Brain-Computer Interfaces (BCIs) are revolutionizing assistive technologies, enabling direct communication between the brain and external devices. Among their many applications, one of the most impactful is motor imagery, where individuals imagine specific movements to send a command. For example, imagining the act of grasping an object can enable users to control a robotic prosthetic hand, allowing them to perform tasks such as picking up a cup or pressing a button. These advancements are transformative, offering individuals with motor disabilities innovative ways to perform tasks previously thought impossible.

A unique platform that highlights the potential of assistive technologies like BCIs is Cybathlon—an international competition that unites researchers, engineers, and individuals with physical disabilities. The event showcases cutting-edge solutions in real-world scenarios across disciplines such as robotic prosthetics and BCIs. Beyond competition, Cybathlon fosters collaboration and innovation, inspiring advancements that push the boundaries of assistive technology.

Inspired by this mission, Team MIRAGE91 was founded in 2014 with the goal of advancing BCI technology and competing on the Cybathlon stage. MIRAGE91 stands for Mental Imagery RAcing Graz Established ‘91, which expresses the beginning of the BCI research at Graz University of Technology back in 1991. Composed of a diverse group of researchers, students, and a dedicated pilot who plays a pivotal role in training and testing, MIRAGE91 has made significant progress in this dynamic field. Since its inception, the team has focused on developing robust, real-time solutions for BCI applications. MIRAGE91 has competed in Cybathlon events since 2016, joining a global cohort of BCI teams from other countries like Germany, Italy, Hungary, Thailand, England, and the USA. In 2024, the team returned with renewed determination and expertise.

Cybathlon Challenges and Tasks

Unlike the previous Cybathlon competitions in 2016 and 2020, where teams developed a four-class BCI system using motor imagery to send discrete commands for controlling an avatar in a navigation racing competition—such as turning left, turning right, resting (no command), and turning on the lights—the 2024 competition introduced a more complex challenge. This time, teams were required to develop a system capable of generating four continuous control signals and two binary discrete signals to interact with the environment.

As a result, the competition tasks became more intricate, shifting from simple navigation to real-life-inspired challenges. Pilots were now required to complete 10 different tasks, each designed to demonstrate the practical applications of BCIs. These included unlocking doors with a virtual key, guiding a virtual glass to catch ice cubes, and controlling a wheelchair through a virtual environment filled with obstacles like flowerpots, chairs, and tables.

While preparing for Cybathlon 2024, Team MIRAGE91 encountered various technical and logistical challenges that required careful planning, problem-solving, and teamwork. One of the most unique experiences for our team was working with a pilot who had a spinal cord injury. Our pilot, Antonio, had sustained his injury in a swimming accident two years ago, making this a completely new experience for our young team members, who had previously only worked with healthy participants. This introduced some complexities in conducting EEG measurements and time planning, as working with individuals with motor disabilities requires additional considerations compared to standard lab setups. However, Antonio’s motivation and enthusiasm for the competition helped minimize these difficulties.

Additionally, the LiveAmp device from Brain Products GmbH played a crucial role in simplifying the measurement process. Its wireless design significantly reduced setup time and effort, allowing us to focus on algorithm development rather than dealing with cumbersome equipment. The high signal quality provided by LiveAmp enabled us to conduct extensive calibration sessions, fine-tuning the system to accurately capture Antonio’s brain activity patterns and improving overall system adaptability.

One major technical challenge was implementing a real-time pipeline for processing EEG signals during the competition. In the beginning, we designed various motor imagery tasks to collect sufficient data for calibrating a model for the online phase. This also allowed us to determine the optimal motor imagery classes that worked best for our pilot.

After multiple measurement sessions, we developed our online system to provide instantaneous feedback on the screen and seamlessly connect to the game for task execution. However, transitioning from an offline experiment to an online system introduced additional challenges, as a drop in accuracy is commonly observed in real-time applications.

Additionally, the low signal-to-noise ratio (SNR) of EEG signals presented further challenges. Extracting meaningful information from noisy signals was particularly difficult when recording across different sessions. Despite our efforts to precisely position the EEG cap each time, slight variations in electrode placement introduced additional variability. Moreover, due to the non-stationary nature of EEG signals, brain activity patterns naturally fluctuate over time, leading to inconsistencies when recording on different days. To mitigate these limitations, we optimized our signal processing algorithms, ensuring reliable performance in an online setting. This process required iterative coding and continuous refinement of the system to minimize differences between offline and online performance. Additionally, to prevent performance degradation in online sessions, we conducted a short calibration session to fine-tune our trained models, ensuring better adaptability in real-time applications.

© MIRAGE91

Team dynamics also played a crucial role in our progress. With a diverse group of phd, master and bachelor students with different backgrounds, we benefited from a wide range of expertise and perspectives. Encouraging student involvement not only brought fresh ideas but also fostered an innovative and collaborative learning environment, where different skills and viewpoints contributed to refining our approach.

Ultimately, our system performed as intended during Cybathlon 2024, allowing us to successfully complete the tasks. The competition itself provided valuable insights into real-world applications of BCIs, as well as opportunities to learn from other teams and improve our methods further. While there are always areas for improvement, our experience at Cybathlon demonstrated that careful preparation, adaptability, and teamwork can help overcome complex challenges in developing assistive technologies.

Collaborative Efforts and Learning Opportunities

A cornerstone of MIRAGE91’s philosophy has been fostering collaboration. The team actively engaged students from diverse academic backgrounds, encouraging them to contribute their unique expertise to the project. This multidisciplinary approach not only strengthened the team’s capabilities but also created a rich learning environment that nurtured innovation and teamwork.

The Cybathlon Experience and Competition Day

All these efforts culminated in MIRAGE91’s successful participation in Cybathlon 2024. The competition was an unforgettable experience, combining intense preparation with the thrill of showcasing our innovations on a global stage. The atmosphere was electric, with teams from around the world demonstrating the incredible potential of assistive technologies.

Beyond the competition itself, the event provided a unique opportunity to engage with researchers from other universities, exchanging knowledge and gaining insights into different approaches to BCI development. These discussions were invaluable, offering fresh perspectives and ideas for future improvements.

The real event was both challenging and inspiring. Unlike the controlled lab environment, the competition introduced real-world factors such as background noise, bright lights, a live audience, and TV broadcasts—elements that made system performance and adaptability even more critical. This experience reinforced the importance of designing BCIs that can function reliably outside of structured research settings.

© MIRAGE91

Looking Ahead

Reflecting on our journey, Team MIRAGE91 is proud of the progress we’ve made and the impact of our work in the field of BCI technology. The Cybathlon experience solidified our belief in the transformative power of BCIs to enhance independence and quality of life for individuals with motor disabilities. As we look toward the future, we remain committed to advancing this technology, inspired by the incredible potential it holds to empower lives around the world. We are also looking forward to participating in the next Cybathlon event.