I'm surprised that Milgrom seemed confident about the existence of such pockets - as I understand it, they would exist in only one version of MOND, one that not much favoured. Further, as MOND is non-relativistic, the non-detection of any MOND effect in any such pocket wouldn't test the relevant theory anyway ... you'd need to consider a more complete theory such as TeVeS.

I'd have thought the more difficult part of any experiment to measure MOND in these pockets would be to find them in the first place, not make measurements of the behaviour of gravity in them ...