Development of a graphical interface for the non-lineal analysis of time series: electroencephalograms

  • Alfonso Medel Rojas Centro Interdisciplinario de Postgrados, Ingeniería Bio-médica, UPAEP, Puebla
  • Alfonso Medel Rojas Centro Interdisciplinario de Postgrados, Ingeniería Bio-médica, UPAEP, Puebla
  • Guillermo De la Rosa Orea Centro Interdisciplinario de Postgrados, Ingeniería Bio-médica, UPAEP, Puebla
  • Juan López Oglesby Centro Interdisciplinario de Postgrados, Ingeniería Bio-médica, UPAEP, Puebla
  • Tomás Morales Acoltzi Centro de Ciencias de la Atmósfera, UNAM, Circuito Exterior, Ciudad Universitaria, D.F.
  • Manuel González Pérez Centro Interdisciplinario de Postgrados, Ingeniería Bio-médica, UPAEP, Puebla

Abstract

 

Objective: We developed a graphical interface for the group of routines in the package executable TISEAN (Time Series Analysis) which allows the study of the behavior of time series (TS) associated to processes of nature. Methodology: Using the MATLAB programming language, we developed a hybrid that allows us to graphically display the results of TISEAN executable routines. The database (DB) consists of 10 EEG signals, people with a genetic predisposition to alcoholism, 5-alcoholics (EEG-A) and 5 controls (EEG-C) and sampled with 64 electrodes, 256Hz.1 Results: We performed numerical experiments with theoretical TS, world famous for testing the proposed methodology. In applying it, we found that both types of ST have a nonlinear behavior. Estimates of invariants of the EEG ST allow to observe differences in the forms and values of the dimensions of the attractors for groups and EEG EEGA-C. Conclusion: A computer program was developed as an hybrid interface between two packets accepted by the scientific community, showing the feasibility of being used by medical to enable them to obtain precursors from the EEG signals.

Published
2018-02-19
Section
Research Articles