Closed cymatn closed 2 years ago
Dear Cyrille,
Which version were you using before? Probably a very old version. There were updates to clean_rawdata that had conflicts, so this option was removed because of that about a year ago. Who told you to use that option? Copying @MakotoMiyakoshi here.
Arno
Hello! I have the same question as Cyrille.
" 'availableRAM_GB' is for clean_rawdata1.10. For any newer version, it will cause error. "
Using the same parameters, the final results from clean_rawdata() fluctuate after ASR. It is demonstrated that adding availableRAM_GB fixes this (in the link below), but now we can't use it if we are using the newer version of clean_rawdata?
https://sccn.ucsd.edu/wiki/Artifact_Subspace_Reconstruction_(ASR)
Is there any way fix to this? This is a huge problem for replicability. Any help would be much appreciated.
Thank you! Armen
Could you try resetting the random seed before calling the function. Also better call cleanArtifact directly which has more options.
We will try to add back the option. @MakotoMiyakoshi maybe you could help?
Dear all,
I re-uploaded my final version (ver. 1.10) here as a workaround. The behavior of the algorithm is identical except for options such as support for Riemanian geometry. https://sccn.ucsd.edu/wiki/Artifact_Subspace_Reconstruction_(ASR)#Download_ver_1.10_.28old.29
Arno, unfortunately I am not available for investigating the problem.
Makoto
On Thu, Jan 7, 2021 at 1:23 PM Arnaud Delorme notifications@github.com wrote:
Could you try resetting the random seed before calling the function. Also better call cleanArtifact directly which has more options.
We will try to add back the option. @MakotoMiyakoshi https://urldefense.com/v3/__https://github.com/MakotoMiyakoshi__;!!Mih3wA!XCnhau4_mp1y3Sxgn1-WSlug_74C6nnjWbN0Y7xLDm7dNWjur38D1N78dPOSpjHxIQ$ maybe you could help?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://urldefense.com/v3/__https://github.com/sccn/clean_rawdata/issues/15*issuecomment-756393572__;Iw!!Mih3wA!XCnhau4_mp1y3Sxgn1-WSlug_74C6nnjWbN0Y7xLDm7dNWjur38D1N78dPMAD9-t8A$, or unsubscribe https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AK7GVNLMMBB75B6FHYXSKZLSYYQ5RANCNFSM4O2UGV7A__;!!Mih3wA!XCnhau4_mp1y3Sxgn1-WSlug_74C6nnjWbN0Y7xLDm7dNWjur38D1N78dPN9JPnmBw$ .
I have created version 2.3, which now deliver reproducible results
http://sccn.ucsd.edu/eeglab/plugins/clean_rawdata2.3.zip
Let me know if it all works fine for you and I will release it
Thanks Arnaud! I believe this is working. But it's hard to tell because first I am removing bad channels using clean_rawdata and it is also giving me inconsistent results. I ran the same file 10 times through clean_rawdata. Sometimes it says there are 3 bad channels, sometimes 4, sometimes 5. Is there a similar fix that can deliver reproducible results for the other functions in clean_rawdata such as removing bad channels? And then I'm assuming this may be a problem for the additional removal of bad data periods function?
Now closing this report. I have added back availableRAM_GB but not 100% it works as before.
Hi,
I recently updated to 2.2 from within EEGLAB, and 'availableRAM_GB' now gives the following error.
EG = clean_rawdata(EEG, -1, -1, -1, -1, 20, 0.25, 'availableRAM_GB', 8); Error using clean_rawdata Too many input arguments.
It seems to be because ARS_calibrate does not accept it as one possible input anymore. I was wondering if it has been replaced by some alternatives (or default value), or why this had been removed.
Thank you, Cyrille