Análise de sinais cerebrais uma abordagem de informação simbólica para distinguir diferentes processos cognitivos

Abstract

This dissertation addresses the analysis of brain signals and their application in different experimental contexts. We introduce the theoretical foundations of signals and time series, as well as techniques for capturing and storing these signals. Next, we discuss the contribution of physicists to the study of neuroscience, focusing on the statistical analysis of discrete time series obtained from brain signals during the performance of different cognitive tasks. We detail the application of Shannon entropy and MPR statistical complexity as quantifiers to characterize brain signals, using Bandt & Pompe's symbolic methodology. We also describe the implementation of Bandt & Pompe's symbolic methodology using the Python programming language. Finally, we explored two distinct experiments: one involving visual cognitive tasks performed by humans and another related to motor tasks performed by monkeys. We show that these information theory tools are useful for characterizing and distinguishing different cognitive processes, as well as the most relevant time scales for processing information related to different tasks.