I am a novice to EEG and I collected some data to look at which position correlated with performance best during a cognitive task (to compare to another technique). I have significant correlations with performance in one position but this is occuring across all frequencies which doesn't make sense to me! Can anyone explain to me why this might be happening? Does this sound like something I've done worng in the analysis? Or could anyone point me in the direction of some literature which might help me explain this? Thank in advance!
Hi,
First of all, you need to explore the literature related to your task, as we don't know it (ie, we don't know what are you looking for and what other scholars achieved previously).
Second, did you get the result you mentioned before cleaning your data up? If no, you need to do so. After that, select the reasonable components and examine the possible correlation(s).
HTH.
Dr. Samer Sarsam
By 'position' do you mean 'time'?
If you have an evoked response (or any other large, rapid event - real or artifactual) locked to the timing of your task in the EEG, it may translate in the frequency domain into a broad band of frequencies, accounting for your observation.
Note that frequency-domain stuff is not independent of the temporal structure of the EEG - it is just a different (and subordinate) characterization of the time evolution of the signals. Thus, you cannot interpret spectral structure without first assessing the temporal structure of your task-related activity.
See for example Yuval-Greenberg et al. Neuron 2008, where 'induced gamma band activity' turned out to be an artifact caused by micro-saccades (which produced short electrical transients, creating broad-band stuctures in the time-frequency domain).
Thank you both for your answers. Dr Sarsam The task was a STROOP and we are interested in the responses during the incongruent trials so I haven't found anything in the literture that would suggest that this is common in a STROOP task. And the result was after cleaning the data do I don't think it was a result of this.
By position I mean electrode position, we were examining just four electrode positions and examined the responses during a 'block' of incongruent trials. Thank you for the paper recommendation I will take a look.
No worries, Kirsty.
Providing some screenshots would give better idea about what's going on. Anywa, it sounds a bit strange if the correlation spans all frequencies. Perhaps there is some artifact (or noise) that is not properly removed? Based on the literature, it would be reasonable to expect that midfrontal theta would correlate with response conflict task performance.
Dr. Samer Sarsam
- Badges
- Science topic
- Similar topics
- Citations