A University of Melbourne spinout, Fluent, is developing an insertable brain–computer interface that sits under the scalp but outside the skull to decode attempted speech, the company said.
The device is positioned above the motor cortex, the brain area that controls speech muscles. Fluent co‑founder and biomedical engineer Dr Tim Mahoney said the device records distinct electrical patterns produced when someone speaks or tries to speak. “If you think of electrical signals as QR codes: when a person speaks, every individual mouth and jaw movement produces different ‘QR codes’ in their motor cortex,” he said.
Fluent is building a machine‑learning model to convert the recorded brain activity into text or audio so non‑verbal users can communicate without selecting letters or using eye tracking. Dr Mahoney said his PhD work showed signal quality is comparable whether electrodes sit beneath the scalp or outside it, allowing the team to avoid drilling into the skull.
Early human testing took place in electromagnetically shielded rooms at the Aikenhead Centre for Medical Discovery at St Vincent’s Hospital Melbourne. In those tests participants wore 144 scalp electrodes while speaking, miming and imagining phrases. Fluent says it built what it calls the largest English dataset of its kind and partnered with a Japanese team with an even larger dataset. The collaboration, the company reported, produced a model that isolated the correct phrase from a set of 128 options with 96% accuracy.
Clinical studies of Fluent’s insertable electrodes are scheduled later this year, the company said. Dr Mahoney described the device’s safety profile as better than a routine cochlear implant.
Fluent has raised more than $2 million from investors including the University of Melbourne Genesis Pre‑Seed Fund, Galileo Ventures, Multiple Sclerosis Western Australia and several international backers, the company said.
Photo credit: www.unimelb.edu.au
Tags: subscalp electrodes, motor cortex, speech BCI, machine learning
Topics: Neurotech industry & startups, Brain–computer interfaces, EEG & neuro-sensing headsets