A high-performance neuroprosthesis for speech decoding and avatar control using ECoG – Kaylo Littlejohn

Talk Description:

Speech neuroprostheses have the potential to restore communication to people living with paralysis, but naturalistic speed and expressivity are elusive. Here we use high-density surface recordings of the speech cortex in a clinical-trial participant with severe limb and vocal paralysis to achieve high-performance real-time decoding across three complementary speech-related output modalities: text, speech audio and facial-avatar animation. We trained and evaluated deep-learning models using neural data collected as the participant attempted to silently speak sentences. For text, we demonstrate accurate and rapid large-vocabulary decoding with a median rate of 78 words per minute and median word error rate of 25%. For speech audio, we demonstrate intelligible and rapid speech synthesis and personalization to the participant’s pre-injury voice. For facial-avatar animation, we demonstrate the control of virtual orofacial movements for speech and non-speech communicative gestures. The decoders reached high performance with less than two weeks of training. Our findings introduce a multimodal speech-neuroprosthetic approach that has substantial promise to restore full, embodied communication to people living with severe paralysis.

About the speaker:

Kaylo Littlejohn is an EECS Ph.D. student at UC Berkeley advised by Professor Gopala K. Anumanchipalli and Dr. Edward F. Chang. His research is focused on restoring lost function to those who have severe paralysis via the use of speech and avatar neuroprostheses. Kaylo’s research enabled the first demonstration of intelligible speech-synthesis from neural activity in a person who has lost the ability to communicate.