Dataset Information

Enhanced estimations of post-stroke aphasia severity using stacked multimodal predictions.

ABSTRACT: The severity of post-stroke aphasia and the potential for recovery are highly variable and difficult to predict. Evidence suggests that optimal estimation of aphasia severity requires the integration of multiple neuroimaging modalities and the adoption of new methods that can detect multivariate brain-behavior relationships. We created and tested a multimodal framework that relies on three information sources (lesion maps, structural connectivity, and functional connectivity) to create an array of unimodal predictions which are then fed into a final model that creates "stacked multimodal predictions" (STAMP). Crossvalidated predictions of four aphasia scores (picture naming, sentence repetition, sentence comprehension, and overall aphasia severity) were obtained from 53 left hemispheric chronic stroke patients (age: 57.1?±?12.3 yrs, post-stroke interval: 20 months, 25 female). Results showed accurate predictions for all four aphasia scores (correlation true vs. predicted: r?=?0.79-0.88). The accuracy was slightly smaller but yet significant (r?=?0.66) in a full split crossvalidation with each patient considered as new. Critically, multimodal predictions produced more accurate results that any single modality alone. Topological maps of the brain regions involved in the prediction were recovered and compared with traditional voxel-based lesion-to-symptom maps, revealing high spatial congruency. These results suggest that neuroimaging modalities carry complementary information potentially useful for the prediction of aphasia scores. More broadly, this study shows that the translation of neuroimaging findings into clinically useful tools calls for a shift in perspective from unimodal to multimodal neuroimaging, from univariate to multivariate methods, from linear to nonlinear models, and, conceptually, from inferential to predictive brain mapping. Hum Brain Mapp 38:5603-5615, 2017. © 2017 Wiley Periodicals, Inc.

SUBMITTER: Pustina D

PROVIDER: S-EPMC5765865 | biostudies-literature | 2017 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Enhanced estimations of post-stroke aphasia severity using stacked multimodal predictions.

Pustina Dorian D Coslett Harry Branch HB Ungar Lyle L Faseyitan Olufunsho K OK Medaglia John D JD Avants Brian B Schwartz Myrna F MF

Human brain mapping 20170807 11

The severity of post-stroke aphasia and the potential for recovery are highly variable and difficult to predict. Evidence suggests that optimal estimation of aphasia severity requires the integration of multiple neuroimaging modalities and the adoption of new methods that can detect multivariate brain-behavior relationships. We created and tested a multimodal framework that relies on three information sources (lesion maps, structural connectivity, and functional connectivity) to create an array ...[more]

PMID: 28782862

Similar Datasets

Project description:Language is not a single function, but instead results from interactions between neural representations and computations that can be damaged independently of each other. Although there is now clear evidence that the language profile in post-stroke aphasia reflects graded variations along multiple underlying dimensions ('components'), it is still entirely unknown if these distinct language components have different recovery trajectories and rely on the same, or different, neural regions during aphasia recovery. Accordingly, this study examined whether language components in the subacute stage: (i) mirror those observed in the chronic stage; (ii) recover together in a homogeneous manner; and (iii) have recovery trajectories that relate to changing activation in distinct or overlapping underlying brain regions. We analysed longitudinal data from 26 individuals with mild-moderate aphasia following left hemispheric infarct who underwent functional MRI and behavioural testing at ∼2 weeks and ∼4 months post-stroke. The language profiles in early post-stroke aphasia reflected three orthogonal principal components consisting of fluency, semantic/executive function and phonology. These components did not recover in a singular, homogeneous manner; rather, their longitudinal trajectories were uncorrelated, suggesting that aphasia recovery is heterogeneous and multidimensional. Mean regional brain activation during overt speech production in unlesioned areas was compared with patient scores on the three principal components of language at both the early and late time points. In addition, the change in brain activation over time was compared with the change on each of the principal component scores, both before and after controlling for baseline scores. We found that different language components were associated with changing activation in multiple, non-overlapping bilateral brain regions during aphasia recovery. Specifically, fluency recovery was associated with increasing activation in bilateral middle frontal gyri and right temporo-occipital middle temporal gyrus; semantic/executive recovery was associated with reducing activation in bilateral anterior temporal lobes; while phonology recovery was associated with reducing activation in bilateral precentral gyri, dorso-medial frontal poles and the precuneus. Overlapping clusters in the ventromedial prefrontal cortex were positively associated with fluency recovery but negatively associated with semantic/executive and phonology recovery. This combination of detailed behavioural and functional MRI data provides novel insights into the neural basis of aphasia recovery. Because different aspects of language seem to rely on different neural regions for recovery, treatment strategies that target the same neural region in all stroke survivors with aphasia might be entirely ineffective or even impair recovery, depending on the specific language profile of each individual patient.

Project description:Background and Purpose The severity of aphasic impairment in chronic stroke survivors is typically thought to be stable by 6 months postonset. However, a recent study showed that stroke survivors with aphasia experience language improvement or decline in the chronic phase, years beyond onset. Little is known about why some individuals improve whereas others remain stable or decline. Additionally, no study has tracked changes in aphasia from assessments completed at multiple time points across many years. The current study offers a comprehensive analysis of potential predictive demographic and health information to determine which factors predict dynamic changes in aphasia severity in chronic stroke. Methods Individuals in the chronic stage of a single-event, left-hemisphere ischemic stroke were identified from an archival database and included for study ( N = 39). Participants were included if they had undergone 2 or more standardized language assessments acquired at time points at least 6 months apart, with the 1st assessment at least 6 months postinjury. A linear mixed-effects model was used to determine the impact of treatment and a variety of demographic and health factors on language change. Results Over time, half of the participants improved (51%), whereas approximately a quarter (26%) decreased, and a quarter (23%) remained stable. A greater number of aphasia treatment hours significantly predicted language improvement ( p = .03), whereas older stroke age was associated with long-term decline ( p = .04). Two interactions were found to be significant in predicting improvement in individuals with diabetes: Increased exercise and younger age at stroke were significant in predicting outcomes ( p < .05). Conclusions Factors that significantly influence language recovery in chronic aphasia include stroke age and receiving aphasia treatment. For those with diabetes, increased exercise was shown to improve outcomes. Results from this study offer clinicians greater insight into the influence of patient factors on long-term recovery from stroke aphasia while suggesting a potential adjunct to language therapy: exercise. Supplemental Material https://doi.org/10.23641/asha.7849304.

Dataset Information

Enhanced estimations of post-stroke aphasia severity using stacked multimodal predictions.

Publications

Enhanced estimations of post-stroke aphasia severity using stacked multimodal predictions.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets