Dear John,
I'm not sure what is going wrong with the pulse ox peak detection here.
Could you upload your physiological recording file to:
https://oxfile.ox.ac.uk/oxfile/work/extBox?id=68312615463381F4C
However, for PNM (and any RETROICOR-based method) it doesn't matter where in the cardiac cycle it marks the change, as long as that is in a consistent point within each cycle. All the EVs (regressors) that are formed will adjust for this. If you don't get a consistent point in each cycle (e.g. sometimes near the start, sometimes near the end) then this is unacceptable for PNM and you'd need to fix it.
You can also use any other software/method that you like to find peaks and then just create a file that contains a value of one at the peaks and zero elsewhere, and then treat it as a pulse ox trigger. However, I'm hoping that if I can see the data then I can suggest an easy way to fix it within PNM (or maybe you can already use your results from the pulse ox trigger, if they pick out a consistent point in each cardiac cycle).
All the best,
Mark
On 30 May 2013, at 20:00, John Graner <[log in to unmask]> wrote:
> Hello,
>
> I have both a pulse ox recording file and a pulse ox trigger file from our scanner. The graph shown in the HTML file after running the pulse ox recording file through the PNM GUI shows that the vast majority of the peaks in the recording were not detected (the script seemed to identify only the largest of the peaks, with only about 1 out of every 30 being correctly marked).
> The pulse ox trigger file I have seems to correctly mark the change in signal associated with every heart beat, but the triggers occur at the initial rise in the signal rather than at the signal peak.
>
> I have two questions:
> 1) Is there a way to adjust the criteria used by PNM to identify peaks in the pulse ox recording (perhaps change the threshold for peak detection or use a more local peak detection method)?
> 2) Does the "Pulse OX Triggers" option assume the triggers occur at the signal peaks, or can I use the file with triggers marking the initial rises?
>
> Thanks,
> -John
|