Use-case 1: EEG Signal analysis

Introduction

This use-case is a simple yet concrete example of real-time EEG analysis usually performed with OpenViBE. The scenario covers the use of oscilloscope, bitmaps, bars and topographic views to display signal, spectrum, and band power.

The scenario

The signal used is a motor imagery session, where the participant performed right and left hand motor imagery trials. For more details, please refer to the official documentation of the OpenViBE motor-imagery bci scenarios, provided with the official release of the software. We chose these data for demonstration purpose only as it is a file provided with the official release of openvibe, and should be available for you anyway.

Signal filtering

We first remove artifacts using temporal filters, especially the common 50Hz noise coming from the electrical installation. The EEG amplifier used for the record we read here is a Mindmedia NeXuS 32b, with one reference channel put on Nz (nose). The Reference Channel box applies this spatial filter to further remove noises.

We then use a Continuous Oscilloscope to display the filtered signal.

UseCase1_1.png
Denoising the signal before display

Spectral analysis

A first pipeline computes two surface Laplacian filters around C3 and C4, the center of the two motor cortices. We then compute the spectrum using FFT, up to 32 Hz, and display it using Instant Bars (spectrum levels) and Stacked Bitmap (Horizontal) (time-frequency map).

UseCase1_2.png
Spectral analysis over filtered data

Topographic display

We compute in a parallel pipeline the alpha band power, averaged over several epochs, and visualize it over the scalp through 2D Topography and 3D Topography.

UseCase1_3.png
Topographic display of the alpha band power over the scalp

Result

Here is the online visualization when we play this scenario on the provided data.

UseCase1_6.png
Signal display
UseCase1_4.png
Spectrum visualization
UseCase1_5.png
2D and 3D Topographies