Image Credit: serrnovik/123rf.com
Researchers at Stanford University have made significant strides in brain-computer interface (BCI) technology, revealing a system capable of decoding “inner speech”—the words individuals think but do not vocalize. Published in the journal Cell, the study presents a groundbreaking approach that could transform communication for individuals with conditions such as amyotrophic lateral sclerosis (ALS).
Traditional BCI systems require users to physically attempt to speak, which can be exhausting and limiting, especially for those with severe motor impairments. The Stanford team’s innovation allows users to simply imagine the words they wish to express, bypassing the physical strain and enabling longer usage of the system.
During trials, four ALS patients, including veteran participant Casey Harrell, engaged with the new BCI. Previously, Harrell had regained some speech capabilities using earlier BCI technology, but this latest phase focused on decoding pure inner thoughts. Researchers trained custom artificial intelligence models to improve the accuracy of translating brain activity into coherent phrases, achieving impressive results in decoding complex sentences.
However, a significant concern arose regarding mental privacy, as the BCI occasionally decoded unintended thoughts. To address this, the researchers implemented a unique “mental safe word” system. Participants were instructed to think of a whimsical phrase—”Chitty Chitty Bang Bang”—to signal when the system should begin or cease listening. This innovative solution yielded a remarkable success rate of 98.75% in controlling the system’s responses.
While still in the early stages, this work serves as a proof-of-concept, demonstrating the potential of BCIs to decode meaningful inner speech while respecting user privacy. The Stanford researchers aim to create a more dignified and less taxing communication method for those affected by debilitating conditions, marking a promising advancement in assistive technology.
Check out the original article here: Source link
