Training a Covert Speech Brain-Computer Interface Through the Passive Perception of Speech
Bibliographic record
Abstract
Covert speech (CS), the mental imagery of speaking, is a ubiquitous trait of human cognition that has received increasing amounts of attention in brain-computer interface (BCI) research. However, collecting large amounts of data are challenging due to fatigue resulting from repeated mental rehearsal. Therefore, a more user-friendly paradigm is needed. Fortunately, investigations have revealed a linkage between CS and speech perception (SP) incurring spatiotemporal similarities. However, it remains unclear whether these correlations can be leveraged to decode CS based on data from SP during online BCI operation. Therefore, the overall goal of the thesis was to train a model to distinguish CS based on SP. Due to the lack of knowledge on how information is processed during CS, the first study compared the neural oscillatory characteristics between SP and CS using a studentized continuous wavelet transform (t-CWT) and phase-amplitude coupling (PAC). Across ten participants, the absence of within task PAC in CS and the presence of cross-task PAC between SP-CS suggested that θ activity was not the latent variable linking SP and CS, but rather that it is the γ-bands (30-60 Hz) that produce the shared features. Using the same data, in Study 2, this interaction was further investigated by measuring phase-coding, a theory of temporal encoding in the brain. Here, significant spatial correlations were observed in the γ-band synchrony profiles as well as in PAC, reinforcing the idea that the γ and θ-bands likely subserve, respectively, shared and distinct encoding processes across tasks. Given the identified correlations in how SP and CS temporally encode information, the objective of Study 3 was to produce an online BCI. Here, utilizing a unique dyadic protocol whereby SP and CS trials were presented in pairs, CS data were linearly transformed using SP and subsequently classified during online BCI operation. Across 10 participants, ternary online accuracies reached up to 93% and averaged at 75% ± 9.02 across all participants. Moreover, model performance correlated to γ-band correspondence between tasks. Collectively, the three studies demonstrated that the shared characteristics between SP and CS can be leveraged to train a CS BCI based on simply hearing words.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".