One of the audio engineering wisdoms of guitar playing is that only comparatively skimpy voicings like power chords go well with heavy distortion because of cross-distortion.
Now one obvious solution would be to just use one pickup per string and send each signal through its own independent distortion unit, but all units sharing the same settings (namely, being "ganged"). This should be much more direct and "natural" than trying to achieve the same feat via signal processing (namely picking the signal apart into its constituents which are then independently distorted).
Of course, a similar result may be achieved by sample-based play: using a polyphonic instrument with "distorted guitar" samples rather one that first puts together clean guitar samples and then adds distortion should be about the same. Experiences/comparisons there?