+- Apnea Board Forum - CPAP | Sleep Apnea (https://www.apneaboard.com/forums)
+-- Forum: Public Area (https://www.apneaboard.com/forums/Forum-Public-Area)
+--- Forum: Software Support Forum (https://www.apneaboard.com/forums/Forum-Software-Support-Forum)
+--- Thread: OSCAR Export of Raw Per-Graph Data (/Thread-OSCAR-Export-of-Raw-Per-Graph-Data)
OSCAR Export of Raw Per-Graph Data - Nonsanity - 11-25-2019
The current CSV export options in OSCAR focus mainly on log, summary, and event-type data. I, and probably others of an analytical frame of mind, would find it handy to be able to export the raw data from each of the displayed graphs.
In my case, I see a distinctive 4 Hz signal in my mask pressure data that has an excellent correlation with my apnea events—better, even, than what OSCAR flags. I want to experiment with automated detection of this signal. For that, I need to extract the raw mask pressure data.
I’ve started writing my own extractor from the SD card data files, and I’ll probably have that long done before this feature could be added to OSCAR, but it seems worth asking for it anyway.
For the export interface, I imagine a fourth radio button for “Raw Data” that, when selected, adds a pull-down menu to select one of the graphs by name. The export file’s default name would also reflect this choice. The data would be CSV with an initial header text line. (Extra points for a binary export option as well.)
The CSV data could be imported into, for example, Excel for special analysis. Or something more creative and specific, like my plan to save it as a WAV file and open it in Audacity to be displayed as a frequency spectrogram—the better to visualize my 4 Hz signal.
(I suppose I could ask for a spectrogram display option to be built into OSCAR itself, but I don’t want to ask for the world...)
RE: OSCAR Export of Raw Per-Graph Data - pholynyk - 11-25-2019
I don't think you want to see many hours of a 25 Hz, or even a half Hz, 16 bit signal transformed into an ASCII csv file. It's going to be pretty huge, for no real benefit. If a .wav file can handle such low frequencies - half Hz ? - it would be a better choice.
The edf format is actually pretty good for handling this kind of data, and ResMed and, now F&P, use it for flow and pressure data. One of the links on the edfinfo web site is to a pretty nice display package, and there may be some analysis packages there, as well.
RE: OSCAR Export of Raw Per-Graph Data - DeepBreathing - 11-25-2019
Nonsanity Wrote:In my case, I see a distinctive 4 Hz signal in my mask pressure data that has an excellent correlation with my apnea events—better, even, than what OSCAR flags.
Welcome to Apnea Board. I think that 4Hz signal would be the forced oscillation technique (FOT) that Resmed uses to distinguish between obstructive and clear-airway apneas. AFAIK only Resmed uses it, and not on all machines (eg the ASV machines don't have it). I don't know if the machine data includes a marker for FOT starting - I doubt it.
Regarding what Oscar flags - it only reports the data that was output by the machine, so if the machine flags an event then so will Oscar.
RE: OSCAR Export of Raw Per-Graph Data - RayBee - 11-25-2019
(11-25-2019, 06:45 PM)DeepBreathing Wrote:Nonsanity Wrote:In my case, I see a distinctive 4 Hz signal in my mask pressure data that has an excellent correlation with my apnea events—better, even, than what OSCAR flags.
Welcome to Apnea Board. I think that 4Hz signal would be the forced oscillation technique (FOT) that Resmed uses to distinguish between obstructive and clear-airway apneas. AFAIK only Resmed uses it, and not on all machines (eg the ASV machines don't have it).
That's exactly what I saw almost two years ago during a titration study on a ResMed AirCurve 10 VAuto. I was wondering what it was back then and found the same info online, that ResMed does this to determine a central apnea by testing for a clear airway. And no, my ASV does not do it.
Mine old SleepyHead data looked like this (4Hz ripple on the Flow Rate and Mask Pressure graphs). I thought it was a crafty and cool technique they used.
RE: OSCAR Export of Raw Per-Graph Data - Nonsanity - 11-25-2019
I bought the BiPAP version of the ResMed 10 S and worked out the pressure values that worked for me while waiting to see my doc for the results of the sleep test. After showing him my data (thanks to OSCAR) he told me to keep using my settings. They took me down to <1 AHI from my test result of 88. They had been having trouble working out what I needed from the test, anyway. Jumping right to BiPAP probably made it easier to identify tolerable settings that worked after only four night’s fiddling.
I suspected that the 4 Hz signal was from the machine, though the doc thought it was me. I pushed the question aside as moot—the signal was there, that was enough.
I suppose my day job of working with voluminous LiDAR data prevents me from thinking much of multi-gigabyte data files.
I’ll withdraw my request, complete my EDF parser, and have it output a WAV file. It seems like a hack, but there are quite a few audio analysis options that would then open up. Hearing my sleep is just a curious bonus.