Start with a good canonical reference: a textbook, a college syllabus (though these are getting harder to find), an encyclopedia article. Look at the references.
Even if you're looking for recent material, those references should be useful as principle studies or background will itself be referenced in more recent research.
Citation counts themselves are at best a highly uneven quality indicator. The real test of research is replication: a study turns up phenomena which are independently validated by other researchers. This of course takes time.
Good current research tends to come from teams and institutions which have produced substantive work in the past. So the references / citations approach I'd opened with should also give you names of individual researchers and/or institutions which have had high-impact work in the field Their more recent work will likely also be of interest. This approach will tend to underrepresent novel work from new teams ... but that work is also often less likely to be significant or robust.
If you're a practitioner in the field yourself, and have the equipment / means to test findings, validating results yourself is another option. Mind the various biases which can come into play here.
