A team of neurobiologists, neurosurgeons, and engineers from Duke University has made a significant leap in biomedical science by creating a device capable of translating brain signals into words—a milestone for individuals suffering from speech-related disorders. This technology, published on November 6 in the journal Nature Communications, holds the potential to enable people with neurological disorders to communicate using a brain-computer interface.

Dr. Gregory Kogan from Duke University School of Medicine, co-leader of the project, explains that many individuals with conditions such as aphasia, ALS, or locked-in syndrome face the loss of speech abilities. Current communication tools are ineffective, exacerbating the isolation of patients. Currently available technologies allow speech decoding at a rate of approximately 78 words per minute, while natural human speech reaches around 150 words per minute.

The team utilized flexible, ultrathin, high-density brain sensors placed on a piece of postage-stamp-sized medical-grade plastic. These 256 microscopic sensors, arranged by Dr. Jonathan Viventi and his team, enable the differentiation of signals from neighboring brain cells, a crucial factor for accurately predicting speech.

In tests involving four patients, conducted in collaboration with neurosurgeons from Duke University Hospital, the device recorded motor speech cortex activity during simple tasks of listening and repeating words. The results were promising, although the decoder was accurate in only 40% of cases. Nevertheless, the experiment's success is significant, considering that the machine learning algorithm used was based on just 90 seconds of speech data collected during a 15-minute test.

Under the direction of Suseendrakumara Duraivela, the team is working on developing a wireless version of the device, a substantial step forward from the wired prototypes. Supported by a $2.4 million grant from the National Institutes of Health, this innovation could provide patients with freedom of movement without the constraints of wires in the future.

Despite the progress, the road to commercial implementation of the technology is long. The current speech rate generated by the device is much slower than natural speech, but efforts are underway to improve speed and accuracy. Dr. Viventi from Duke Magazine emphasizes that challenges persist, but the team is on a promising path to achieving the goal of enabling fluid communication for individuals affected by speech disorders.