Positron emission tomography was used to investigate changes in regional cerebral blood flow (rCBF) in neurologically normal subjects during word reading and word repetition. The blood flow in these conditions was compared with control conditions where subjects were presented with stimuli of comparable auditory and visual complexity to real words and said the same word on presentation of each stimulus. The control condition for word repetition (hearing spoken words presented backwards) resulted in bilateral activation of the superior temporal gyrus. Word repetition caused a significant increase in rCBF over this control condition in the left superior and middle temporal gyri. The control condition for word reading (seeing stimuli written in ‘false fonts’, i. e. non-existent letter-like forms) resulted in significant changes in rCBF bilaterally in the striate and extrastriate cortex. Word reading caused a significant increase in blood flow relative to this control in the posterior part of the left middle temporal gyrus. The implications of these results are discussed, and it is argued that they are consistent with localization of a lexicon for spoken word recognition in the middle part of the left superior and middle temporal gyri, and a lexicon for written word recognition in the posterior part of the left middle temporal gyrus.

The neural systems involved in hearing and repeating single words were investigated in a series of experiments using PET. Neuropsychological and psycholinguistic studies implicate the involvement of posterior and anterior left perisylvian regions (Wernicke's and Broca's areas). Although previous functional neuroimaging studies have consistently shown activation of Wernicke's area, there has been only variable implication of Broca's area. This study demonstrates that Broca's area is involved in both auditory word perception and repetition but activation is dependent on task (greater during repetition than hearing) and stimulus presentation (greater when hearing words at a slow rate). The peak of frontal activation in response to hearing words is anterior to that associated with repeating words; the former is probably located in Brodmann's area 45, the latter in Brodmann's area 44 and the adjacent precentral sulcus. As Broca's area activation is more subtle and complex than that in Wernicke's area during these tasks, the likelihood of observing it is influenced by both the study design and the image analysis technique employed. As a secondary outcome from the study, the response of bilateral auditory association cortex to ‘own voice’ during repetition was shown to be the same as when listening to ‘other voice’ from a prerecorded tape.

Brain activity during reading tasks was investigated using PET. The aim was to account for differences in the results of two previous studies [those of Petersen et al. (Science 1990; 249: 1041–4) and Howard et al. (Brain 1992; 115: 1769–82)] by systematically varying the type of reading task and the exposure duration of the word stimuli. Both variables strongly influenced patterns of brain activity. There were three types of task: (i) reading aloud; (ii) reading silently; and (Hi) lexical decision on visually presented words and pseudowords. Reading aloud and reading silently engaged the left middle and superior temporal regions, confirming the important role of these areas in visual word processing.The areas principally engaged during lexical decision were the left inferior and middle frontal cortices and the supplementary motor area; activity in these areas suggests that the subjects were using a phonological strategy to perform the task. There was also a significant effect of exposure duration, with activity being greater for short (150 ms) exposure durations than for long (1000 ms or 981 ms) exposure durations. We conclude that until we understand how subtle variations in experimental design influence brain activity during reading tasks, the association of specific processing functions with individual anatomical areas activated during reading is premature.

Processing of speech is obligatory. Thus, during normal speech comprehension, the listener is aware of the overall meaning of the speaker’s utterance without the need to direct attention to individual linguistic and paralinguistic (intonational, prosodic, etc.) features contained within the speech signal. However, most functional neuroimaging studies of speech perception have used metalinguistic tasks that required the subjects to attend to specific features of the stimuli. Such tasks have demanded a forced‐choice decision and a motor response from the subjects, which will engage frontal systems and may include unpredictable top‐down modulation of the signals observed in one or more of the temporal lobe neural systems engaged during speech perception. This study contrasted the implicit comprehension of simple narrative speech with listening to reversed versions of the narratives: the latter are as acoustically complex as speech but are unintelligible in terms of both linguistic and paralinguistic information. The result demonstrated that normal comprehension, free of task demands that do not form part of everyday discourse, engages regions distributed between the two temporal lobes, more widely on the left. In particular, comprehension is dependent on anterolateral and ventral left temporal regions, as suggested by observations on patients with semantic dementia, as well as posterior regions described in studies on aphasic stroke patients. The only frontal contribution was confined to the ventrolateral left prefrontal cortex, compatible with observations that comprehension of simple speech is preserved in patients with left posterior frontal infarction.