[FieldTrip] Motor beta activity - DICS solution more noisy than sensor data?
Eelke Spaak
eelke.spaak at donders.ru.nl
Fri Nov 22 08:41:35 CET 2013
Hi everyone,
Thanks very much for your great input so far! Actually, as Jörn
suggested, I did the very straightforward check (thanks for the tip :)
) of running one of our tutorial test scripts
(test_beamforming_extended), and it turns out this one does not
produce the same results as depicted here:
http://fieldtrip.fcdonders.nl/tutorial/beamformingextended . So, most
likely somewhere a bug has been introduced...
Hopefully I can find out what it is today and fix it. Will keep you posted!
Best,
Eelke
On 22 November 2013 08:06, "Jörn M. Horschig" <jm.horschig at donders.ru.nl> wrote:
> Hi Eelke,
>
> since everyone jumped on the train, here my 2 cents:
> To verify whether this is a newly introduced bug, maybe run a tutorial test
> script that includes beamforming. If they look alright, it gets more likely
> that it is you or your data and not fieldtrip :) It's not definite evidence
> of course though. SinceVitoria also experiences strange things, it might be
> something worthwhile to investigate.
>
> My initial guess from the plots is that there is something wrong with the
> forward model. All unit problems should have been resolved, but just to be
> sure you could check whether all objects are in the same unit (make it 'cm'
> as the grads are).
>
> Best,
> Jörn
>
> Charidimos Tzagarakis wrote:
>>
>> Eelke,
>> Thinking again about my second suggestion (regarding individual
>> variability) I actually can't think of a case where this could realistically
>> produce what you get. On the other hand, looking at TF maps per subject and
>> channel (on the "helmet" layout), normalised with a "rest" epoch, may help
>> spot something unusual.
>> Best,
>> Haris
>>
>> Charidimos [Haris] Tzagarakis MD, PhD, MRCPsych
>> University of Minnesota Dept of Neuroscience and Brain Sciences Center
>>
>>
>>
>> On 21 November 2013 18:09, Charidimos Tzagarakis <haristz at gmail.com
>> <mailto:haristz at gmail.com>> wrote:
>>
>> Hi Eelke,
>> Provided there is no major recent revision of the DICS code, I
>> would have expected motor desynchronisation to show up pretty
>> well. Are the maps shown at source and channel level straight
>> differences of L and Right hand conditions at the beta band (I
>> hope I am correctly interpreting your paradigm) ? If so it might
>> be helpful in pinpointing the problem/as a sanity check to see
>> what happens when you use beta desynchonisation (ie change
>> relative to the baseline) instead for each condition, and see
>> source/channel maps of that separately for L and R and then when
>> you take the difference. I suppose the main element this checks
>> for is whether L and R conditions have the same baseline.
>> This doesn't immediately explain why source and channel results
>> are different but in the absence of any other clues it may be a
>> way to 2ble check the whole process.
>>
>>
>> Another point to consider is that, although beta changes should
>> appear in all subjects, it is possibly true that there are
>> individual differences in the actual beta range and frequency bin
>> of maximum effect. If you are using the same settings for all
>> subjects when you beamform with DICS you may be missing some of
>> the effect (true, this is also the case for channel data but there
>> may be subtle differences that add up - there are many voxels and
>> few channels). I believe it may be useful to see what happens when
>> you run the beamformer tailored to each subject's particular beta
>> characteristics (ie change the "foi" for each subject, keep the
>> tapsmofrq the same - possibly smaller) and then combine everything
>> (you'll need of course to come up with a relative metric such as
>> perc. change when you combine all subjects to account for the
>> slightly different frequencies you used )
>>
>> Best,
>> Haris
>>
>> Charidimos [Haris] Tzagarakis MD, PhD, MRCPsych
>> University of Minnesota Dept of Neuroscience and Brain Sciences Center
>>
>>
>>
>> On 21 November 2013 10:36, Eelke Spaak <eelke.spaak at donders.ru.nl
>> <mailto:eelke.spaak at donders.ru.nl>> wrote:
>>
>> Fellow FieldTrippers,
>>
>> Currently I am looking at a contrast for left- versus
>> right-hand index
>> finger button presses. As expected, on sensor level (combined
>> planar
>> gradient, grand average) I see a clear lateralisation in beta band
>> power starting at least 0.5s before the button press (see
>> https://db.tt/Rtch3Qjy). Both 'blobs' are significant; there is
>> clearly more beta power ipsilateral to the response hand. I would
>> prefer to do further analyses on source level, so I attempt to
>> reconstruct the sources for this effect using DICS beamformer
>> (common
>> filter, applied to both conditions separately; fixedori and
>> realfilter
>> = 'yes'). The grand average results for this (again contrast
>> left vs
>> right response hand) are shown at https://db.tt/IBQZG0d8 .
>> (Ignore the
>> R/L-flip, this is radiological convention.)
>>
>> As you can see, the source level solution is much more blurry
>> than on
>> sensor level. This picture is without using any regularisation
>> (lambda
>> parameter), the results are even worse when I use lambda =
>> '5%'. The
>> negative blob (right hand higher power than left) becomes
>> 'marginally
>> significant' on source level (p ~ 0.06) where it was p < 0.001 on
>> sensor level. The positive blob is nowhere near significant.
>> Also, the
>> individual results are much less topographically consistent on
>> source
>> than on sensor level (explaining the worse statistics).
>>
>> I have checked the segmentation of my MRIs, the 'gray' seems to be
>> nicely within the head all the time. Also, I have manually
>> verified
>> the alignment of headmodel, sourcemodel, and gradiometer
>> information
>> for all subjects.
>>
>> As a final note, the above sensor-level plot was taken from a
>> 'slice'
>> out of a planar-gradient time-frequency analysis (mtmconvol). The
>> ingredient for the beamformer was an mtmfft fourier spectrum
>> on the
>> axial gradiometer data, obtained for just the time-frequency
>> range of
>> interest (subselect toilim [-0.5 0], mtmfft foi = 23,
>> tapsmofrq = 7).
>> When I compute condition-averaged power based on these fourier
>> spectra
>> and look at the contrast, the results are again as expected:
>> https://db.tt/n2P3UKcQ (of course less localised because of axial
>> gradient vs planar). The freq structures underlying this
>> contrast are
>> exactly the same as those going into ft_sourceanalysis, so the
>> problem
>> must be in the source analysis step (and/or in the preparation
>> of the
>> geometric information, although these seem fine by visual
>> inspection).
>>
>> Does anyone have any idea that might explain these seemingly
>> contradictory results? I would have expected demixing to improve
>> signal-to-noise ratio, rather than worsen it.
>>
>> Thanks!
>> Best,
>> Eelke
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl <mailto:fieldtrip at donders.ru.nl>
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
More information about the fieldtrip
mailing list