Hello,
temporal concatenation now uses migp by default which means that the input to the main melodic algorithm is ( essentially ) a PCA of the input data. Depending on exactly what is needed you could regress the output spatial maps against the original session data to obtain timecourse information.
Kind Regards
Matthew
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
> On 16 Oct 2017, at 12:11, Honami Sakata <[log in to unmask]> wrote:
>
> Dear Experts,
>
> Hello.
> I'm trying to analyze my data with FSL 5.0.10 on macOS Sierra version 10.12.6.
> When I perform Multi-session temporal concatenation of MELODIC, the following problems occur;
>
> 1) The maps of each independent component seems to be correct, but time series and power spectra are incorrect.
> Time series becomes looks like electromyogram, and power spectra looks like loose waveform.
>
> 2) The sum of time points may be wrong.
> If 3 sessions, each containing 600 points, are concatenated,
> the time series should be have 1800 time points.
> However, there are only 1200 time points in concatenated time series.
>
> In this case, what should I do?
> I would like to know some information about it.
>
> Best regards,
> Sakata
|