Decision Tree to Analyses EEG Signal: A Case Study Using Spatial Activities

Abstract

The Electroencephalogram (EEG) is based on records of brain electrical activity measured through the scalp, and it is commonly used for studies and neurological examinations. It has now been used in conjunction with BMI (Brain Machine Interface) Systems, which allow the communication of an individual and external equipment, such as a computer, through brain stimulation, without any muscle movement. These tools typically present data in the form of graphs or topographic maps for monitoring or analysis of brain activities. In this work, a methodology of EEG analysis is proposed through data mining, which tries to extract knowledge from a large database. This paper presents a case study using EEG signals from visually impaired and sighted individuals during the execution of an activity that stimulated spatial ability. In order to verify the hypothesis that sighted use the sense of sight, even with blindfold, and visually impaired use the sense of touch to identify spatial geometric objects. An experiment was made and through the data of the recorded brain signals, decision tree techniques were applied to understand the main areas involved in the brain activities during the execution of the recognition task. The results suggest that the hypothesis is true confirming that visuall and sighted subjects activate different brain areas.