Machine Learning Aphasia
Morphosyntactic Production in Stroke-induced Agrammatic Aphasia: A Cross-linguistic Machine Learning Approach.
Valantis Fyndanis (photo: UiO/Mathias Fossum)
About the project
The aim of the project is to gain insight into aphasia, a condition characterized by language/communication problems as a result of brain damage. In particular, Machine Learning Aphasia focuses on stroke-induced agrammatic aphasia, which usually occurs following damage to Broca's and neighbouring areas in the left hemisphere of the brain, and investigates grammatical (morphosyntactic) aspects of sentence production related to the verb, such as subject-verb agreement (e.g., Every morning John walks to work), tense/time reference (e.g., Yesterday John went to the cinema), and sentential negation (e.g., George does not like chocolate).
Verb-related morphosyntactic impairment is one of the hallmarks of agrammatic aphasia. Although many studies on morphosyntactic production in aphasia have been conducted thus far, little is known on the factors that determine the relative preservation or impairment of a given verb-related morphosyntactic phenomenon/category in a given person with agrammatic aphasia (PWAA) in a given language.
Inspired by the developments in machine learning, the project aims at filling this gap by taking an original and innovative methodological approach. Machine Learning Aphasia addresses two important, yet unanswered questions:
- Which factors determine the level of performance of a given PWAA, native speaker of a given language, on verb-related morphosyntactic production?
- What is the hierarchy of factors/predictors of successful verb-related morphosyntactic production in agrammatic aphasia?
A large number of subject-, morphosyntactic category- and experiment-specific factors will be considered. More than 100 Norwegian-, Italian-, Greek-, English- and Russian-speaking PWAA will be tested.
Addressing the questions above will advance our understanding of the complexities underlying morphosyntactic production in agrammatic aphasia, which will be a significant contribution to cognitive science. This study will also have significant clinical implications, as the findings about the best predictors of morphosyntactic production in agrammatic aphasia will inform and improve treatment programmes for PWAA.
This project has been funded by the Research Council of Norway (Project #287745 – FRIPRO (FRIHUMSAM) grant). Total budget: NOK 9,993,000.
- David Caplan (Harvard Medical School/Massachusetts General Hospital, Boston, USA)
- Gabriele Miceli (University of Trento, Trento/Rovereto, Italy)
- Olga Dragoy (Higher School of Economics–National Research University, Moscow, Russia)
- Monica I. Norvik (University of Oslo, Norway)
- Haris (Charalambos) Themistocleous (University of Gothenburg, Gothenburg, Sweden & Johns Hopkins University, Baltimore, USA