
mac again , O2's are operated open / closed loop .Open loop is startup to a pre-determined time-out and/or engine warmup plus adequate O2 warmup ( most O2's are electrically "heated") - during this time , they are OUT of the equation .Whearas your problem seems to be in this mode . After so much time that the O2's do come into play would seem to not be when your problem is at it's height - for it WILL take time to recover even after initiating "closed loop" (normal running -all sensors + O2's IN the equation) - as you say it becomes fine.
Next point is if your O2 test turns out Ok ,intermittantly ?? ,yet at one time showed in OBDII (i'm guessing)badly - the inspection of plugs can especially show that one BANK of cylinders is "off" and to go ahead and further check O2's (you have front&rears ea. side?) on a specific side ; but again if in closed loop you got a "greenlight" on the O2's for both banks ,you see how now O2' seem a distant possibility and a plug inspection , and I forgot to mention a very easily done COMPRESSION Chk. at this time . Just so you can avoid the jump to "shotgun approach" too soon(very expensive & no real logic either that way!!)

Many different very serious problems do "clear up" after warmup ; coolant leak into cyl. onto a spkplug ,oil splattering(valve seals or rings) onto & fouling a plug - until they're hot and "burn-off" the oil ,head gasket losing seal - only to gain tightness after aluminum "expansion" closes up the problem , on, and,on and on!.Plugs wil help indicating this - even your new plugs - nobody is saying they were'nt replaced when you did ,it's just the FACT that all plugs do tell a STORY !!
Respectfully Glen 2/10/18