Hello Stefan,
The <blah>_ts.txt files are the values from the input data - e.g. in the case of a group study, the “timeseries” is the subject estimates from the lower-level analysis. The values in the report.txt are the values for the higher-level model which has been run, and are the outputs from featquery that you wish to report. For higher-level analyses the timeseries values may potentially have similar scalings to the values in the report.txt, but are _not_ the outputs from the model.
Hope this helps,
Kind Regards
Matthew
> On 17 Jun 2016, at 10:18, Stefan <[log in to unmask]> wrote:
>
> Hello everybody,
>
> Sorry for this adapted re-post (https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;6555892b.1606), but I am still struggling with the output of featquery.
>
> I am interested in the mean parameter estimate and in the individual parameter estimates of my subjects. To this end, I used an ROI (finding from a group-level contrast) and extracted parameter estimates for this COPE in featquery.
>
> If I understand correctly, the mean parameter estimate is given in the column "mean" in "image statistics" in the html-output (and of course, in the respective "report.txt" file). I also thought (and was told and read in some threads) that the individual parameter estimates for this region can be found in the "mean time series" (or "mean_mask_ts.txt" file) or, likewise, if I am not interested in the individual averages for the whole ROI but rather in the peak maximum then I would look up in "max_pe1_ts.txt".
>
> However, I am not sure if this is correct, because I noticed that if I calculate the mean of these individual estimates it's not the same as the mean parameter estimate given in the "mean" column of "image statistics" (furthermore the latter seems to better capture the effects what I see in the zstat-maps, etc.). I wonder how this could be? Or am I wrong in assuming that the individual estimates can be found in the "mean_mask_ts.txt" file? (but what are these?) And how can I extract the correct estimates? In some threads I also found that people extract them with featquery on the second level for each participants separately, but I thought one already has them in the featquery output for the group-level COPE.
>
> Thanks in advance!
>
> Best, Stefan
|