Determination of the Cognitive Model: Compressively Sensed Ground Truth of Cerebral Ischemia to Care

摘要

Reported research concerns optimized methods of computerized clinical decision support on the example of stroke care management. New paradigm of compressive cognition is presented and discussed including implementation of the proposed empirical model designed to improve ischemia description and aid reperfusion therapy. The concept of semantic compressed sensing was developed to analyze clinically conditioned consensus of ground truth formulated basing on semantic descriptors of objectified expert ratings, interview data analysis, monitoring of vital signs, lab measurements, the results of physical examinations and imaging studies. The designated sparse model allows determining the interrelationship between subjective interpretations of physicians completing comprehensive picture of pathology in emergency conditions. According to the experiments carried out, the obtained effectiveness of stroke diagnosis and prediction of the effects of applied therapy is very high. The potential benefit is not only important for the patient and the physician, but also for the whole society, by significantly reducing the socio-economic costs of caring for a stroke patient.