Post Stroke Rehabilitation with Brain-Computer Interfaces (BCI)

General description of the project

Voluntary muscle movement originates in the brain, where different regions exchange information in a coordinated manner to plan and execute motor action. From this central processing, the neural signal descends through the spinal cord and is transmitted through the efferent pathways of the peripheral nervous system to the muscle. Subsequently, the sensory information generated during the execution of the movement travels the reverse path through the afferent pathways, returning to the central nervous system and allowing the generation of new motor responses in a continuous feedback process.

In contrast, current knowledge about the neural processing of motor signals remains limited. For this reason, in recent years, experimental approaches have been developed and consolidated to study the relationship between movement and its representation in the brain by measuring cortico-kinetic coherence. This analysis is carried out using non-invasive techniques, including the use of accelerometers to characterise the frequency, intensity and duration of movement, as well as the recording of brain activity using electroencephalography (EEG) and magnetoencephalography (MEG), techniques that measure, respectively, the electrical activity and magnetic fields generated by the brain during motor execution.

However, all these techniques have significant limitations, which form the starting point for this project. These include the high complexity of the recorded signals and the presence of noise from external physiological sources, such as blinking or heartbeats, as well as brain activity not directly related to the motor process of interest. These factors hinder the interpretation of the data and reduce the reliability of the analyses, making it necessary to develop methods to improve the quality and usefulness of the information obtained.

Objectives

The aim of this project is to design, develop and apply advanced Artificial Intelligence models capable of jointly analysing electroencephalography (EEG) signals and muscle activity in order to characterise the relationship between movement and brain activity.

To this end, brain and muscle activity will be recorded simultaneously while participants perform motor exercises, allowing the models developed to identify the execution of the exercises and distinguish between different types of motor activity. This approach will provide a deeper understanding of the neurophysiological mechanisms underlying movement control and execution, as well as the dynamic interaction between the central nervous system and the muscular system.

These models will be applied specifically to the study of patients who have suffered a stroke during their rehabilitation processes, through the analysis of brain activity recorded while performing therapeutic exercises. Based on the extraction of relevant characteristics from neurophysiological signals, the project aims to evaluate the functional evolution of patients throughout therapy, identify which exercises and intervention modalities generate the most significant changes in brain activity and contribute most effectively to motor recovery, and improve understanding of the mechanisms of brain reorganisation after injury.

Ultimately, the goal is to lay the groundwork for the design of more specific and personalised rehabilitation protocols, tailored to the type of brain damage and the individual response of each patient.

Scientific production

J. P. García, M. Zivanovic and M. Gómez, ‘Application of Computer Vision and Explainability in EEG Classification for Motor Imagery Therapy after Stroke’, poster presented at the Annual Computational Neuroscience Mitin, Florence, Italy, Jul 5–9, 2025.

Funding

This project forms part of the collaboration agreement signed with Fundación Caja Navarra as part of its programme of grants for data and information fusion.

More projects