Binary classification algorithm for non-invasive brain-computer interface systems based on electroencephalography

• Grant number: 24/17737-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: July 01, 2025
• End date: June 30, 2026
• Field of knowledge: Engineering - Biomedical Engineering - Bioengineering
• Principal Investigator: Leonardo Andre Ambrosio
• Grantee: Carolina Elias de Almeida Américo
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

The main purpose of this Scientific Initiation (IC) project is to study machine learning-based binary classification methods and to develop algorithms in Python, Matlab or Julia for such classifications, with validation tests applied directly to brain signals acquired through a non-invasive brain-computer interface (BCI) based on electroencephalography (EEG) signal acquisition and motor imagery. Among the systems to which the classification method and, consequently, the referred IC are of interest, two stand out as the group´s main focus: the first being the control of a 3D rotating LED matrix display, and the second being the determination of routes and paths for a smart and collaborative wheelchair within a home automation environment, both through motor imagery. In order to achieve these goals, activities are proposed involving basic studies on human brain anatomy and physiology; invasive and non-invasive brain electrical signal acquisition methods; electroencephalography and electrode placement on the scalp; key techniques on filtering and noise reduction; and finally, binary classification with machine learning techniques (LDA and RDA), with the development and testing of algorithms as well as validation tests using electrode kits and OpenBCI's free softwares all available at our laboratories. As a result, in addition to the necessary formation of human resources qualified for high-level research at the Electric and Computer Engineering Department (SEL) of the University of São Paulo (USP), in the areas of assistive technologies for mobility and inclusion, as well as virtual and/or augmented reality, it is expected that, by the end of this work, the scholarship recipient - currently in the 4th year of her undergraduate degree in Computer Engineering - and the research group to which she is affiliated will have developed a classification algorithm that will assign binary class labels to vectors of characteristics extracted from EEG acquired brain signals, originating from motor imagery neurocognitive processes. In the future, this algorithm will be extended to multi-class classification, allowing additional degrees of freedom in controlling 3D displays or wheelchairs in home automation environments.