Journal
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
Volume 29, Issue -, Pages 2605-2614Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TNSRE.2021.3133853
Keywords
Visualization; Electroencephalography; Task analysis; Grasping; Thumb; Stroke (medical condition); Steady-state; Action observation (AO); brain-computer interface (BCI); event-related desynchronization (ERD); steady-state motion visual evoked potential (SSMVEP); rehabilitation after stroke
Categories
Funding
- National Natural Science Foundation of China [31771069, 31800824]
- National Key Research and Development Program of China [2020YFC2004200]
- Chongqing Science and Technology Program [cstc2018jcyjAX0390]
- China Postdoctoral Science Foundation [2021M700605]
Ask authors/readers for more resources
The study introduced a novel online action observation (AO)-based BCI with visual stimuli inducing steady-state motion visual evoked potentials (SSMVEP) and activating sensorimotor regions to enhance rehabilitation training. Results showed improved classification accuracy and enhanced event-related desynchronization (ERD) compared to the AO-alone scenario, suggesting increased attentiveness as a key factor in the AO-based BCI's effectiveness.
Brain-computer interface (BCI)-based stroke rehabilitation is an emerging field in which different studies have reported variable outcomes. Among the BCI paradigms, motor imagery (MI)-based closed-loop BCI is still the main pattern in rehabilitation training. It can estimate a patient' motor intention and provide corresponding feedback. However, the individual difference in the ability to generate event-related desynchronization (ERD) and the low classification accuracy of the multi-class scenario restrict the application of MI-based BCI. In the current study, a novel online action observation (AO)-based BCI was proposed. The visual stimuli of four types of hand movements were designed to simultaneously induce steady-state motion visual evoked potential (SSMVEP) in the occipital region and to activate the sensorimotor region. Task-related component analysis was performed to identify the SSMVEP. Results showed that the amplitude of the induced frequency in the SSMVEP had a negative relationship with the stimulus frequency. The classification accuracy in the four-class scenario reached 72.81 +/- 13.55% within 2.5s. Importantly, the AO-based closed-loop BCI, which provided visual feedback based on the SSMVEP, could enhance ERD compared with AO-alone. The increased attentiveness might be one key factor for the enhancement of the ERD in the designed AO-based BCI. In summary, the proposed AO-based BCI provides a new insight for BCI-based rehabilitation.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available