This is a paper that I've been citing and talking about for sometime. I was interested in it originally because it clearly showed a dissociation between the ability to comprehend speech and perform syllable discrimination, the task effect that David and I have been talking about for a decade. I recently re-read the paper because Gabriele Miceli, another major player in aphasia research in the 1970s and 1980s (and still very active!), pointed out to me that the article is also quite relevant to the issue of the relation between perception and production of speech, as the title previews.
What the authors did was measure VOT in aphasics' own speech utterances (they read a set of words aloud) and compared it their ability to discriminate syllables that differ in terms of the same VOT dimension. You can see the implication for motor theories: If a patient cannot reliable produce correct VOTs, then this should affect their ability to perceive (discriminate) VOT.
So what did they find? I'll let Blumstein et al. summarize:
It is quite clear, at least for the anterior aphasics (Broca and Mixed Anterior), that the ability to perceive the VOT continuum relates in no way to the ability to produce voiced and voiceless stops. Thus, the anterior aphasics maintain the ability to perceive this distinction, but make both phonemic as well as phonetic substitutions.
This is a cool result because is zooms in on one feature, VOT, and shows a direct non-correspondence between perception and production of this speech dimension.
Blumstein, S., Cooper, W.E., Zurif, E.B., & Carmazza, A. (1977). The perception and production of Voice-Onset Time in aphasia Neuropsychologia, 15 (3), 371-372 DOI: 10.1016/0028-3932(77)90089-6