Error in RELAX_drift

[izzycourtney26]

Hello,

I'm currently trying to implement RELAX to process data from a go-nogo experiment. RELAX ran for 60 out of 70 of the data files, but for the other 10, this error keeps happening:

Attempt to grow array along ambiguous dimension.

Error in RELAX_drift (line 145)
CumulativeExceedsUpperBoundPerEpoch(1,e)=NaN;

Error in RELAX_Wrapper_beta (line 425)
[continuousEEG, epochedEEG] = RELAX_drift(continuousEEG, epochedEEG, RELAX_cfg); %
Use epoched data to add periods showing excessive drift to the mask

Error in pop_RELAX_beta (line 600)
RELAXProcessing_ICA_AllParticipants, RELAXProcessingRoundThreeAllParticipants,
RELAX_issues_to_check, RELAXProcessingExtremeRejectionsAllParticipants] =
RELAX_Wrapper_beta (RELAX_cfg);

All of our data is pretty noisy, including the ones that RELAX crashed on, but they don't seem that much noisier than the others. Any advice would be much appreciated!

[NeilwBailey]

Hi there I haven't seen that error before, so I don't have a ready-made solution sorry. I would recommend saving the pre-processing steps for one of the files causing the issue (which will save the data just prior to this error), then sequentially applying the functions in RELAX_Wrapper_beta that occur after the blink cleaning to that saved file until RELAX_drift. Then applying the code within RELAX_drift to the file showing the error until the line that causes the error. Then inspect the relevant variables in the workspace. It looks like there might be a size mismatch between the following two variables: epochedEEG.RELAX.EpochsMarkedAsNaNsForExtremeEvents and CumulativeExceedsUpperBoundPerEpoch. Although I'm not sure why that might occur. Hopefully inspection of the variables and code leading up to the error will help. If you follow these steps and gain a helpful insight, but still haven't managed to solve the problem, let me know and I might be able to help further. Otherwise, if your study design allows it, you could set a higher high-pass filter for all your data (1Hz would likely take care of the drift that is causing the issue). Alternatively, the error seems to be occurring in the lead-up to the MWF cleaning. If the wICA_ICLabel setting provides sufficient cleaning by itself, you could consider only applying that cleaning (in reasonably clean data, I think wICA_ICLabel might be the preferred method. Although it leaves more artifacts in the data, that cost might be more than offset by a better preservation of signal). This wouldn't resolve the problem, but instead might by-pass it so that you don't need to solve it (while still providing sufficient data pre-processing if your experimental design allows for these settings). Kind regards, Neil

…

On Sat, Mar 23, 2024 at 3:13 AM izzycourtney26 ***@***.***> wrote: Hello, I'm currently trying to implement RELAX to process data from a go-nogo experiment. RELAX ran for 60 out of 70 of the data files, but for the other 10, this error keeps happening: Attempt to grow array along ambiguous dimension. Error in RELAX_drift (line 145) CumulativeExceedsUpperBoundPerEpoch(1,e)=NaN; Error in RELAX_Wrapper_beta (line 425) [continuousEEG, epochedEEG] = RELAX_drift(continuousEEG, epochedEEG, RELAX_cfg); % Use epoched data to add periods showing excessive drift to the mask Error in pop_RELAX_beta (line 600) RELAXProcessing_ICA_AllParticipants, RELAXProcessingRoundThreeAllParticipants, RELAX_issues_to_check, RELAXProcessingExtremeRejectionsAllParticipants] = RELAX_Wrapper_beta (RELAX_cfg); All of our data is pretty noisy, including the ones that RELAX crashed on, but they don't seem that much noisier than the others. Any advice would be much appreciated! — Reply to this email directly, view it on GitHub <#17>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGM6ZADOJAB76DRUSAVK5N3YZRKDLAVCNFSM6AAAAABFDTH3RCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGIYDEOBUGYYDOMA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[izzycourtney26]

Thanks! I stepped through the RELAX_Wrapper function and saw that for those subjects, after the call to RELAX_epoching in line 393, the size of epochedEEG.RELAX.EpochsMarkedAsNaNsForExtremeEvents was bigger than the number of epochs in epochedEEG. I was able to resolve the issue by setting OnlyIncludeTaskRelatedEpochs to 0 instead of 1.

[NeilwBailey]

Great to hear you were able to resolve the issue, thank you for letting me know.