Perhaps a few well-placed stacks of rebels, perhaps not full stacks, but very-experienced and/or well-equipped small ones will deter the AI from certain paths. Also, since it will undoubtedly diminish over time with the various assaults, the path will eventually open up.
For example: 4 units of experienced Celtic Professional Swordsmen with cavalry support in the mountain pass north of Venice will keep the Romans at bay until the Sweboz and Getai can develop into that void north and east of the Alps. Then the three can struggle over Pannonia, I would guess somewhere around 180 BC.
Only if there is a way to keep the rebels from moving, d'accord. Have them garrison a fort, perhaps? I think the general could be given a hidden trait that limits his movement to zero? It could be a lot of fun researching what minor tribes were in various areas and discussing which ones would help game-play. It beats having a permanent 'wall of forests', or invisible desert walls that are unrealistic. I can't imagine (though I may well be wrong) that historically Carthage left it's eastern border un-garrisoned, which is what the AI might do if it thought a barrier was there keeping Egypt out. I think Carthage should suffer to keep some sort of garrison to block a, however unrealistic, advance from the Ptolemies. Maybe that's a contributing factor as to why they didn't struggle there? Not my AoE.
I imagine a huge force of experienced desert warriors south of Cyrene would keep the two major factions down there from engaging in this 'desert war' I have heard so much about.
There are also 'region debilitations' such as troop losses to disease or heat available in MTW2, from remaining for extended periods of time in the desert or swamp (such as the Pripet Marshes or the desert crossings) Mountain passes in and around Bactria could have the same effect.