Functional mapping of language pathways in the brain:
a combined intraoperative electrical stimulation and neuroimaging approach
Language processing mechanisms have been a matter of extensive research for the last 150 years. While classical language organization models suggested the involvement of specific brain areas in language production, comprehension, repetition and existence of a “concept centre”, the contemporary models hypothesize an alternative view. Accordingly, language acts upon a division of labour between semantic and phonological processing and relies on dorso-ventral brain organization. Importantly, the substantial role of WM matter structural connectivity has been highlighted in these contemporary, dual stream models.
Electrical stimulation mapping (ESM) allows identifying critical functional sites at both cortical and WM levels in patients undergoing awake brain surgery. Up to date, a large number of ESM works have elucidated discrete cortical areas involved in speech and language processing. However, the evidence provided by these studies with regard to WM connectivity and its relevance for language processing is yet to be explored.
In this thesis we focused on investigating the involvement of the left perisylvian WM pathways in language processing, using ESM combined with structural imaging in patients with selective left-hemispheric intrinsic brain lesions. I present our results in form of four articles: three disclosing our findings on structure-to-function relationships and one reporting patients’ language and cognitive results before and after the awake surgery. In study 1, we have observed that electrical stimulation applied at the level of the left frontal white matter tract (FAT) interfered with correct performance in verb generation task, suggesting the involvement of FAT in language production. FMRI and DTI in vivo dissections corroborated with the intraoperative observations. Studies 2 & 3 both supported the dual stream model of language processing, evidencing that the electrical stimulation applied at the level of arcuate fasciculus (AF; dorsal pathway) disrupted a correct transfer from speech sound to articulatory representations, whereas the same stimulation applied at the level of the level of the inferior fronto-occipital and inferior longitudinal fasciculi (IFOF/ILF; ventral pathways) resulted in specific difficulties in semantic matching. Additionally, symptom-lesion mapping revealed involvement of areas located in the left temporal and frontal lobe in semantic processing and track-wise analyses revealed relationships between the integrity of the left IFOF integrity and performance on one of the semantic processing tasks (semantic pairs). The final study in this patients’ series (N=63) revealed that before surgery age, educational level, tumor grade and lesion size may all influence patients cognitive and language performance. Importantly, the tasks tapping phonological and semantic processing allowed detecting impairments that passed unnoticed in standard assessment. After surgery, all patients declined in naming and repetition. In addition, patients with low-grade gliomas and patients under chemoradiation appeared to be the most vulnerable to decline in comprehension and working memory, also post-surgically. Finally, changes in verbal fluency task suggested that at long-term certain aspects of executive function may substantially improve in (one year after surgery).
All the studies presented here substantially help in elucidating the role of the left perisylvian WM connectivity in language processing. Moreover, they offer a set of clinical applications for the intra- and perioperative care of neurosurgical patients. Although future studies are needed, we suggest that the task of verb generation may be useful for language monitoring in surgeries compromising the integrity of the left FAT. Nonwords repetition appeared to be the most suitable for the intraoperative monitoring of AF integrity, whereas semantic pairs task may be used to monitor the integrity of IFOF/ILF. The results outlined from patients’ follow-up recommend which tasks are the most suitable for patients’ follow-up and highlight which facets of language and cognition should be taken into account in patients’ neurorehabilitation programs.