Spring School 2024: Motor imagery, P300, SSVEP, ECoG data sets @gtecmedicalengineering

Uploaded April 2025 | Updated September 2026, 1 day ago

The talk presented various datasets related to motor imagery, P300, SSVEP, and ECoG (electrocorticography) collected by g.tec medical engineering. These datasets serve as valuable resources for researchers and developers working in the field of brain-computer interface (BCI) technology. They offer opportunities for studying brain signals associated with different cognitive tasks and assist in the development and evaluation of BCI algorithms and applications.

The Spring School offers an exclusive focus on cutting-edge neurotechnology. Immerse yourself in 140 hours of dynamic education spread across 10 enriching days, featuring insights from 85 distinguished speakers hailing from 86 esteemed institutions. Engage in 44 innovative Hackathon projects facilitated by 15 passionate hosts, all united in their mission to pioneer advancements in neurotechnology.

Learn more about Unicorn Hybrid Black: https://www.gtec.at/product/unicorn-hybrid-black/
Spring School 2024: https://www.gtec.at/spring-school-2024/

Spring School 2024: Motor imagery, P300, SSVEP, ECoG data sets

Before/After recoveriX - stroke - WDF WPF - ST4

Before/After recoveriX - Parkinsons - TUG - ST43

Cognitator: Brain-Controlled Virtual Reality Game with Unicorn Hybrid Black

Game Control and Interaction with g.tec: From 2D and 3D to VR and AR

OSCAR LIVE Animal EEG: Freely Moving Alpaca with Unicorn BCI Core-8

Transmitting EEG data from the Unicorn BCI-Core-8 to the Samsung Galaxy

Before/After recoveriX - stroke - 9HPT - ST8

$Unlockers: High-Accuracy P300 Classification for Brain-Computer Interface Communication BR41N.IO Hackathon during IEEE SMC 2025 - Second place winner - Data Analysis How can Brain-Computer Interfaces become more practical for locked-in patients when only limited training data is available? At the BR41N.IO Hackathon, Team Unlockers developed an advanced machine learning pipeline designed to improve P300 classification accuracy while reducing the amount of data required from users. Built within just 24 hours, the project focused on one of the most important challenges in BCI research and assistive communication: achieving reliable P300 detection without requiring hundreds of repetitive trials. This is especially important for locked-in patients, where lengthy calibration sessions can be physically demanding and limit practical use. The team developed a complete EEG signal processing and machine learning workflow that combined artifact removal, data augmentation, feature engineering, and advanced classification techniques. Multiple Linear Discriminant Analysis (LDA) approaches were evaluated, including Fishers LDA, Regularized LDA, and Stepwise Regularized LDA, alongside several EEG preprocessing strategies designed to maximize signal quality while preserving valuable training data. A key innovation of the project was the use of targeted EEG data augmentation techniques, including peak shifting, amplitude scaling, and temporal transformations to generate additional training examples from limited datasets. The team also explored spiking neural networks for synthetic EEG generation, demonstrating how artificial intelligence and machine learning can improve Brain-Computer Interface performance when training data is scarce. The final classification pipeline achieved an impressive 94% accuracy for P300 versus non-P300 detection, outperforming many traditional baseline approaches while requiring only a fraction of the training data often used in BCI systems. These results highlight the potential for more efficient calibration procedures, faster user onboarding, and improved patient experience in assistive communication applications. This project demonstrates important concepts in ERP research, P300 spellers, cognitive neuroscience, assistive communication, machine learning, artificial intelligence, EEG signal processing, and Brain-Computer Interface development. It also shows how rapidly interdisciplinary teams can build and validate advanced neurotechnology solutions when they can focus on algorithms and innovation rather than hardware integration and signal acquisition challenges. BR41N.IO Hackathon projects repeatedly demonstrate that researchers, students, startups, and developers can build high-performance Brain-Computer Interfaces, ERP applications, machine learning pipelines, assistive technologies, and real-time EEG solutions within just 24 hours. The Unicorn Hybrid Black platform enables teams to focus on classification, signal analysis, and application development instead of spending valuable time on EEG hardware setup, artifact correction workflows, or low-level data acquisition. More about Unicorn Hybrid Black: https://www.gtec.at/product/unicorn-hybrid-black-bci-platform More about BR41N.IO: https://www.gtec.at/hackathon/$

Before/After recoveriX - stroke - BBT - ST10

Before/After recoveriX - Spinal Cord Injury - TUG - ST44