LAUSR

Gajardo-Vidal et al. reported data from a sample of 134 stroke patients with relatively circumscribed damage to the left frontal lobe. Contrary to the received view, mapping lesion to deficit using an ecologically-valid measure of speech production (i.e. spontaneous connected speech elicited with a picture description task) revealed that, irrespective of lesion extent, damage to Broca’s area does not contribute to long-term speech production outcome, whereas damage to the white matter in the vicinity of the anterior part of the arcuate fasciculus (AF) is strongly implicated. Our results therefore address a matter of critical clinical importance: understanding the causes of inter-patient variability in speech production outcome post-stroke. We welcome the opportunity to reply to two Letters to the Editor on our paper. Both Letters rely on a source of data different from ours: direct electrical stimulation (DES) and/or resection carried out during neurosurgical procedures on awake patients (e.g. with glioma). A key behavioural outcome measure in such work is the presence or absence of speech arrest (e.g. an inability to name a visually presented picture) following DES over a given cortical region. Inferences drawn from different clinical populations and methodological approaches that converge on a common scientific understanding (consilience) bolster confidence in that understanding, and so we strongly endorse the contribution of research in neurosurgical patients to the existing body of knowledge regarding the regions involved in speech production. Consilience is important because different methods have different strengths and weaknesses. For instance, the challenge for lesion-deficit mapping research in stroke patients is to dissociate the contribution of brain regions that are commonly damaged together. In the case of DES, an observed behavioural effect is not necessarily the consequence of direct stimulation of the targeted region; it may instead reflect the contribution of stimulation that spreads along white matter pathways into remote regions. In the case of resection (and DES), the absence of a behavioural effect might reflect functional reorganization that has occurred overtime prior to neurosurgery, due to an underlying neurological condition (e.g. tumour or epilepsy). This contrasts with the sudden onset of stroke where functional reorganization is not expected to occur until after the insult. In what follows, we identify the main contentions of each Letter first and then address them one by one. Mandonnet and Duffau propose that prior data, acquired in patients undergoing neurosurgery, already refute the relevance of Broca’s area to persistent speech production impairments. They highlight, for example, the work of Benzagmout et al. where focal resection of Broca’s area only resulted in transient speech production impairments. They also enjoin the community to understand the apparent neglect of such data. We recognize the perpetuation of the myth of Broca’s area as the critical region for speech production, but challenge the decisiveness of current neurosurgical data with respect to the research question we addressed. For example, the Benzamount et al. data may establish that damage to Broca’s area is not sufficient to cause long-lasting speech production impairments, but they do not refute its necessity in combination with surrounding regions (including the underlying white matter), which was the focus of our work.