Towards a Brain-Computer Interface (BCI) for Improving Phonological Processing in Developmental Dyslexia: An Exploratory Study
Date:
Towards a Brain-Computer Interface (BCI) for Improving Phonological Processing in Developmental Dyslexia: An Exploratory Study
Purpose Brain–computer interfaces (BCIs) have considerable potential for enhancing automatic features of neural processing, but have typically been developed for non-linguistic disorders. Here, we developed a non-invasive EEG-BCI designed to train neural processing related to phonology during natural speech listening in developmental dyslexia. As a first step towards developing a child-focused intervention, the BCI was tested with adults with and without dyslexia.
Method An operant learning paradigm using mobile EEG (Araujo et al., 2023) was designed to help participants reduce their theta/delta oscillatory ratio during natural speech listening. According to Temporal Sampling theory, theta (~4–8 Hz) and delta (~0.5–4 Hz) neural oscillations support the encoding of rhythmic and acoustic information in speech, while a higher theta/delta ratio has been associated with poorer phonological awareness in dyslexic children (Araujo et al., 2024). Participants completed 16 BCI runs across eight training days. During story listening, they learned to keep a spaceship in the upper half of a computer screen, with a lower theta/delta ratio moving the spaceship upwards. Participants completed pre- and post-tests assessing phonological awareness, single-word and nonword reading, rapid naming, and amplitude rise-time discrimination. BCI learning was assessed using both real-time online scores and offline scores calculated following more stringent EEG artefact removal.
Results Learning to reduce the theta/delta ratio was observed in both the dyslexic and control groups, although fewer participants met the learning criterion following more stringent offline EEG preprocessing. Among dyslexic offline learners, significant pre–post improvements were observed in syllable-stress recognition, rapid naming, and single-word and nonword reading. In the principal analysis pooling all 16 offline BCI learners across groups, stronger theta/delta reduction was significantly associated with greater improvements in single-word reading (r = .61), nonword reading (r = .49), and amplitude rise-time discrimination (r = −.55; all FDR-corrected p < .05).
Conclusions Adults with and without dyslexia were able to learn to regulate theta/delta ratio during natural speech listening. The associations between stronger BCI learning and improvements in reading, and auditory rise-time measures suggest that the theta/delta ratio is a promising neural target for the development of BCIs for children with developmental dyslexia, consistent with Temporal Sampling theory (Zheng et al., 2026).