Dynamic classification applied to brain recognition of context exploration

SEBASTIÁN MARTÍNEZ; VIRGINIA MAZZONE; DEMIAN GARCÍA VIOLINI; AZUL SILVA; BELLUSCIO, MARIANO; PIRIZ, JOAQUIN; RICARDO S. SÁNCHEZ PEÑA

Congreso; Congreso Argentino de Control Automático; 2020

Asociación Argentina de Control Automático

A methodology for the characterization of cognitivestates using local field potential, stemmed from electrical signalsin the hippocampus, is presented in this study. To this end, theproposed methodology employs classical theoretical results fordynamical systems, particularly based on the Dynamic ModeDecomposition method, in conjunction with binary clusteringroutines based on supervised learning. All the analysis and resultsshown here consider real data measured in the hippocampus,particularly in the CA1 region of rats performing explorationtasks. Thus, using the measured local field potential, this study indicatesevidence that the context exploration can be clustered intofamiliar or unfamiliar environments, depending on the particularcontext and previous training of each rat. All the results shownin this paper, are presented using different statistical measuresto assess the results and their confidence range.