I've had this issue before and I'm having it now. I'm thinking it's a combination of power to fuel injectors AND an O2 sensor issue. The original issue years ago was fixed w/ just an O2 sensor but the fuel injector wiring might still been a problem. What stands out for your issue is your graphs and the odd changes in the voltage.... It's a bit of a story but here's where I am:
Years ago it was basically a stock 302 w/ MAF. It would race around NP but when it came to any hill it would starts boggin out and eventually could barely limp. I just pulled over so didn't give it much time to die but I'm certain it would. Even when getting back to all flat road, if this issue happened, it would never recover until I shut it down for about 10-15 minutes. 1st time she ever went to a mechanic and they said she had a broken cam...
I just replaced my O2 sensor w/ a brand new Bosch but I dropped it from like 6 inches before this issue started. I read the box and it said DO NOT DROP IT!!!. A 6" drop can't hurt an O2 sensor!!! BS Yes it can! >:( A new O2 sensor fixed it and I was VERY careful not to drop it :)
Now I've got a brand new everything which should be in my sig and this issue is back. It's all tuned really nicely and hauls ass! Until I hit a f'in hill..... Now I can monitor things w/ QH/TunerPro and I see my A/F Err correction go way out of bounds and keeps going even when I hit a flat road. It runs lean and keeps getting worse until I shut down for a bit then it's happy. ECT is 178-180 solid too.
I also have a brand new Bosch O2 sensor and it did slip out of my fingers installing it from about 6inches above the ground again....
Also my MAF conversion was soldered in beautifully over 20 years ago and the solder joints still look beautiful. PO def did a solid bench job. I checked continuity on either side of the joints and the continuity was very low despite she can haul ass. One injector wasn't working at all which is why I discovered the wiring issue. It would barely light up a noid/test light.
I fixed the 1 injector's wiring and the noid light lights up the whole garage now. I did not bother with the others tho. I was able to get everything tuned and running AWESOME!!! (more fine tuning left) and that's when I began the hill tests and failing again...
Now I can see stats w/ the tuner, I figured it was going into some Open Loop Limp Mode but it doesn't; it's Closed Loop the whole time. I'm using Decipha's default log data set too.
To your graph concern and possibly the relation to my issue, I see the mVolts for the injectors (running state/bottom graph) typically is very mountainous and constantly changing when it's happy but when it craps out, it's more hills and valleys with extreme changes WHEN it does change. I also see the Volts for the O2 sensor (Dyno/top graph) react at the exact same time so mountainous or hills/valleys, both O2 and Injectors coincide all the time. When it craps out, it tends to be extreme variances vs. normally they tend to sit "somewhere in the middle". I posted 2 graphs below, 1st is when it craps out, the 2nd is when it's running normal. I have a lil more fine tuning w/ MAF AD Counts but you can see it's pretty good.
I wonder if you see something similar?
Current Theory at least for me: My O2 sensor probably failed due to the VERY short drop AND my injector wiring is heating up more on the hill due to load and killing the power connection.