Audio / Sound Files

[SleepyTimeTime]

Out of curiosity, I started making audio recordings of my sleep using SnoreLab. It's an impressive app and it's useful to both hear what I'm doing while sleeping and see a chart of dB level vs time.

Unfortunately, you can't export data from SnoreLab, so I put my Olympus voice recorder to use with an external microphone which I tape to my headboard above my head. I am getting about 128 MB of sound files per night.

There is excellent agreement between the audio data viewed and played back in Audacity and the Flow Rate chart in Oscar. On the audio data, I can clearly see Central Apnea patterns and apneas occurring. These match very nicely to the Event Flags and Flow Rate in Oscar.

I can find events during the night when I lift the mask in my sleep to wipe away my mouth or scratch my nose. These always show us as gray zones on the Leak Rate chart. In my 12 days of using CPAP and OSCAR, the program has not once misidentified a leak as an apnea event.

I can hear what are probably arousals when I toss and turn over. I hear an occasional cough.

Most of the time, I am not snoring. For the 12 nights of data collection, I have averaged 44 Centrals/night vs 6 Obstructives/night. Typically, I do have night time yawns or big inhalations/exhalations. SnoreLab confirms this.

Questions to sleep experts and OSCAR Team:

1. Do you see any added benefit to recording sleep using the voice recorder? Or is it just barely a curiosity for me at this stage?
   
2. If you think there IS added benefit to capturing 64 or 128 kbps audio files, can you add a feature to OSCAR to synchronize and play audio files directly within OSCAR? Right now I have to run the two programs simultaneously and switch between them. I could see another row of sound data with the existing rows and the option to play or not play the sound associated with the area where you are zoomed in.
   

[sawinglogz]

I know there are some sleep monitoring devices that have a microphone used simply to measure dB in an effort to detect snoring.

Are you finding the audio itself useful, or just the dB chart?

[SleepyTimeTime]

I'm finding both the chart and the audio to be very useful. On the chart, I see normal rhythmic breathing and I can see it degrade to Cheyne-Stokes Respiration or apneas. It's easy on the sound chart to find irregularities, i.e., sounds that are obviously not normal respiration or CSR. With Audacity, it's a simple click to play the sound at the location where the chart shows something is going on. I manually captured the times of these tosses & turns for a couple of nights. The breathing chart from the sound recording and the respiration flow on OSCAR tell the same story. There's greater detail in the OSCAR chart. The audio chart tells me when I'm tossing & turning or lifting the mask in my sleep.

I've been using an "O2 Ring" pulse oximeter which has an accelerometer and indicates the times of motion. So I can correlate motion with both the sound and the O2 Ring. I can get more information about motion from the sound recording than from the O2 Ring (Note: I'm changing to a Contec pulse oximeter so I can import that data into OSCAR).

I'm also using an Emfit QS detector under the mattress to provide an indication of sleep class (which I doubt it is very accurate).

I'm trying to correlate SpO2, sleep classes, and tosses & turns to the onset of CSR. So far, the only thing I've been able to conclude is that a half to full dose of Unisom (doxylamine succinate antihistamine) really exacerbates both apneas and CSR. When I suffer from a few consecutive nights of insomnia, I take a Unisom to try and break the cycle. That looks like something to avoid.

I'm new to all this and trying to find the best set of tools that might provide me some guidance on improving sleep. Right now it seems the best thing I can do is keep up the CPAP therapy (AirSense 10) and try to avoid Unisom. 

I think I'll change back to using 5 or 10 mg of Melatonin to test it out.

[sawinglogz]

We're hoping to add support for the O2 Ring style oximeters in the near future, though I don't know how accurate they are: I think fingertips react to desaturation faster than the base of your fingers. Also, the new models only report every 4s; maybe that's why?

I've also heard that one should avoid sedatives with centrals. I've found the occasional VERY low dose (about 1/6 the "official" dose of 2 pills!) of benadryl to be a nice way to reset things. It doesn't throw my centrals too far out of whack, and it helps reset disrupted sleep.

Maybe someone will take up the banner to add an audio importer for OSCAR (along with adding playback). I already have too many things on my plate! As an aside, I find using a spectrogram in Audacity a very good way to visually identify distinctive sounds.

[SleepyTimeTime]

sawinglogz...thanks for the tip on benadryl to do a reset after a bout of insomnia. I'll give that a try. I've had pretty good luck weaning myself off the half-tab Unisom after starting CPAP therapy.

Glad you are going to add support for ring-style oximeters.

OSCAR (and previously SH) are amazing tools. I'm really impressed with the signals analysis that is able to identify CSR, OA, CA, RERA so reliably. That is not trivial filtering!

[jaswilliams]

The event detection is done by the machines not OSCAR

[A KLERK]

The event detection is done by the machines not OSCAR

But... OSCAR shows them very nicely. You will never miss one!

[sawinglogz]

...well, you won't as of the next release...