CogniX: EEG & Eye-Tracking BCI Research Project
This page presents CogniX as an example application developed and deployed in the Interactive Technologies Lab using EEG and eye-tracking equipment. The project was supported by HFRI funding and demonstrates how existing infrastructure equipment can be combined to support a multimodal research application.
Project Context
CogniX was developed as part of a research project carried out in the lab under funding from the Hellenic Foundation for Research and Innovation (HFRI). The project explored the use of EEG and eye-tracking within a single interactive application, supported by a visual, node-based software environment.
In this documentation, CogniX is used as an example showcase of:
- multimodal data acquisition,
- synchronized data streaming,
- real-time and offline processing pipelines,
- integration with external applications.
Multimodal Setup
The application uses two primary signal modalities:
EEG
EEG is used to capture motor imagery activity, which is later classified and mapped to discrete actions in the application.
Relevant device documentation:
Eye-Tracking
Eye-tracking is used for gaze-based target selection within the application interface.
Relevant device documentation:
Data Synchronization with LSL
All data streams in CogniX are synchronized using Lab Streaming Layer (LSL). EEG signals, eye-gaze data, and task markers are streamed in parallel and aligned in time.
This approach allows:
- device-agnostic data acquisition,
- synchronized recording across modalities,
- reuse of the same setup for offline analysis and real-time applications.
Software Architecture Overview
CogniX follows a node-based, flow-oriented architecture that supports both offline and real-time processing.
At a high level, the system is structured as follows:
-
CogniX Core
Handles node execution, data flow, graph evaluation, and communication. -
CogniX Editor
Provides a graphical interface for building and running processing graphs. -
External Application
A Unity-based application receives processed signals and provides user feedback.
The system separates the workflow into three distinct phases:
- Data acquisition
- Model calibration (offline)
- Real-time deployment
This separation allows the same infrastructure setup to be reused across different stages of an experiment.
Application Example: EEG & Eye-Tracking BCI Puzzle Game
As a concrete application, CogniX was used to support a simple puzzle game that combines: - eye gaze for selecting game elements, - EEG motor imagery for triggering actions.
The game serves as a demonstration of how: - multimodal signals can be combined, - offline-trained models can be deployed in real time, - the infrastructure can support interactive applications beyond data collection.
The video below shows the CogniX system in use during the EEG and eye-tracking puzzle game.
