作者: Ali Yousefi , Angelique C. Paulk , Thilo Deckersbach , Darin D. Dougherty , Emad N. Eskandar
DOI: 10.1109/EMBC.2015.7320205
关键词:
摘要: Behavioral tests are widely used to quantify features of cognitive processing. For a large class behavioral signals, the observed variables non-Gaussian and dynamic; classical estimation algorithms ill-suited modeling such data. In this research, we propose mathematical framework predict state variable related which best modeled using Gamma distribution. The proposed algorithm combines Smoother EM in prediction process. is applied reaction time recorded from subjects performing Multi-Source Interference Task (MSIT) dynamically their flexibility through course experiment.