Brain Connectivity Supporting Language Recovery in Aphasia

Abstract
Aphasia is one of the most common neurological deficits after a stroke, typically resulting from injury to
cortical brain regions related to language processing in the dominant hemisphere. However, many individuals
with aphasia can exhibit language impairments that are out of proportion to the degree of gray matter injury,
with severe deficits from relatively smaller subcortical lesions, or less severe deficits in spite of relatively large
lesions. This discrepancy is frequently attributed to the notion that language relies not only on the integrity of
gray matter regions, but also on the white matter pathways supporting their ability to act in concert.
Nonetheless, white matter disconnections beyond the necrotic or gliotic post-stroke brain lesions are not
always measured or taken into account in models of brain-behavior relationships. To fully understand the
neurobiology of aphasia, brain damage should be quantified as the combination of direct necrosis / gliosis as
well as cortical disconnection. The overarching purpose of this research proposal is to comprehensively map
residual white matter networks in stroke survivors to determine their role in the neurobiology of language
processing and aphasia recovery. Using newer advancements in structural neuroimaging, our group developed
Connectome-Lesion Symptom Mapping (CLSM) to test specific questions related to aphasia mechanisms and
aphasia recovery. During the first cycle of this project, this research yielded 15 high-impact peer-reviewed
publications. Based on this success, the novel research proposed in this project will build on these
achievements to evaluate three independent new conceptual topics related to aphasia: we will define
multimodal network dynamic modeling approaches to elucidate the relationship between structural and
functional neuronal network integrity post-stroke, including direct and indirect neuronal communication, and
their relationship with aphasia (Aim 1). The dual stream model is a promising new theoretical framework for
language processing, however, it is still an oversimplification and our recent data suggests that each stream is
composed of finer grained sub-networks. Using the connectome approach, we will define the sub-networks that
form the dorsal and ventral streams of language processing (Aim 2). We will determine stream-specific white
matter microstructural network plasticity supporting aphasia recovery after treatment (Aim 3). To accomplish
these aims, we will leverage a large baseline behavioral and imaging chronic aphasia dataset from the Center
for the Study of Aphasia Recovery (C-STAR) (n-199) (Aims 1 and 2), and data from the ongoing treatment
study Predictor of Outcome of Language Rehabilitation (POLAR) (n=150) (Aim 3). Overall, this project will build
on connectome and network science to advance translational and personalized research in aphasia. It will
advance knowledge on neuroimaging methods, provide mechanistic information about language processing,
and determine markers for therapy-related language improvement.

Grant Number: 7R01DC014021-12
NIH Institute/Center: NIH
Principal Investigator: Leonardo Bonilha