Metsfan, I apologize because I'm going to talk around you with Rich and Vaughn for a minute. The data posted is presented very well, but it's confusing to me, so feel free to ignore the conversation here, because I need to speculate on what's happening.
Rich, I disagree the events indicated as UA are central. They look obstructive. The machine goes all the way to 25 cm and there is no flow response until the recovery breath. As usual, it is very unique and I have not seen this same flow pattern before. Let's use the graph below and describe what is happening (I hope you and Vaughn are in a collaborative mood because I'm not all that certain of my interpretation).
At 14:30 exhale transitions into an inhale which terminates in a flat line (0-flow) and no exhale until the UA is flagged at 14:45, followed by strong inhale and then a weak 2-part exhale. The next inhale is followed by a partial quick exhale and zero flow from 14:56 to 14:47 at the UA flag slight exhale followed by strong inhale. This is repeated, with most apena appearing to be breath holding after inhale. The machine increases pressure to 25 and cannot make a dent in the flow. This is pretty much a hallmark of obstruction, as there is plenty of pressure in the mask to induce inhale in an open airway.
The mask pressure shows the machine attempts to maintain breathing rhythm and volume, but the user does not respond in spite of an apparent PS of about +15.
Since significant parts of the night do not exhibit any apnea or hypopnea, I speculate that the periods of apnea are strongly related to sleeping position. These are obstructive apnea following inhale, and do not respond to any pressure stimulus.
Could a lower EPAP pressure encourage exhalation, and prevent the breath holding? Is higher pressure needed to prevent OA? I dunno!
What we can say is the the apena are all 10 to 15 seconds in length, and may not actually result in significant SpO2 desaturations. I'm interested in Vaughn's insights as well.