Name of the project
Cognitive and neurolinguistic technologies for personalized learning and rehabilitation of children
Key idea (slogan) of the project
Synergetic Linguistics for the new generation
Arizona State University
University of Helsinki
Global scientific/technological goal (challenge) of the project
The project is consistent with the fourth Priority Directions of Development of science, technology and engineering in the Russian Federation, “Life Sciences”, approved by the Decree of the President of the Russian Federation dated July 7, 2011 № 899 as amended on December 16, 2015.
Academic supervisor of the project
Description of the project
The project addresses the global challenge of Expanding Human Potential. As many people today fall outrageously short of their potential, the project is aimed at developing capabilities latent in everybody by harnessing powerful technologies that accelerate children’s learning and development. Worldwide it is considered urgent to develop new diagnostic and rehabilitation technologies for children with developmental disorders, as well as innovative technologies of personalized learning pathways. The existing practice of writing textbooks and developing tests often lacks scientific foundation of identifying texts complexity level and is based on authors’ intuition. The most urgent issues of Applied Linguistics at present include assessing readability of texts used in educational and examination discourses and selection of cognitively and linguistically appropriate texts which could promote students learning and development. Lower as well as higher than appropriate readability levels of texts used in class or testing inhibit learning and result in distorted interpretations of children competences and state of development. The project envisions design and development of an Automated text complexity analyzer (ATCA) to identify a text complexity level and correlate it with academic levels (ages) of children. As a computational tool, ATCA produces texts indices and predicts comprehension levels which are highly demanded in diagnostic and teaching children with speech disorders. Speech, cognitive, emotional and social development disorders are currently recognized by WHO and UNESCO as most severe challenges in health care. Such disorders are detected in over 1/5 of children and the percentage is steadily increasing. The project implies a large-scale fundamental research of linguistic and psychophysiological mechanisms of language acquisition and addresses the problems of linguistic components of speech disorders of various genesis as well diagnostics and rehabilitation practice. The project also solves the problem of validation techniques of texts used in clinical practice. Linguistic components of psychometric tools, including assessment of speech development, cognitive functioning, pain syndromes, particularly in children and adolescents remain crucial research issues expecting validation in Russian environment. Psychophysiological correlates of speech perception will be assessed with combined use of event-related potentials and measures of induced EEG oscillatory changes providing a more comprehensive view of electrophysiological brain response to speech. The project database is to enable testing such methods of EEG data analysis as deep mining and machine learning algorithms. As a joint international study the current project also fills the existing gap of describing risk factors (sociodemographic, cultural, family, etc) influencing children development. The data are planned to be used to design personalized paths for learning and health care for children with different cognitive and linguistic abilities.
Approaches to the implementation of the project
The main approach applied in the project is integral: text comprehension and speech development in healthy children and children with speech disorders are to be studied with linguistic methods as well as with the use of EEG, MRI, eye tracking, different neuropsychological and neurocognitive tests. The project population comprises monolingual and bilingual (Russian, Tatar) children aged 4 – 18 years. The Project Tests Database is compiled of original tests, designed by the project team, and those developed by other researchers but translated (if necessary), adapted and validated, collaboratively with producers, for the current environment. Prior to be used in the Project and included into the Database for the Analyzer, all testing blocks are piloted, cross-validated on a wide range of population and statistically processed. The Project implies elaborating techniques of investigating expectancy interruptions and mismatch negativity aimed at assessing speech perception.
University’s experience, research and technological preliminary preparatory work
We have already developed a structure of created computerized versions of language assessment instruments, established collaboration with partners abroad, partly launched pilot projects on blocks of tests, got an order from institutions of health care to perform this work and we are working on the diagnosis and treatment of speech disorders of generated speech (within SAE “Translational 7P Medicine”). We have a certain number of instruments for psychometric validation and neurocognitive tests (more than 20 batteries and scales). The partners of the project have developed a set of methods for analyzing induced EEG changes and tested experimental tasks, optimized for the combined use of the method of event-related potentials and induced EEG oscillatory changes, which is of particular importance in the context of the proposed research plan. The projects of this magnitude require extensive neuro-, psycho- and sociolinguistic data on 1000 and more people, including those with disorders of speech development, which is impossible without well-established contacts with a network of institutions of health, education and correctional pedagogy. Within the framework of the projects, carried out by the laboratory soLet, headed by D.McNamara, we have developed and implemented a program of automatic processing of natural languages, such as Coh-Metrix, TERA, SiNLP, TAALES, TAACO, as well as automated tutors for learning iStart and WritingPal. Currently, prof. D.McNamara is engaged in several projects: Using Writing Pal program in adult education (Exploration of Automated Writing Strategy Instruction for Adolescent Writers using the Writing Pal), Modeling social contacts and activities under integrative learning.
Our team has sufficient experience in conducting research and examining the level of test complexity: the members of the team participated in the 5-year project of the European level (TEMPUS) on the development and validation of tests, they are members of the federal team elaborating the Unified State Test, they have experience in transcribing texts and conducting experiments, in creating large electronic corpora of texts; their research results have been presented at a number of international conferences and have been approved by experts, we have established collaboration with foreign partners. Our foreign partners have got experience in identifying the level of English texts complexity.
Achieving of global leadership (excellence), as one of the project results
Global genuine leadership is achieved due to the synergistic character of the research, its strong linguistic component and the highly qualified team able to think and act ‘out-of-the box’. The excellence area is the integration area of computer science, neuroscience, psycholinguistics, cognitive linguistics, language assessment and related studies. The main project outcomes are cost-effective, user-proven and “ready to use” technologies with a high potential to boost personalization in rehabilitation and learning. We benefit from the same foundation and database to create tools to be in common use in education and healthcare.
During the Project the team grasps and develops new insights and competences: methods of Integrative Linguistic to be used in speech disorders treatment and teaching; development of combined psycho-physiological methods and techniques to model neurobiological processes of speech perception, etc. The project findings are to be in-demand: A) in daily practice of neurologists, neurolinguists, clinical linguists, speech pathologists, psychologists, neuropsychologists, physicians, therapists, rheumatologists, general practitioners, pediatricians; B) by designers and developers of psychometric instruments of evaluating various parameters of human activity; C) by language test developers; D) by authors and compilers of textbooks in identifying the target audience of readers (listeners).
Significance, relevance and academic novelty of the project
As multidimensional by character, the Project contributes to health care and education by developing and applying new approaches to linguistic data analysis. Diagnostic and assessment tools designed and accomplished within the Project are valid and reliable as they are based on findings of Cognitive Sciences and Applied Linguistics (Neurolinguistics, Computational Linguistics, Language Assessment, Corpus Linguistics). The novelty of the project is predetermined by its synergetic nature and possibility to use new and old data as well as technologies to expand human potential (a global challenge). The testing site of the Project is expected to become a unique first-class experimental facility due to the combination of a very broad spectrum of techniques used: from traditional to rare psychophysiological methods. The latter, including EEG method, are expected to provide high informative values and precision of measurements. At the moment, speech disorder treatment practices require a computerized diagnostic battery which could automatically generate individual valid stimuli for each new testing. To the best of our knowledge, the database of linguistically valid Russian tests for assessing and identifying the degree of pain in children and teenagers has never been compiled.
Therefore, as practitioners face problems while diagnosing and providing care to patients in pain, there is an urgent need to develop individual therapeutic programs. The corpora of educational and examination discourses compiled as implementation of the project task is to serve as a testing area and a search engine of discourse markers and the unique balance of texts indices determining the level of text complexity. As a publicly available search system it is to become a significant contribution to General and Applied Linguistics. The existing Russian text processers (including Antiplagiat and TextMining) enable only to detect regular discourse markers and calculate quantitative text indices (number of letters, syllables, words, sentences). They lack the function of evaluating text complexity (readability) based on such discourse parameters as syntactic simplicity, concreteness/abstractness, cohesion and coherence. The mathematical model of Russian text complexity and Automated Text Complexity Analyzer designed on its basis can be used worldwide to develop similar engines for other synthetic languages. As modern schoolchildren population is heterogeneous, the problem of linguistic appropriateness of texts used in teaching and learning is crucial. Therefore, educational materials developed with an appropriate academic and cognitive level (age) in mind are highly demanded.
Relevance and significance of the project in technologies
The project is aimed at the development of linguistic technologies to tailor individual rehabilitation and learning needs. Its outcomes comprise not only fundamentally new data on neurobiological aspects of speech perception, but also the development of a wide range of neurolinguistic and psychophysiological methods for studying standard and pathological speech perception conditions. Another significant outcome is improving EEG recording techniques and data analysis of patients with cochlear implants. Development of experimental tasks profiled for combined use of event-related EEG potentials methods (requiring a relatively large accumulation of similar incentives implementations) and induced changes in the EEG rhythmic activity (requiring a longer interstimulus interval), is expected to provide a unique technological solution.
The additional significance of the Project is attributed due to developing neurolinguistic and neurocognitive methods for speech disorders assessment which enable far more accurate diagnosis in a shorter period of time. The planned publications of the research outcomes in peer-reviewed journals, as well as extensive scientific collaboration of the project team are to ensure dissemination of the developed techniques and diagnostic tools. This will enable to gain and strengthen leadership in the research area. As the Project is also aimed at designing and developing Russian text analytical processing technology, its implementation in teaching and language assessment is to change a number of traditional practices of item-writing and test development. The consecutive implementation of the technologies developed at each stage of the project is to provide ongoing process of personalizing learning materials, as well as improvement of the existing system of language assessment in schools. The project is directly related to the global scientific tasks set to generate new technologies for expanding human potential (health promotion and improving learning). The estimated costs of the Project outcomes implementation comprise: in-service development programs in healthcare and education, hosting summer/ winter schools and master classes at Kazan Federal University and on the venues of the Project partner universities, writing and publishing ‘Best Practices’ Manuals. As the Project findings are to be publicly available, there are expected costs on maintaining a server to support databases, corpora and the Project web site.