Hello,
> Our EPIs had few signal drop out areas.
>
> I performed 1st level and got significant signals on the area where was
> adjacent to the signal drop out area.
>
> Could this significant signal be the result of our fMRI experiment? Or just
> some artifacts?
>
> How should I discern such significant signals from artifacts?
Typically areas of significant signal dropout will get masked out by
the 1st level estimation proportional thresholding. So generally I
wouldn't be too concerned that any in-mask effects would be due to
artifacts. For example, we reliably observe significant activations
in sentence-processing tasks in posterior inferior temporal and
fusiform regions adjacent to areas of signal dropout.
Obviously, having group results might also help, as if you're seeing a
consistent group effect in the same region it also lessens the
likelihood that the results are influenced by some strange artifact.
Finally, a generally good thing to do is probably do some sort of
timeseries analysis to identify possible spikes in the data. There
are many toolboxes out there to help with this, including tsdiffana
(http://imaging.mrc-cbu.cam.ac.uk/imaging/DataDiagnostics).
That all being said, if you're still worried, it would probably help
to know more about the specific task design, and your analysis steps,
etc.
Hope this helps!
Best regards,
Jonathan
|