Let my try to stretch the analogy to anticyclonic tornados: The vast majority of all tornados move cyclonic, i.e. counterclockwise on the Northern Hemnisphere of our blue planet. There is e.g. a publication on anticyclonic tornados by Howard Bluestein et. al. which says:
It is also possible that the source of vorticity in the anticyclonic tornadoes is not from a parent mesoanticyclone (produced via tilting), but rather from (hypothesis 2) preexisting anticyclonic shear vorticity on the anticyclonic-shear side of the low-level jet associated with the rear-flank downdraft [...]
In simple words: It is not fully clear yet, how exactly these atypical atmospheric vortices form, but there are different possible explanations why tornados exist which rotate against the direction supported by Coriolis force. This reminds me somehow of the Coriolis Force Effect on Drains:
The notion that the Coriolis force determines which direction water spirals down drains is one of the most prominent scientific myths.
Some educational German television show has been able to actively influence the rotation of water spiraling down a drain with appropriate boundary conditions like little wings in the drain (but the clip is not online). In the metereology of planets with solid surface, the effect of the ground on the atmospheric motion is usually termed orographic effect. I suppose the different rotations of the Juno IR image of Jupiter's north pole have similiar "guide wings" in the lower atmosphere, most probably density fluctuations.
Another important factor is that eddies in fluids live really long, plus these eddies can interact with each other. I remember listening to a scientific talk on some kind of "algebra of eddies", but I did not manage to figure out the reference. However, a similar idea can be found within the references of a preprint on the network of solar flares. The idea of this kind of meta-modelling of Navier-Stokes equation is to save computer-time and nevertheless predict eddy dynamics. What that means for your questions: It is not too unprobable that a single anticyclonic eddy formed, and due its interactions with the other eddies in the storm system, other anticyclonic eddies could have been branched of this one. Let me conclude with the words of @AtmosphericPrisonEscape
One can speculate much and say it should have been A or it should have been B, but unless those scenarios are tested in a 3D global circulation model, nobody will know for sure.