[FieldTrip] dftinvert option on ft_preprocessing

[Schoffelen, J.M. (Jan Mathijs)]

Hi Yasir,

The dftinvert option is not deprecated.

If you would have played detective a bit, and don’t shy away from reading some code, you’d see that this option operates on line 465 in https://urldefense.com/v3/__https://github.com/fieldtrip/fieldtrip/blob/master/private/preproc.m__;!!HJOPV4FYYWzcc1jazlU!5BqIlA_JvJpo-_fVl4JnnB59sByGtAK2i3zgm7Sql6fVZffKshSMZ1jfXKXVynXNl9jIyDlCgBIwTb7Uj1xwK7_am93mtI3Lm8Iehg$ . 
In other words, this option returns the actual estimate of the powerline fluctuations. 

My assumption is that, in combination with the option cfg.boxcar and cfg.rectify, cfg.dftinvert = ‘yes’, will result in an estimate of the envelope of the powerline signal. This can be used to evaluate the sensor-by-sensor contamination. 

Whether or not you can omit this analysis step from your processing pipeline, that’s up to you.

Note that additional documentation about the megconnectome pipeline is in a document with reference documentation https://urldefense.com/v3/__https://www.humanconnectome.org/software/hcp-meg-pipelines__;!!HJOPV4FYYWzcc1jazlU!5BqIlA_JvJpo-_fVl4JnnB59sByGtAK2i3zgm7Sql6fVZffKshSMZ1jfXKXVynXNl9jIyDlCgBIwTb7Uj1xwK7_am93mtI39rx7yUw$ , and that there’s a dedicated HCP e-mail discussion list, which is probably the more appropriate venue for questions about these data.

Good luck with your analysis,

Jan-Mathijs


> On 24 Jan 2023, at 02:14, Yasir Çatal via fieldtrip <fieldtrip at science.ru.nl> wrote:
> 
> Hello,
> 
> I was reading the HCP MEG data preprocessing pipeline in megconnectome to adapt to my MEG data. In the pipeline hcp_datacheck.m lines 241-248,
> 
> % for the time course of the line noise
> cfg2           = [];
> cfg2.dftfilter = 'yes';
> cfg2.dftfreq   = lfreq + [-1 0 1].*(hdr.Fs./1024);
> cfg2.feedback  = 'none';
> cfg2.dftinvert = 'yes';
> cfg2.rectify   = 'yes';
> cfg2.boxcar    = 0.2;
> 
> This cfg2 is an input to ft_preprocessing on line 308:
> 
> tmp = ft_preprocessing(cfg2, tmpdata);
> 
> I specifically wanted to ask about dftinvert and boxcar options. These options seem to be deprecated and I couldn't even find them in the fieldtrip-r10442 which comes with megconnectome. May I ask what the aim was of these options? Would my preprocessing be affected if I leave these out?
> 
> Thanks,
> Yasir
> _______________________________________________
> fieldtrip mailing list
> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!9_4KKPjOqhyB5h7zucw1S-dBoHTw1Hqsg5FHl2FDB6oNKtGdP8dckzZMrVOhia03DxWPvsceLlye9HcrRdcBXWPYQLXaiXNbkPr6Ng$