[Eeglablist] Meaning of ITC

Thu Apr 16 19:48:02 PDT 2009

Dear Ernesto,

Thank you for your response.
I tried to read Martinez-Montes et al. (2008) but gave it up (sorry).
Instead, I checked Yeung et al. (2004) and found it important. They
reported the result of simulation that adding a 'non-phase-locked' 5Hz
half-cycle sine wave to the background noise can cause ITC increase. I'm
not surprised to see the result (though this time I could be wrong here).
Since 5Hz is 200 ms/cycle, and the given jitter was 60 plus or minus 32 ms;
as long as you use a usual window length (several hundred milliseconds) or
wavelet cycles at 5 Hz (>1), you should find ITC increase there. 

We should be aware of the issue anyway, otherwise we may be mislead when
interpreting data. Thanks for the info.

Makoto

--- Ernesto Palmero <e.palmero at googlemail.com> wrote:

> Dear Makoto Miyakoshi:
> 
>  I have read you post at the EEGLab forum and maybe I am completely
> wrong
> but ITC in at least in the way is computed in the EEGLab can be
> related with
> ERPS and has to analyze with muh caution. ITC has been or phase
> consitency
> has been propose as one of the possible mechanism to generate the
> ERPs. If
> for example you have a singal in the single trial which phase are
> random
> except at lets say 30ms in wich ideally all the phase have the same
> value,
> then when you perform the average you will see a nice ERPs come up.
> The only
> problem here  is that in the case that you have an increase in power
> or an
> ERPS you could have a similar behaviuor but just because a fix
> latency and
> amplitude component will produce the exactly the same behaviour. Thus
> why
> you have to be very carefull when applying and interprete the ITC
> results
> because you can confound effects and maybe the effect you are looking
> is
> just related with an increase in the EEG signal over the single trial
> or an
> ERP in this case ITC will have the same interpretation (because it
> could be
> generater) like the ERP. I will advice you to paper in this file that
> I
> think are really illustrative of this problem and can give some ligth
> on
> this issue:
> 
> STATISTICS IN MEDICINE
> Statist. Med. 2008; 27:2922–2947
> Published online 13 December 2007 in Wiley InterScience
> (www.interscience.wiley.com) DOI: 10.1002/sim.3132
> Exploring event-related brain dynamics with tests on complex
> valued timeËÇrequency representations.
> by Eduardo Martinez-Montes et.al
> 
> and
> 
> Detection of synchronized oscillations in the
> electroencephalogram: An evaluation of methods
> Psychophysiology, 41 (2004), 822–832. Blackwell Publishing Inc.
> Printed in
> the USA.
> Copyright r 2004 Society for Psychophysiological Research
> DOI: 10.1111/j.1469-8986.2004.00239.x
> by NICK YEUNG et. al
> 
> I will before trying to find any possible explanation I will rule out
> that
> something like the one describe in this paper are not happening and
> then you
> can really try to find the reason for the ITC. If you have a mix
> situation
> like the one this paper describe it will be very difficult to do it.
> 
> Have a nice day and sorry if this note does not answer your question
> or you
> already knew about this issue
> Ernesto Palmero Soler
> 
> 2009/4/16 Makoto Miyakoshi <mataothefifth at yahoo.co.jp>
> 
> > Dear Pal,
> >
> > >From this reply I hope our communication is normally logged on the
> > eeglablist.
> >
> > Yes, the Bereitshcafts Potentials seems a good example, if the
> potential
> > were constructed by phase locking (I'm not sure on this point). By
> the way,
> > could you specify 'Alan Gevins works from around 1984' which I
> could not
> > find? I want to take a look at the illustrations.
> >
> > The immediate meaning of ITC is 'trial-by-trial EEG phase coherency
> at a
> > certain IC/scalp channel, time bin, and frequency range, is likely
> to be
> > this much'. It only reflects (at least in princiele) the timing of
> neural
> > firing (large portion of neurons), and does not reflect the amount
> of it
> > (though ERSP and BOLD do, in constrast, in my understanding).
> Certainly, a
> > physiological meaning of ITC is a event-related 'timing' of
> neuronal
> > firing.
> >
> > So, my question again may be: what is the 'psychological' meaning
> of ITC
> > increase/decrease, given that it does not reflect increase/decrease
> of
> > involved neuronal activity? Maybe I'm too much used to a simple
> schema that
> > associates ERSP/BOLD increase with 'activation' of a certain
> portion of
> > neurons.
> >
> > Makoto
> >
> >
> >
> >
> >
> >
> >
> > --- P $B>-%% (Bl Gunnar Larsson <Pal.Gunnar.Larsson at epilepsy.no>
> wrote:
> >
> > > Dear Makoto
> > >
> > > > Dear Pal,
> > > >
> > > > Thank you for the reply. It is very informative and
> interesting...
> > > > almost exciting.
> > > >
> > > > It seems to me, however, that the term 'synchronization' is
> used in
> > > > different meanings between you and me. Maybe you mean it in a
> sense
> > > of
> > > > 'synchronized firing among neurons at a given moment', right?
> There
> > > are
> > > > actually papers that report inter-channel coherency, and that
> is
> > > > certainly a kind of synchronization. But the concept of ITC is
> > > 'inter-
> > > > trial'
> > > > synchrony: it compares EEG phases AMONG TRIALS and measure
> phase
> > > > variance, so it had nothing to do with inter-channel phase
> > > synchrony.
> > >
> > > You do have a point. I don't know your research at all. Is inter
> > > trial between stimuli or between sessions or between persons? I
> have
> > > not thought this through, but I was thinking between stimuli and
> > > something not unlike to bereitschaft-potentials, but mainly my
> pont
> > > was on the relation betwen bold and EEG.
> > >
> > > >
> > > > Maybe we'd better create a term 'Inter-neuronal coherency
> (INC)' in
> > > > contrast with inter-trial coherency (ITC).
> > > Coherency in EEG was very popular, but it did not lead to very
> much
> > > more than some papers. One of hte main reasons was that spatial
> > > filters were not applied (except by Gevins) and hence there were
> too
> > > much blurring of the findings. I guess you should look at Alan
> Gevins
> > > works from around 1984. He did 128 channel EEG and made arrows
> > > showing coherences and latencies between brain areas. His
> > > illustrations have been widely used in the literature. To my
> > > knowledge he did not look at ITC.
> > > >
> > > > I'm analyzing my EEG data and found some experimental effect in
> > > ITC,
> > > > but I have no idea how to interpret it... because theoretically
> it
> > > is
> > > > not related to neither ERSP (EEG power) nor BOLD.
> > > Back to bereitschaft - there is no expectance in this?
> > > >
> > > > I will ask EEGLABlist technicians how to post our
> communications to
> > > the
> > > > list. By the way, so far you are the only person who replied.
> My
> > > > gratitude is all yours (haha).
> > > >
> > >
> > > Regards
> > >
> > > P $B%;%e<+ (Bl
> > >
> > > > Makoto
> > > >
> > > >
> > > >
> > > > --- P $B<@</<'<' (Bl Gunnar Larsson
> <Pal.Gunnar.Larsson at epilepsy.no>
> > wrote:
> > > >
> > > > > Dear Makoto
> > > > >
> > > > > First - I thought it went back to the list, but did not check
> to
> > > see
> > > > > that it did.
> > > > >
> > > > > My points was mostly towards the Nunez take on this;
> > > synchronization
> > > > > does not require energy. Hence synchrony comes out with high
> EEG-
> > > > power
> > > > > and no BOLD-effect. On the other hand, interneurons in cortex
> are
> > > so
> > > > > symmetrical that they do not show up on EEG, but consumes
> quite a
> > > bit
> > > > > of energy and will be well seen in fMRI.
> > > > >
> > > > > According to the above, your phase locking(?) would be
> expected
> > > to
> > > > > influence EEG-power but not BOLD (or ?).
> > > > >
> > > > > Another point from Nunez: If you have a million stochastic
> firing
> > > > > neurons, they would give a resultant amplitude of A*sqrt(n)
> where
> > > A
> > > > is
> > > > > the amplitude and n is number of neurons. However, when they
> are
> > > > > synchronized, the amplitude is n*A. So a million neurons will
> > > show
> > > > > same amplitude as 1000 synchronized neurons! This gives that
> > > small
> > > > > systems may dominate the EEG you record.
> > > > >
> > > > > P $B<'<=<'t@<- (Bl
> > > > >
> > > > > -----Opprinnelig melding-----
> > > > > Fra: Makoto Miyakoshi [mailto:mataothefifth at yahoo.co.jp]
> > > > > Sendt: 14. april 2009 03:54
> > > > > Til: P $B<'<=<'t@<- (Bl Gunnar Larsson
> > > > > Emne: RE: [Eeglablist] Meaning of ITC
> > > > >
> > > > > Dear Pal,
> > > > >
> > > > > Thank you for your fast response and information.
> > > > > Neurovascular coupling is the essential phenomenon with BOLD
> and
> > > it
> > > > > has totally different time constant than EEG... that is fine
> with
> > > me.
> > > > > That said, what I'm wondering is the meaning of ITC. On the
> one
> > > hand,
> > > > > ERSP is a measure of EEG power, which most likely reflects
> sum of
> > > > > regional neuronal activities, which should be associated with
> > > BOLD
> > > > via
> > > > > neurovascular coupling (in principle). On the other hand, ITC
> has
> > > > > nothing to do with EEG power, since it is a measure of
> > > inter-trial
> > > > > phase variance, therefore it should not affect BOLD either.
> Then,
> > > > what
> > > > > is the meaning of ITC in contrast with ERSP/BOLD?
> > > > >
> > > > > By the way, your response does not seem to be sent to EEGLAB
> > > list.
> > > > > Why don't we re-send it to the list?
> > > > >
> > > > > Makoto
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > --- P $B<><B<><1<><)<><) (Bl Gunnar Larsson
> > <Pal.Gunnar.Larsson at epilepsy.no>
> > > wrote:
> > > > >
> > > > > > I think you should read the paper by PL Nunez on the
> relation
> > > > > between
> > > > > > EEG and fMRI from 2000 (or 1999?). EEG and BOLD is very
> > > different
> > > > > > entities. Also, the EEG- parameters are direct functional
> > > measures
> > > > > > which gives responsen from a few miliseconds to some
> seconds
> > > > > > (Pfurtscheller). BOLD on the other hand, is the change in
> > > > > oxygenation
> > > > > > due to change in bloodflow due to change in energy
> consumption
> > > due
> > > > > > change in activity. Hence the effect is very indirect and
> shows
> > > a
> > > > > > maximum after some 5s. E.G. AM Dale et al has shown some
> nice
> > > use
> > > > > of
> > > > > > the BOLD and EEG/MEG so there no reason to discard one,
> just
> > > use
> > > > > with
> > > > > > care.
> > > > > >
> > > > > > P $B<><)<><?<><)<#<(<></ (Bl
> > > > > >
> > > > > > -----Opprinnelig melding-----
> > > > > > Fra: eeglablist-bounces at sccn.ucsd.edu
> > > > > > [mailto:eeglablist-bounces at sccn.ucsd.edu] P
> $B<><)<><?<><)<#<(<></
> > (B
> > vegne av
> > > > Makoto
> > > > > > Miyakoshi
> > > > > > Sendt: 10. april 2009 11:25
> > > > > > Til: eeglablist at sccn.ucsd.edu
> > > > > > Emne: [Eeglablist] Meaning of ITC
> > > > > >
> > > > > > Dear experts,
> > > > > >
> > > > > > Let me take a question about a basic physiology.
> > > > > > In my intuition, ERSP (i.e. EEG power) is related to
> neuronal
> > > > > > activation, either firing frequency or number of firing
> cells.
> > > On
> > > > > the
> > > > > > other hand, ITC (i.e. inter-trial EEG phase) is NOT
> directly
> > > > > related
> > > > > > to them. So, by observing ITC, we may be observing some
> already
> > > > > > systematized temporal responsibility of the given network,
> > > which is
> > > > >
> > > > > > too complicated and abstract to imagine, compared to the
> case
> > > of
> > > > > > neuronal firing simply increasing/decreasing.  Does it
> mean,
> > > then,
> > > > > ITC
> > > > > > is a totally different index than, for example, BOLD?
> > > > > >
> > > > > > Makoto
> > > > > > _______________________________________________
> > > > > > Eeglablist page:
> http://sccn.ucsd.edu/eeglab/eeglabmail.html
> > > > > > To unsubscribe, send an empty email to
> > > > > > eeglablist-unsubscribe at sccn.ucsd.edu
> > > > > > For digest mode, send an email with the subject "set digest
> > > mime"
> > > > > to
> > > > > > eeglablist-request at sccn.ucsd.edu
> > > > > >
> > > > >
> > > > >
> > >
> > >
> >
> >
> > _______________________________________________
> > Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> > To unsubscribe, send an empty email to
> > eeglablist-unsubscribe at sccn.ucsd.edu
> > For digest mode, send an email with the subject "set digest mime"
> to
> > eeglablist-request at sccn.ucsd.edu
> >
> 