Open Research Within Brain-Computer Interfacing

There is not a one fits all approach [to Open Research], so it is important to find practices that are the best and tailored to your own project.

- Alicia Falcon Caro

Tell us a bit about you and your research.

Hi, I am Alicia Falcon Caro and I am a lecturer and a researcher in Cognitive Computing and Brain Informatics. I am involved in delivering lectures and lab sessions for a broad range of modules across most of the courses in the Computer Science Department. Each academic year, I propose and supervise several UG Final Year Projects and MSc Major Projects, mostly related to mobile application development, robotics, medical applications, AI applied to biomedical engineering, and brain-computer interfaces (BCI). My main research interest is in the area of artificial intelligence (AI) applied to medical applications

Tell us about a project you were involved in which used Open Research practices and principles?

The project presented in this case study is the project I am currently working on as part of my PGR programme.

The aims of this project is the development of advanced signal processing and machine learning techniques applied to brain-computer interfacing (BCI) for the rehabilitation of stroke patients with reduced mobility.

The research focus has been on adaptive cooperative networks and subspace analysis as two challenging areas in signal processing with applications to multi-subject (also known as hyperscanning) electroencephalography (EEG) based BCI.  We reformulated the common spatial pattern subspace technique for the hyperscanning scenario and deployed the concept of adaptive cooperative networks to multi-subject BCI for the first time. As part of the objectives of the project, we recorded and publicly released EEG hypercanning motor tasks signals from a number of volunteers for the validation of our techniques and to further encourage and support research in this field.

Describe the open research practice(s) employed in your study. Why did you select them?

The lack of publicly available EEG hyperscanning datasets, encouraged us to record, prepare and release our own dataset.  This gave us more control over the experiments and brought our research closer to the community through volunteers from different ages, genders and backgrounds, so we could obtain more inclusive data . The released of the dataset [1, 2] allowed us to contribute to the field and facilitate research on multi-brain EEG motor task applications.

We disseminated our research outputs through the publication of research papers, including open access [3], the publication of code and data through GitHub, and the participation in outreach activities. For traditional journals [4], we made the pre-print available via GitHub and deposited the accepted manuscript in IRep, the NTU institutional repository. The publication in open access and pre-prints   allows us to bring our research to a wider community, making it more accessible. On the other hand, the publication of code and data allows for a more transparent, reproducible, and reliable research.

Did you face any challenges in the project, and how did you overcome them?

One of the main challenges we encountered was the high publication fees for open access journals, which in non-funded projects like ours, is a discouragement to publish in these journals. Even so, we were able to overcome this challenge by selecting journals with open access policies permitting pre-prints and self-archiving of accepted manuscripts in institutional repositories and open platforms.

One of the other challenges we encountered was the involvement of more diverse participants in our recordings, which limited the involvement of a wider community in our research. Although our recording systems are non-invasive and low risk for the participants, each recording session requires the volunteers to be available for a considerable long time (from 1 to 5 hours approximately in most cases) and we are unable to provide any compensation (monetary or university credits for example) for their participation in the recordings. Even so, to overcome this challenge, we focused on reducing the time of each recording session and decided to develop techniques for small datasets. We also allowed the participants to look at their own brain recordings to spike their interest in our research and encourage their participation.

What has been the impact of adopting open research practice(s)in your project?

All the open research practices we are employing in our research are improving the visibility of our research. Working on a project that focuses on improving the rehabilitation and therefore, the lives of a minority of the society that have lost or partially lost the mobility of some of their limbs, allows us to work towards a more inclusive society. The use and publication of data from a wide range of people allows us not only to develop more inclusive techniques but also to promote a more inclusive research in this field.

On the other hand, through the sharing of pre-prints, accepted manuscripts and fully open journal papers, we were able to bring our research to a wider community, improving its accessibility and the visibility of our research group. This has led to an increase in possible collaborations for our group. Furthermore, the publication of code and data made our research more reliable and easily reproducible.

What did you learn from making this project ‘open’? Do you have any advice for others considering adopting open research practices?

Through my experience applying open research practices in this project, I learnt about the advantages that it brings and how publishing in open access journals is not the only option when trying to follow open research.

I would like to encourage everyone to consider being more open and remind them that there are different approaches to open research. There is not a one fits all approach, so it is important to find practices that are the best and tailored to your own project. Still, there are some open research practices that are easier to follow for most people, so I would advice them to start with these practices. Some of these practices are the dissemination of the research topic through public outreach activities, or the publication of code and data.

References

[1] Falcon Caro, A., Shirani, S., Ferreira, J. F., Bird, J. J., and Sanei, S., 2024. HyperCSP (Version v1.3) [Computer software]. https://github.com/AliciaFalconCaro/HyperCSP/

[2] A. Falcon-Caro, 2024. Formulation of Common Spatial Patterns for Multi-task Hyperscanning BCI Dataset. IEEE Transactions on Biomedical Engineering. https://doi.org/10.5281/zenodo.10523774.

[3] Falcon-Caro, A.; Peytchev, E., and Sanei, S., 2024. Adaptive Network Model for Assisting People with Disabilities through Crowd Monitoring and Control. Bioengineering, 11, 283. DOI: 10.3390/bioengineering11030283

[4] Falcon-Caro, A., Shirani, S., Ferreira, J. F., Bird, J. J., and Sanei, S., 2024. Formulation of Common Spatial Patterns for Multi-task Hyperscanning BCI. IEEE Transactions on Biomedical Engineering. doi: 10.1109/TBME.2024.3356665

Publications

• Falcon-Caro, A., and Sanei, S., 2021. Diffusion Adaptation for Crowd Analysis. 2021 International Conference on e-Health and Bioengineering (EHB), Iasi, Romania, 2021, pp. 1-4, doi: 10.1109/EHB52898.2021.9657730.
• Falcon-Caro, A., and Sanei, S., 2022. Cooperative Networking Approach to Assisting Blinds in a Crowd Using Air Trackers. 2022 4th International Conference on Emerging Trends in Electrical, Electronic and Communications Engineering (ELECOM), Mauritius, 2022, pp. 1-5, doi: 10.1109/ELECOM54934.2022.9965262.
• Falcon-Caro, A., and Sanei, S., 2023. Gesture Recognition via Estimation of Information Exchange between Muscles. In 2023 24th International Conference on Digital Signal Processing (DSP) (pp. 1-5). IEEE.
• Falcon-Caro, A., Frincu, M., and Sanei, S., 2023. A Diffusion Adaptation Approach to model Brain Responses in an EEG-based Hyperscanning Study. 2023 IEEE Statistical Signal Processing Workshop (SSP), Hanoi, Vietnam, 2023, pp. 393-397, doi: 10.1109/SSP53291.2023.10207972.

Participation in outreach activities:

• SoapboxScience in Nottingham: http://soapboxscience.org
• Open conferences at NTU (CIRC Research Showcase 2023) and UoN (2nd Workshop - Electric&Electronic Engineering - Career Path to Net Zero)
• International conferences:
  • Online International Conference on e-Health and Bioengineering (EHB 2021)
  • Hybrid International Conference on Emerging Trends in Electrical, Electronic and Communications Engineering (ELECOM 2022)
  • International Conference on Digital Signal Processing (DSP 2023)
  • IEEE Statistical Signal Processing Workshop (SSP 2023)

Alicia Falcon Caro | Nottingham Trent University