While brainstorming... could this be a bad/clogged injector? Or leaky injector o-ring? If one cylinder was being starved of fuel as demand went up then the O2s would still look about right (I think). What strikes me as weird is the KR (90 % of the time) seems to start when the IPWs hit the 7.5-8.0ms range.
Just throwing ideas out there. I'm going to pull the fuel rail and visually inspect what I can this weekend.