SBIR Phase I: Intelligent tutoring system with EEG-based instructional strategy optimization

This SBIR Phase I project's broader/commercial impact is a proposed tutoring system with the potential to meet the need for truly scalable personal tutoring by consistently equaling the performance of an above-average tutor. One-on-one instruction by qualified personal tutors is the optimum form of teaching. Well-trained personal tutors, however, are a limited resource, particularly in disadvantaged areas and for science, technology, engineering and mathematics, and are outside the budget of the vast majority of school districts and parents. However, unlike human tutors, the system proposed here can be available at any time outside school; it can be stopped and resumed at will, can be utilized at home, and is blind to economic, gender and other social cues. The system will be offered to the $34 billion after-school education, test-prep and eLearning market, with an initial entry point in Advanced Placement (AP) exam preparation services to high school students. Subjects suitable for the learning paradigm of the proposed system account for 1 million annual exams, resulting in to an initial Total Available Market of $400 million annually. This beachhead market will be expanded to include general year-round tutoring for both high school and higher education. The project will develop a computerized tutoring platform that adapts its teaching strategy to students in real-time by monitoring their brain activity. This innovation will meet shortfalls in current computer-based adaptive training technologies, which rely on generalized learner state models and inferences of student engagement and cognitive workload, and thereby lack some of the key abilities of human tutors. This platform will combine the strengths of an Intelligent Tutoring System modeled after expert human tutor behavior and has an innovative scaffolding mode with newly-developed wearable brain monitoring technology. The Phase I objectives are to evaluate EEG-based gauges of cognitive workload and engagement on students executing learning tasks in a classroom environment, evaluate the gauges in response to learning tasks of defined difficulty, and define methods of using the gauges to modify instructional content. A trial will be carried out on students at a partner high school, the anticipated outcome of which is the measurement of cognitive gauges with high classification accuracy. A structural framework of how the cognitive state gauges will interface with the tutoring system and optimize its instructional strategy will be developed, and user requirements for the prototype system will be defined.