Maybe something simpler and more reproducible to attempt first (and
bear with me, because I know nothing about EEG) would be to stimulate
the brain with some external sensory input that might produce
recognizable patterns in the EEG output. If NuPIC can recognize
patterns coming out of an EEG that are being injected artificially,
that is evidence that it might be able to recognize patterns that
originate organically from internal thought.

The simplest thing would be to make the wearer stare at a screen that
is alternating between black and white every few seconds. An
interesting experiment would be to see if NuPIC learns the patterns
well enough to throw a high anomaly score if the screen unexpectedly
displays a polkadot green screen instead of black (or something like
that).

Once there is evidence that NuPIC can recognize patterns in EEG
output, more involved experiments can be done that involve the subject
alternatively thinking about calm vs stressful situations. Perhaps
NuPIC can pick up that difference, too?
---------
Matt Taylor
OS Community Flag-Bearer
Numenta


On Fri, Jan 10, 2014 at 6:38 AM, Hans Czylok <[email protected]> wrote:
> Hello!
>
> Does anyone has got some experience with NeuroSkys Mindwave Mobile?
>
> I have got one here since 2 days and wonder which signal of EEG or NeuroSkys
> would be useful to measure stress relief after some certain mental and body
> training.
> The advantage is that the price is rather low and SDK is free. The
> disadvantage is that I can not wear it more then 15 minutes, because
> forehead electrode and the adjustable head band press too much. Maybe my
> head is simply too big :).
>
> It would be interesting, if CLA could be used too learn certain EEG patterns
> too identify "too much stress in the brain" not only in the actual situation
> as Neurosky self calculated indicators for "attention" and mediation" do,
> and already too identify the average risk for burn out or other mental
> disease before they occur or to identify good relaxed condition of brain.
> Could it be possible to measure the "quality" of brain's condition on basis
> of EEG signals?
>
> best regards
> Hans Czylok
>
>
> 2014/1/9 Jay Gokhale <[email protected]>
>>
>> Oops didn't reply all:
>>
>> On Jan 8, 2014 4:20 PM, "Jay Gokhale" <[email protected]> wrote:
>>>
>>> Hello Matt,
>>> My lab is getting a bunch of EEG supplies in the coming months.  I'll let
>>> you know if I have a spare to bring.
>>>
>>> On Jan 7, 2014 3:16 PM, "Matthew Taylor" <[email protected]> wrote:
>>>>
>>>> So does anyone have a spare 8-channel EEG cap lying around that they
>>>> aren't using? We could use it at our next hackathon! :)
>>>> ---------
>>>> Matt Taylor
>>>> OS Community Flag-Bearer
>>>> Numenta
>>>>
>>>>
>>>> On Sat, Jan 4, 2014 at 11:19 PM, Chetan Surpur <[email protected]>
>>>> wrote:
>>>> > Sounds like a great hackathon project :)
>>>> >
>>>> >
>>>> > On Fri, Jan 3, 2014 at 5:34 AM, Hideaki Suzuki <[email protected]>
>>>> > wrote:
>>>> >>
>>>> >> Hi Marcos,
>>>> >>
>>>> >> I agree.  BCI is an interesting possibility of CLA for me.
>>>> >>
>>>> >> I *hope* CLA can automatically capture useful special/temporal
>>>> >> patterns of
>>>> >> inputs (EEG and so forth) and convert them to outputs.  Nobody,
>>>> >> however, has
>>>> >> tried it yet AFAIK.  How good CLA can classify patterns out of noise
>>>> >> is
>>>> >> unknown. (please correct me in case someone knows).  We may need
>>>> >> filters to
>>>> >> reduce unwanted signals.     We may need CLA hierarchy...  I have no
>>>> >> clue.
>>>> >> Please don't receive a wrong impression from me like BCI is already
>>>> >> considered.
>>>> >>
>>>> >> Applying CLA in the field of BCI must be a good research topic.
>>>> >>
>>>> >> Before starting, we could say something about the feasibility by
>>>> >> feeding
>>>> >> EEG to Nupic and see abnormality score.  If the score is low, IIUC,
>>>> >> CLA is
>>>> >> predicting some patterns detected.
>>>> >>
>>>> >> B.R.
>>>> >>     Hideaki Suzuki.
>>>> >
>>>> >
>>>> >
>>>> > _______________________________________________
>>>> > nupic mailing list
>>>> > [email protected]
>>>> > http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>>>> >
>>>>
>>>> _______________________________________________
>>>> nupic mailing list
>>>> [email protected]
>>>> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>>
>>
>> _______________________________________________
>> nupic mailing list
>> [email protected]
>> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>>
>
>
> _______________________________________________
> nupic mailing list
> [email protected]
> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>

_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to