Event Abstract

Back to Event

Establishing brain-to-behaviour prediction models of post-stroke aphasia: A systematic investigation of brain parcellations, multimodal imaging, and machine learning algorithms.

Ajay D. Halai1*, Anna M. Woollams2 and Matthew A. Lambon Ralph1
  • 1 MRC Cognition and Brain Sciences Unit, University of Cambridge, United Kingdom
  • 2 Neuroscience and Aphasia Research Unit, School of Biological Sciences, University of Manchester, United Kingdom

In recent decades, structural and functional neuroimaging have radically improved our understanding of how speech and language abilities map to the brain in normal and impaired participants, including the diverse, graded variations observed in post-stroke aphasia (Butler et al., 2014; Halai et al., 2017). Despite the potential for a paradigm shift in neuroscience theory and clinical practice, only recently have a handful of studies begun to explore the reverse inference: creating brain-to-behaviour prediction models in post-stroke aphasia (Halai et al., 2018, Hope et al., 2015, Hope et al., 2018; Pustina et al., 2017, Yourganov et al., 2016, Yourganov et al., 2015). In order to establish some key foundations for successful prediction models, in this large-scale study we systematically investigated four critical issues to determine the optimal: (1) behavioural measures to use as targets; (2) partitioning of the brain space for use as predictive features; (3) combination of structural and connectivity measures from multimodal neuroimaging; and (4) type of machine learning algorithms to generate predictions. There is increasing agreement that binary aphasia classifications are limited; furthermore, while it is possible to predict performance on individual neuropsychological tests, any assessment taps multiple underlying component abilities and hence performance across tests is often inter-correlated over patients (Halai et al., 2018). An alternative approach places patients as points in a continuous multidimensional space, where the axes represent primary neuro-computational processes. Therefore, we explored the influence of the core model factors while predicting four principal dimensions of language and cognition variation in post-stroke aphasia. There has not been any formal comparison on how models perform with different brain parcels (i.e. anatomical, functional or theory driven). Similarly, although many machine-learning algorithms exist, it is not clear if one should be preferred to others in a particular context (i.e. support vector is widely used but it is not suitable for multi-modal data). Finally, recent studies have produced conflicting evidence for the utility of adding structural connectivity information to typical clinical MRI scans (Pustina et al., 2017; Hope et al., 2018). The goal for this study was to determine the ‘optimal’ recipe to prediction post-stroke aphasia deficits. The results showed that across all four behavioural dimensions, we consistently found that the best prediction models were derived from structural measures extracted from T1 scans, which suggested that adding information on white matter connectivity (in vivo patient-specific diffusion weighted data) did not improve the models. The winning models used a larger number of parcellations and submitted to a multi-kernel learning algorithm, which conducts as many local multivariate models as there are parcels. This enhances the sensitivity of the models by assigning zero weights to poorly contributing parcels. Our results provide a set of principles to guide future work aiming to predict outcomes in language disorders from brain imaging data. From a clinical implementation perspective, it suggests that the majority of the information needed for effective outcome prediction in stroke aphasia can be obtained using standard clinical MR scanning protocols, obviating the need for acquisition and processing of diffusion weighted data.

Acknowledgements

We are especially grateful to all the patients, families, carers and community support groups for their continued, enthusiastic support of our research programme. This research was supported by grants from The Rosetrees Trust (A1699) and ERC (GAP: 670428 - BRAIN2MIND_NEUROCOMP).

References

Butler, R. A., Lambon Ralph, M. A., & Woollams, A. M. (2014). Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. Brain, 137(12), 3248-3266. Halai, A. D., Woollams, A. M., & Lambon Ralph, M. A. (2017). Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of speech fluency, phonology and semantics. Cortex, 86, 275-289. Halai, A. D., Woollams, A. M., & Lambon Ralph, M. A. (2018). Predicting the pattern and severity of chronic post-stroke language deficits from functionally-partitioned structural lesions. NeuroImage: Clinical, 19, 1-13. Hope, T. M. H., Leff, A. P., & Price, C. J. (2018). Predicting language outcomes after stroke: Is structural disconnection a useful predictor? NeuroImage: Clinical, 19, 22-29. Hope, T. M. H., Parker Jones, Ō., Grogan, A., Crinion, J. T., Rae, J., Ruffle, L., . . . Green, D. W. (2015). Comparing language outcomes in monolingual and bilingual stroke patients. Brain, 138(4), 1070-1083. Pustina, D., Coslett, H. B., Ungar, L., Faseyitan, O. K., Medaglia, J. D., Avants, B., & Schwartz, M. F. (2017). Enhanced estimations of post‐stroke aphasia severity using stacked multimodal predictions. Human Brain Mapping, 38(11), 5603-5615. Yourganov, G., Fridriksson, J., Rorden, C., Gleichgerrcht, E., & Bonilha, L. (2016). Multivariate Connectome-Based Symptom Mapping in Post-Stroke Patients: Networks Supporting Language and Speech. The Journal of Neuroscience, 36(25), 6668-6679. Yourganov, G., Smith, K. G., Fridriksson, J., & Rorden, C. (2015). Predicting aphasia type from brain damage measured with structural MRI. Cortex, 73, 203-215.

Keywords: Aphasia (language), prediction, Principal compenent analysis(PCA), multivariate pattern analysis (MVPA), Model performance

Conference: Academy of Aphasia 57th Annual Meeting, Macau, Macao, SAR China, 27 Oct - 29 Oct, 2019.

Presentation Type: Platform presentation

Topic: Not eligible for student award

Citation: Halai AD, Woollams AM and Lambon Ralph MA (2019). Establishing brain-to-behaviour prediction models of post-stroke aphasia: A systematic investigation of brain parcellations, multimodal imaging, and machine learning algorithms.. Front. Hum. Neurosci. Conference Abstract: Academy of Aphasia 57th Annual Meeting. doi: 10.3389/conf.fnhum.2019.01.00092

Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.

The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.

Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.

For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.

Received: 05 May 2019; Published Online: 09 Oct 2019.

* Correspondence: Dr. Ajay D Halai, MRC Cognition and Brain Sciences Unit, University of Cambridge, Cambridge, England, CB2 7EF, United Kingdom, Ajay.halai@mrc-cbu.cam.ac.uk