Evaluation of stimulation conditions for the design of Brain-Computer Interfaces based on visual stimuli.

RUBEN ISAAC CARIÑO ESCOBAR (2023)

https://orcid.org/0000-0003-3377-0813

Brain-computer interfaces (BCIs) have emerged as promising technologies for enabling direct communication and control between the human brain and external devices. The effectiveness and efficiency of BCIs heavily rely on the selection of optimal specificity criteria, which determine the design parameters used to elicit and interpret brain signals. This thesis presents a comprehensive investigation into the selection of specificity criteria for BCIs based on visual Event-Related Potentials (P300) and Steady-State Visual Evoked Potentials (SSVEP).

The study focuses on three main findings. Firstly, it demonstrates the superiority of the Cartoon Face (CF) visual scheme over the Standard Flash (SF) in eliciting a more prominent P300 brain response and achieving better single-trial classification of target elements from non-target elements. Secondly, the investigation reveals that the combination of either the On-Off rectangular (OOR) or sinusoidal (OOS) luminance-modulated visual stimulus, along with the Filter-Bank Canonical Correlation Analysis (FBCCA) as a detection method, yields the best performance for SSVEP-based BCI assessments.

Finally, the research highlights the superiority of the sitting posture over the supine posture in P300 paradigm evaluations. The findings of this research have significant implications for the design and implementation of BCIs. Practitioners and designers are recommended to consider incorporating the Cartoon Face visual scheme as it enhances P300 responses and single-trial classification effectiveness. For SSVEP-based BCIs, the use of OOR or OOS visual stimuli with FBCCA as I the detection method should be prioritized. Moreover, adopting the sitting posture during P300 paradigm assessments can lead to improved BCI performance.

Despite the valuable insights gained, this research acknowledges certain limitations. The selection of specificity criteria may exclude other relevant factors that could impact BCI performance. Additionally, the research methodology and data collection process may have potential biases that can arise if we have preconceived expectations or preferences for certain outcomes.

To address these limitations, future investigations should explore alternative specificity criteria, conduct unbiased data collection and analysis, and evaluate the generalizability of findings in real-world applications.

In conclusion, this thesis provides a comprehensive exploration of specificity criteria for BCIs based on P300 and SSVEP. The identified findings underscore the importance of considering visual stimuli, detection methods, and posture in BCI design. By implementing the recommended specificity criteria and addressing the limitations, practitioners and researchers can advance the development of more effective and reliable BCIs, leading to enhanced user experiences and expanding the potential applications of these transformative technologies.

Doctor of Philosophy in Engineering Sciences

Artículo

INGENIERÍA Y TECNOLOGÍA CIENCIAS TECNOLÓGICAS OTRAS ESPECIALIDADES TECNOLÓGICAS OTRAS