The single-case approach of this study, with its high internal validity, contributes to an understanding of the architecture of auditory cognition. However, the generalisability of its findings is constrained given that inferences are largely based on a modest control sample size. Although MR's performance was referenced to a matched control group, the group was marginally more musically experienced than him. There is debate as to whether musical training causes cognitive differences in individuals. According to Marin and Perry (1999), tasks that are overlearned or automatised in musicians may require careful analysis and orderly sequential programming in non-trained individuals. Thus, the possibility exists that comparison to a potentially more musical control group may have exaggerated MR's impairments. A larger, more strictly matched group of controls would mitigate this limitation to some extent. However, it should be noted that the potential inter-individual variance concerning the representation of musical ability in the brain is a problem of music psychology generally (Repp, 1991; Sloboda, 1985). Such uncertainty cannot be ameliorated by using a more representative sample, no matter how large it is.
Converging data of single case studies with previous cases in the literature and experimental studies of normal subjects can give rise to strong conclusions (Peretz, 1993a). However, inconsistencies between studies mean that formal conclusions should not be drawn. Such inconsistencies include different methods of task administration, different lesion aetiologies, and discrepant criteria used to select patients/participants (Andres & Van der Linden, 2001). For instance, Peretz's studies have tended to focus on patients with large bilateral lesions compared with MR's circumscribed unilateral lesion. Another inconsistency concerns variability of the time lag between the onset of the lesions and the testing phase (Andres & Van der Linden, 2001). Although MR was examined after a post-stroke period long enough to exclude the possibility of a "mass effect", MR reported some recovery of his music cognition over the course of the study, such as his slightly improved appreciation of music. If MR's abilities were unstable at the time of testing, unrepresentative results may have been obtained, highlighting the need for ongoing assessment.
A methodological problem of this study concerns the use of some materials which contained imperfections. The instrumental and environmental sound identification tasks were based on poor recordings that may have been distracting or prevented optimal performance from participants. However, there is no reason to suppose that MR would have been more affected than controls, or vice versa.
A potentially informative auditory domain - voice recognition - was not examined in this study, which could form part of future research with MR. His reported difficulty to recognise voices when the speaker is not visible warrants further enquiry in order to complete investigation in the non-verbal auditory domain. The possibility exists for a voice recognition disorder underpinned by impaired pitch perception or sequential streaming. In the area of auditory scene analysis, further testing of MR's sequential and simultaneous streaming for slow tones is necessary for strong inferences to be drawn regarding the influence of auditory scene analysis on musical and speech prosody domains of audition. Another area for further experimental work concerns the analysis of vowels or diphthongs in speech. As these aspects of speech require accurate temporal resolution of pitch perception at a fine-grained level (Griffiths, 2001), MR's impaired fine temporal processing suggests that he may be unable to detect these subtle timing differences in speech stimuli.
A question that arises immediately out of Griffiths' model is why all individuals with acquired-deafness, of the degree shown by MR or others with musical hallucinosis, do not experience musical illusions. If the illusions are due to the activity of a normal module cut off from its normal input, one would expect the same phenomenology from all those with peripheral auditory hearing loss. The answer to this question may lie in a threshold for spontaneous or triggered activity within the perception/imagery module. Although speculative, this notion highlights a possible avenue for future research