Reconstructing kinematics trajectories during walking from EEG signals.

Lower-limb robotic exoskeletons have emerged as aids for over-ground, bipedal ambulation for individuals with motor limitations. The usability and clinical relevance of these robotics systems could be further enhanced by brain-machine interfaces (BMIs). Different approaches have been explored in the last decade to interact with robotics exoskeletons by means of BMIs based on EEG. One of the approaches explored is based on the decoding of kinematics trajectories during walking from EEG. Although walking is automatically based on reflexes governed at the spinal level, there are evidences that suggest that the motor cortex is particularly active during specific phases of the gait cycle. In addition, recent studies claim that EEG signals are directly related to the value of joint angles involved in human gait. In this regard, our team has verified that it is possible to get a relation between lower- limb angles and EEG signals by using linear regression models. However, despite current efforts for reconstructing kinematics trajectories from EEG signals, more research is still needed to improve the performance of current decoding algorithms. Furthermore, there is a huge lack of EEG data available for researchers to develop, test and compare their algorithms.

The main goal of this proposal is to register the EEG signals and the kinematics trajectories of lower-limbs of a high number of subjects during walking to: (1) integrate all these data into the EUROBENCH database; (2) improve our algorithm for reconstructing kinematics trajectories from EEG by using the recorded information; and (3) include in the EUROBENCH database the results of our decoding algorithm. This information will be a powerful resource for future researchers to develop, test and compare their algorithms.


  • PI: Jose M. Azorin
  • Research Center: Universidad Miguel Hernández de Elche (UMH)
  • Duration: April 1st, 2021 – December 21th, 2021
  • Funding: This project has received funding from the European Union’s Horizon 2020 research and innovation programme, via an Open Call issued and executed under Project EUROBENCH (grant agreement No 779963)