BCI4ALL: Interfaces Cérebro-Computador para Pessoas em Estado de Encarceramento Mental e com Perturbações da Consciência

BCI4ALL is an interdisciplinary project that aims at the R&D of new approaches, methods, theoretical and computational models, and respective clinical experimental validation, to advance EEG-based BCIs. These BCIs are intended to be used as Assistive Technology for people in LIS and CLIS, as well as an Assessment Tool for individuals in DoC. It is essential to note that BCIs in this project are designed to facilitate two levels of communication. For LIS patients, the expectation is that interfaces will enable communication through spellers, allowing letter-by-letter writing or word selection from a medium to large lexicon. For CLIS and DoC patients, the focus shifts to providing binary BCI responses in the form of YES/NO answers. Regardless of the communication level, BCIs must be personalized/tailored to each participant (personalized medicine). The project encompasses two general goals that are break down into subgoals: 1 – Research and develop BCIs for use by patients with LIS and transitioning from LIS to CLIS, with a focus on improving BCI usability. Specifically, our aims include: a) achieving BCIs that are calibration-free (or minimal calibration) across-sessions (i.e. generalizable classification models) b) enhancing natural interaction, encompassing: 1) self-paced control to enable patients to communicate at their own pace; 2) two-way (bidirectional) interaction, allowing patients to respond to caregiver prompts, but also to initiate conversation; and 3) using LLMs, generate communication content from sparse BCI output commands and synthesizing speech using the personality traits and patient’s “old” voice (this has never been done in this context) c) develop a multimodal interface that seamlessly integrates eye-tracking with the BCI, enabling concurrent or complementary use. 2 – Research and develop methods and approaches to enhance BCI performance for patients in CLIS, LIS with intermittent DoC and for assessing patients with DoC. The following sub-goals will be pursued: a) research the use of anodal tDCS to enhance discrimination of P300 event-related potentials (ERP) in CLIS patients (this has never been tested in CLIS population) b) design neural/BCI paradigms that can be used as off-the-shelf tools to assess patients at reflexive, perceptual and cognitive levels applicable to both CLIS and DoC patients; and automatic assessment of vigilance and awareness fluctuations. One of the challenges in current studies involving CLIS patients is the timing of BCI experiments, which typically start only after patients have transitioned into CLIS, often when they can no longer control eye-tracking systems. By the time BCIs are introduced, patients may lack the ability to provide feedback on the interface design and its personalization, posing difficulties in assessing their comprehension and usability. Furthermore, the period of non-communication during this transition phase raises concerns about the potential reduction of goal-directed thinking, suggesting a deterioration in executive function, arousal, and vigilance. To address this, we aim to conduct longitudinal studies beginning while patients still have communicative capabilities with eye-trackers, enabling the introduction of BCIs before they enter CLIS and providing insights into neuronal changes during the transition. Achieving this requires BCIs to offer improved usability to encourage patient acceptance and seamless integration with standard interfaces like eye-trackers.