Response to Wilson: What Does Motor Cortex Contribute to Speech Perception?

Figure 1. Coarse schematic models of speech perception illustrating the fundamental difference between auditory and motor theories of speech perception. (a) Schematic of an auditory theory. Acoustic speech input activates Although the main goal of our paper [1] was to argue against mirror neurons as a possible instantiation of the Motor Theory of speech, we also presented evidence in support for an alternative auditory theory of speech perception. That is, we promoted a model as in Figure 1a and against that represented in Figure 1b. Wilson [2] does not dispute this central position. Instead he argues that speech production regions could have a top-down influence on perception. We agree wholeheartedly and would add that speech production systems are not the only source of topdown information. As Wilson hints, lexical-semantic information can also influence perception, and visual speech information is known to have dramatic effects [3] – arguably to a much greater extent than motor information. Although some authors attribute the influence of visual speech entirely to motor activity [4], there is evidence that ‘direct’ cross-sensory integration (visual-to-auditory) is the more robust source of influence [5]. It seems that the only point of dispute raised byWilson is one of terminology. We suggested that the motor system is not ‘necessary’ for speechperception;Wilson suggests that it is. By our use of the termwemean that it is possible, at least undersomecircumstances, foraccurate speechperception to occur without the influence of the motor system. Evidence for this claim comes from the fact that even large left frontal lesions that reduce speech production to nil or stereotyped output do not produce considerable impairments in speech recognition [6]; that deactivating the entire left hemisphere in Wada procedures produces mutism yet results in only a 7.5% error rate in discriminating minimal phonemic pairs (hearing ‘bear’ and pointing to a matching picture among phonemic distractors [7]); that the failure to develop speech production does not preclude normal receptive speech development [8,9], and that infants as young as 1-month-old exhibit sophisticated speech perception ability including categorical perception well before they acquire the ability to speak [10]. It is a fair criticism that many studies demonstrating preserved auditory comprehension in Broca’s aphasics do not implement tight controls on contextual information. However, (i) this indicates the auditory system in concert with non-motor contextual cues is sufficient to support accurate speech perception at a high level, further reinforcing our claim that themotor system is not ‘necessary’ even in everyday noisy conditions and (ii), in studies that remove non-motor contextual cues, such as theWada study