Periodic Reporting for period 1 - DONUT (European Doctoral Network for Neural Prostheses and Brain Research)
Okres sprawozdawczy: 2024-01-01 do 2025-12-31
DONUT, the European Doctoral Network for Neural Prostheses and Brain Research (https://donut-project.eu/(odnośnik otworzy się w nowym oknie)) aims to provide a multidisciplinary, intersectoral training environment for talented young researchers. It also serves as a springboard for expanding EU partnerships in the rapidly evolving field of Brain-Computer Interface (BCI) technology and related scientific disciplines.
The objective of DONUT is to train and mentor 10 doctoral candidates (DCs) to bridge the gap between BCI research and real-world applications. The overarching ambition is to enhance the quality of life for individuals with paralysis, support motor and language rehabilitation after stroke, and contribute to the early detection of Alzheimer’s disease (AD).
Previous BCI research has demonstrated the potential of these systems to offer alternative communication and control pathways for paralyzed individuals by directly interpreting brain signals and bypassing damaged neural circuits. This holds promise for improving autonomy and well-being. However, widespread adoption of BCI technologies remains limited by several challenges: lengthy calibration sessions, inconsistent performance, the phenomenon of BCI illiteracy (where systems fail to work for all users), and a lack of affordable, high-quality, and user-friendly brain recording devices.
Beyond communication, BCIs are increasingly being integrated into rehabilitation programs for motor and language functions, though typically only in specialized centres due to the need for expert staff and resources. Additionally, researchers are exploring BCIs as a tool for identifying early-stage EEG-based biomarkers of AD pathology, potentially enabling screening of at-risk individuals even outside clinical environments.
To fulfil its mission, DONUT brings together the complementary expertise of 7 academic institutions and 8 associated partners across 8 EU countries, working collaboratively to tackle fundamental challenges in brain science and to develop cutting-edge BCI systems and applications.
The objective of DONUT is to train and mentor 10 doctoral candidates (DCs) to bridge the gap between BCI research and real-world applications. The overarching ambition is to enhance the quality of life for individuals with paralysis, support motor and language rehabilitation after stroke, and contribute to the early detection of Alzheimer’s disease (AD).
Previous BCI research has demonstrated the potential of these systems to offer alternative communication and control pathways for paralyzed individuals by directly interpreting brain signals and bypassing damaged neural circuits. This holds promise for improving autonomy and well-being. However, widespread adoption of BCI technologies remains limited by several challenges: lengthy calibration sessions, inconsistent performance, the phenomenon of BCI illiteracy (where systems fail to work for all users), and a lack of affordable, high-quality, and user-friendly brain recording devices.
Beyond communication, BCIs are increasingly being integrated into rehabilitation programs for motor and language functions, though typically only in specialized centres due to the need for expert staff and resources. Additionally, researchers are exploring BCIs as a tool for identifying early-stage EEG-based biomarkers of AD pathology, potentially enabling screening of at-risk individuals even outside clinical environments.
To fulfil its mission, DONUT brings together the complementary expertise of 7 academic institutions and 8 associated partners across 8 EU countries, working collaboratively to tackle fundamental challenges in brain science and to develop cutting-edge BCI systems and applications.
The ongoing joint project has made significant strides in advancing BCI technologies, with a focus on enhancing user interaction and performance. The research team is dedicated to improving the usability and effectiveness of various BCI systems, particularly regarding visual feedback mechanisms and intuitive control interfaces, including visual feedback mechanisms for cVEP-BCI systems to enhance user interaction and developing an intuitive object-control method for smart-home environments using cVEP flickering technology. Research is also being conducted on designing 3D stimuli for virtual reality applications utilizing the SSMVEP stimulus protocol, which aims to improve user engagement. A significant achievement includes the development of a VEP transaction approval system that integrates real-time user feedback, achieving high accuracy rates. A key objective of the project is to develop a collaborative BCI framework for motor imagery rehabilitation, which incorporates real-time cognitive state monitoring to enhance therapeutic interactions and facilitate the control of lower-limb exoskeletons.
Additional goals include designing a custom integrated circuit (IC) in 65 nm for concurrent measurements of EEG and electrode-skin impedance, conducting systematic reviews of hybrid EEG–EMG biosensing, deep learning, and VR/Serious Games, and developing a collaborative BCI system for motor imagery within a virtual reality rehabilitation environment. Knowledge exchange and collaboration are prioritized through seminars and guest lectures, ensuring effective dissemination of findings within the academic community.
Research activities also involve extensive data collection and analysis, leading to valuable insights that inform the design of future BCI applications. The team actively engages in various dissemination activities, including workshops and presentations, to share findings and methodologies with a broader audience.
In the field of human emotion recognition using a hybrid BCI, we conducted a systematic mapping study of recent work at the intersection of hybrid EEG–EMG biosensing, deep learning, and VR/Serious Games, screening 2,128 records and including 28 eligible studies. The analysis revealed that no existing work combines EEG+EMG with deep learning in VR, with evidence instead clustering in partial combinations. Based on synthesized pipelines and evaluation practices, we identified key gaps and proposed a roadmap, including VR-native EEG–EMG datasets and subject-independent benchmarking protocols.
Regarding AD biomarker development, we examined the relationship between simultaneously recorded scalp EEG and hippocampal activity using depth electrodes in patients with temporal lobe epilepsy (TLE), a population relevant to Alzheimer's disease (AD) research due to overlapping pathological features. In addition, we successfully exploited the potential of scalp EEG to identify individuals with amyloid and tau positivity, the two primary biomarkers of AD alongside neurodegeneration.
Additional goals include designing a custom integrated circuit (IC) in 65 nm for concurrent measurements of EEG and electrode-skin impedance, conducting systematic reviews of hybrid EEG–EMG biosensing, deep learning, and VR/Serious Games, and developing a collaborative BCI system for motor imagery within a virtual reality rehabilitation environment. Knowledge exchange and collaboration are prioritized through seminars and guest lectures, ensuring effective dissemination of findings within the academic community.
Research activities also involve extensive data collection and analysis, leading to valuable insights that inform the design of future BCI applications. The team actively engages in various dissemination activities, including workshops and presentations, to share findings and methodologies with a broader audience.
In the field of human emotion recognition using a hybrid BCI, we conducted a systematic mapping study of recent work at the intersection of hybrid EEG–EMG biosensing, deep learning, and VR/Serious Games, screening 2,128 records and including 28 eligible studies. The analysis revealed that no existing work combines EEG+EMG with deep learning in VR, with evidence instead clustering in partial combinations. Based on synthesized pipelines and evaluation practices, we identified key gaps and proposed a roadmap, including VR-native EEG–EMG datasets and subject-independent benchmarking protocols.
Regarding AD biomarker development, we examined the relationship between simultaneously recorded scalp EEG and hippocampal activity using depth electrodes in patients with temporal lobe epilepsy (TLE), a population relevant to Alzheimer's disease (AD) research due to overlapping pathological features. In addition, we successfully exploited the potential of scalp EEG to identify individuals with amyloid and tau positivity, the two primary biomarkers of AD alongside neurodegeneration.
This project has pushed the boundaries of current BCI technology by introducing innovative frameworks that integrate cognitive state assessment into the rehabilitation process. Unlike traditional BCI systems, which often operate under fixed parameters, the methodologies developed in this project allow for dynamic adjustments based on real-time user feedback. This adaptability enhances the therapeutic experience and supports more personalized interventions.
Furthermore, the project has pioneered the use of single-stimulus paradigms in EEG-based authentication systems, demonstrating their feasibility and user-friendliness compared to conventional multi-target approaches. By reducing the complexity of user interactions and minimizing the need for extensive training, the project has made significant strides toward more accessible BCI applications.
The integration of virtual environments with BCI systems marks another significant achievement of the project. By enabling collaborative neurorehabilitation scenarios, this project sets a new standard for future developments in adaptive BCI technologies, with the potential to transform therapeutic practices and improve user outcomes.
We have demonstrated that scalp EEG recorded in response to a set of dedicated visual stimulation paradigms can distinguish cognitively unimpaired individuals who are amyloid-negative from those who are amyloid-positive, and can also detect tau positivity. These findings point toward the development of a cost-effective preclinical screening tool.
Furthermore, the project has pioneered the use of single-stimulus paradigms in EEG-based authentication systems, demonstrating their feasibility and user-friendliness compared to conventional multi-target approaches. By reducing the complexity of user interactions and minimizing the need for extensive training, the project has made significant strides toward more accessible BCI applications.
The integration of virtual environments with BCI systems marks another significant achievement of the project. By enabling collaborative neurorehabilitation scenarios, this project sets a new standard for future developments in adaptive BCI technologies, with the potential to transform therapeutic practices and improve user outcomes.
We have demonstrated that scalp EEG recorded in response to a set of dedicated visual stimulation paradigms can distinguish cognitively unimpaired individuals who are amyloid-negative from those who are amyloid-positive, and can also detect tau positivity. These findings point toward the development of a cost-effective preclinical screening tool.