Nome do Pesquisador: Birajara Soares Machado (Instituto do Cérebro - Centro de Pesquisa Experimental - Centro de Pesquisas do Hospital Albert Einstein)
Agência de fomento: Sem financiamento
Vigência: 08/2018 à 08/2020
Pesquisadores Docentes da UFABC:
Andre Ricardo Oliveira da Fonseca
Resumo: Detecting signal transitions that relate to behavioral changes is a fundamental step on the study of brain activity. Nevertheless, currently there is no standard automatic staging technique and the process is often manually performed. In this project, we propose a method that employs the complexity of time series and cluster-weighted modeling for unsupervised data segmentation. The technique has several advantages: it is suited for short-length signals; it is well adapted to detect changes of rhythmic oscillations and has no modeling restrictive conditions. It provides an automatic and efficient staging process for simulated and biological data sets.