A Speech Neuroprosthesis in the Frontal Lobe and Hippocampus: Decoding High-Frequency Activity into Phonemes

Area of Science:

• Neuroscience
• Biomedical Engineering
• Speech Technology

Background:

• Speech loss due to injury or disease significantly impacts quality of life.
• Existing neuroprostheses have limitations in harnessing diverse brain activity.
• Novel brain regions, including anterior cingulate and orbitofrontal cortices and the hippocampus, hold potential for speech control.

Purpose of the Study:

• To develop and evaluate a novel speech neuroprosthesis.
• To explore the use of high-frequency brain activity from previously unharnessed areas for speech articulation.
• To demonstrate silent control of a speech neuroprosthesis.

Main Methods:

• A speech neuroprosthesis was developed utilizing high-frequency activity from the anterior cingulate cortex, orbitofrontal cortex, and hippocampus.
• A neurosurgical epilepsy patient with intact speech participated in the study.
• The participant silently controlled the neuroprosthesis to produce vowel sounds.

Main Results:

• The neuroprosthesis achieved 85% accuracy in producing different vowel sounds during initial trials.
• Participant performance improved consistently over time, suggesting neuroplasticity.
• Silent control of the neuroprosthesis was demonstrated, even when trained on overt speech data.

Conclusions:

• This novel speech neuroprosthesis is clinically feasible for decoding high-frequency brain activity, including spiking activity.
• The technology enables silent phoneme production, offering a potential pathway for restoring speech.
• Early implantation in disease stages like amyotrophic lateral sclerosis could facilitate improved training and allow for silent control later on.