Validation of non-invasive EEG speech decoding in individuals with speech disabilities and attempted speech

Validate whether high-accuracy non-invasive EEG-based speech decoding, trained on overt speech from healthy individuals, can be achieved in individuals with speech disabilities, including using non-invasive recordings during attempted speech under data collection protocols analogous to those used for invasive measurements.

Background

The approach demonstrates strong performance using non-invasive EEG during overt speech from a healthy participant, suggesting potential for practical speech neuroprostheses without invasive procedures.

However, no data from individuals with speech disabilities or attempted speech were included, and the authors state that this hypothesis remains to be tested; they also stress the importance of validating non-invasive decoding using attempted speech paradigms commonly used in invasive studies.