A realistic assistant for patients with paralysis based on a multimodal and explainable Brain-Computer Interface, developed using zero-calibration, paradigm-independence and stimulus minimization technologies

​Grant Funding for Young Researchers 2025-2027

The lifetime of a paraplegic patient such as stroke, ALS, spinal cord injuries is very difficult due to inflicted by the illness restrictions on movement and communication. Fifteen million people suffer stroke worldwide each year, and all patients are highly limited in the social activities, entertainment, and communication with their family, friends and caregivers[1-3]. They basically require a long-term care that leads to immeasurable economic and social expenses to support their minimum quality of life.
Brain-Computer interfaces (BCIs) use brain activity to control external devices and facilitating paralyzed patients to interact with external environments such as an exoskeleton, speller, robot arm, AR/VR game, brain monitoring and etc [4-5]. BCI represents a promising and sole strategy to establish communication with paralyzed ALS patients as it does not need muscle engagement for its use. Distinct techniques have been explored to assess brain neurophysiology to control BCI for patient's communication, especially electroencephalography (EEG) and more recently near-infrared spectroscopy (NIRS) that are non-invasive, low-risk, and easy of use. In addition to assisted communication, BCI is also being extensively studied for the health care monitoring system by measuring all the life signals (heart rate, respiration, brain/muscle activities, etc.) in daily life [6-8]. Presentation of the stimuli during BCI communication incurs an eye fatigue and stress for the patients. Hence, in this project we aim to achieve a goal of constructing state-of-the-art BCI system that relies on solely diminutive stimuli with a very small size, which will have a positive effect on usability of the developed interfaces [9-10]. BCI can also be used to recover the brain functions by intensive rehabilitation tasks by motivating the patient's intention to move the paretic limb with the contingent sensory feedback of the paretic limb movement guided by assistive devices[11-12, 15,18]. The aim of the project is to develop real-life brain-computer interface assistive tools to help patients suffering from neurodegenerative and movement disorders. However, current BCI techniques have diverse challenging issues that must be overcome to be used in a real-world environment, and thereby we aim to solve the following problems that limit the current BCI adoption in a clinical setting: 1) zero-calibration/training 2) stimulus-diminutive BCI paradigms 3) data/paradigm-independent BCI tools 4) explanation/interpretability of classification results and 5) multi-modal BCI open-dataset.