you are asking.
    As far as I am aware, it depends on the brain area, and what type of processing is occurring - but we're only just starting to unravel the neural code really...

    What level of detail of neural activity would you want to record? Firing of every cell? Release of neurotransmitter at every synapse? Dendritic spiking? Gap junctional communication? Glial processes? Gene expression? I would imagine on a low level, it would vary hugely between individuals, but at a much higher level, similarities would be observed.
    I would say not exactly, because the the activity of the brain doesn't just depend on the input, but also on the brain's current state, which is constantly changing. However, if you could make your artificial stimulus in a way to account for this, then I guess perhaps theoretically...
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook