Language networks in aphasia and health: a 1000 participant Activation Likelihood Estimate analysis

Abstract

Aphasia recovery post-stroke is classically and most commonly hypothesised to rely on regions that were not involved in language premorbidly, through neurocomputational invasion or engagement of quiescent homologues. Contemporary accounts have suggested, instead, that recovery might be supported by under-utilised areas of the premorbid language network, which are downregulated in health to save neural resources ( variable neurodisplacement). Despite the importance of understanding the neural bases of language recovery clinically and theoretically, there is no consensus as to which specific regions are activated more consistently in post-stroke aphasia (PSA) than healthy individuals. Accordingly, we performed an Activation Likelihood Estimation analysis of language functional neuroimaging studies in PSA and linked control data. We obtained coordinate-based functional neuroimaging data for 481 individuals with aphasia following left hemisphere stroke (one third of which was previously unpublished) and for 530 healthy controls. Instead of the language network expanding by activating novel right hemisphere regions de novo post-stroke, as would be predicted by neurocomputational invasion/quiescent homologue engagement mechanisms of recovery, we found that multiple regions throughout both hemispheres were consistently activated during language tasks in PSA and controls. Multiple undamaged regions were less consistently activated in PSA than controls, including domain-general regions of medial superior frontal cortex and right fronto-temporal cortex. In the reverse direction, the right anterior insula and inferior frontal gyrus were more consistently activated in PSA than controls, particularly for executively-demanding comprehension tasks. These regions overlap with control networks known to be recruited during difficult tasks in healthy individuals and were more consistently activated by patients during higher than lower demand tasks in this meta-analysis. Overall, these findings run counter to neurocomputational invasion of the language network into new territory or engagement of quiescent homologues. Instead, many parts of the pre-existing language network are less consistently activated in PSA, except for more consistent use of spare capacity within right hemisphere executive-control related regions (cf. variable neurodisplacement). This study provides novel insights into the language network changes that occur post-stroke. Such knowledge is essential if we are to design neurobiologically-informed therapeutic interventions to facilitate language recovery.