Background

Edward Chang is Chair of Neurological Surgery at UCSF and a researcher in speech neuroprosthetics. His laboratory has advanced the use of electrocorticography (ECoG) electrodes to decode intended speech from brain signals in paralyzed patients, establishing new methodological foundations for speech restoration BCIs.

Speech Neuroprosthetics Innovation

Chang developed recording and decoding techniques using ECoG electrode grids positioned on the brain’s surface to identify cortical activity encoding speech intentions. His approach directly targets speech motor representations in the motor cortex and adjacent language areas.

Real-time Speech Decoding

His laboratory demonstrated real-time decoding of intended speech from paralyzed patients, with a 2023 Nature paper (Metzger et al.) reporting 78 words per minute from a 1,024-word vocabulary. The approach translates decoded neural signals into synthesized speech or text output, enabling communication to family members and caregivers. In September 2026 his group reported in Nature Neuroscience (Brosler et al.) simultaneous real-time decoding of speech and upper-body gestures from a single 253-channel ECoG array, driving a virtual avatar for two BRAVO-trial participants.

Clinical Translation

Chang’s work establishes ECoG as a viable recording modality for clinical BCIs, with the advantage over penetrating intracortical electrodes that a single array resting on the brain’s surface, without entering tissue, can cover a broad area of speech cortex. His clinical investigations at UCSF bridge neurosurgical expertise with speech neuroscience.

Impact on the Field

By prioritising speech over gross motor control, Chang expanded the scope of BCI applications and addressed a distinct aspect of paralyzed patients’ most profound communication limitations. His methodological innovations inform competing commercial companies developing speech BCIs.