Scientists from Columbia University are a step closer to developing a method for translating brainwaves directly into human speech.
The new technology will be aimed at assisting patients who are unable to speak, and although this development is still a long way from being used as a standard method of communication for all non-verbal patients, it is a massive step in the right direction to overcoming speech difficulties, thus enabling communication for those who can’t talk for whatever reasons.
The scientists at Columbia University were given access to patients at Northshore University Hospital who were undergoing brain surgery. The patients were asked to participate in the study, which saw them undergo chronic intracranial encephalography (iEEG) monitoring prior to surgery, so that researchers could obtain high resolution brain scans.
In order to translate the brainwaves into language, the researchers asked each patient to listen to verbal communications while their iEEG monitors recorded brain activity. An artificial intelligence (AI) was then trained to study the brainwave patterns in a bid to infer meaning from them.
After many trials and adjustments, the AI was able to correctly deduce what each patient was hearing, solely from the reading on the iEEG, and reproduce the words with clarity so that someone with normal hearing could understand each AI translated word, and could even suggest if they thought the voice was that of a male or female.
For example, each patient listened to a recital of counting from zero to none, which the AI then reproduced clearly enough to be understood by the listeners, who heard the translation through headphones.
Although this may sound slightly intrusive to some (a machine that can ‘read your mind’ is the stuff of futuristic thrillers after all) the AI isn’t actually accessing the private thoughts of each patient, but is instead translating the overheard speech by interpreting brainwaves.
This means that mental communication (or reading another’s mind) isn’t possible with this technology as it stands at present, but the study does show that it is possible to understand brainwaves translated from an iEEG into a verbal language.
The technology used in this study still needs improvement in order to be used in the mainstream. An advanced translation algorithm and a deeper understanding of the brain itself, could eventually lead to giving a voice to those who have no power of speech.
Recent Comments