With the help of artificial intelligence can speak?
Some of them are extremely serious diseases cause the Affected to lose in the end, the possibility of speaking. Researchers now succeeded to the language of the test subjects successfully to restore, by the brain activity of the people was read out, and this whole spoken sentences were generated.
In a recent study by the University of California, San Francisco the language could be made of volunteers, successfully, by all the sentences read by a new technology directly from the brain and then as a language were converted. The results of the study were published in the English journal “Nature”.
The process is complex and invasive
“For the first Time, a study shows that we can generate all the spoken sentences by the brain activity of a Person. This is a gratifying proof that we should be using the technology that is already in range of being able to build a device, which is suitable for patients with language loss clinically,“ explains study author Dr. Edward Chang from UC San Francisco in a press release. It is the process a complex and invasive process, it is not exactly decoded, what do you think of the subject, but what it actually said.
Where were the spoken words?
In the Experiment, which was conducted by the linguist Gopala Anumanchipalli, took subjects part, which have already been implanted with large arrays of electrodes for a different medical procedure in the brain. The researchers these people were several hundred sentences to read aloud while they are recorded by the electrodes, the recorded signals exactly. The authors of the study were able to detect a particular pattern of brain activity, which then arises, when subjects thought of words and certain cortical areas were associated with, before the final signals from the so-called motor cortex of the tongue and mouth muscles are to be sent. A kind of signal can be used to Reconstruct the language, say the researchers.
Virtual voting system, the party Concerned was created
Through the direct analysis of the audio material, the Team was able to determine which muscles and movements are affected, and it was then built a kind of virtual model of the voting system of the Person. Then the researchers transferred the during the session, recorded brain activity using a machine learning system on this virtual model. This allowed Essentially to the recording of brain activity, which is used to control the debates of the mouth.
The System determines which words are formed by facial muscles
It is important to understand that this process transforms abstract thoughts into words. It will be understood, rather, the specific instructions from the brain to the facial muscles, and is determined from the well-known words, which words would form these movements. The resulting synthetic speech is not crystal clear, but quite understandable. Set correctly, it could be the System in the location, of 150 words per Minute of a Person utter, which is not otherwise capable of language.
Accuracy achieved is an amazing improvement
In the future, should be worked on the ability to mimic the spoken language perfectly. Nevertheless, the accuracy that was achieved within the study, an amazing improvement in the real-time communication in comparison to the currently available options. For comparison: A Person with a degenerative disease of the muscles, for example, often speaks by spelling words with her eyes. This leads to approximately five to ten words per Minute, and other methods for disabled people are even slower. It is a kind of miracle, that the people affected can at all speak of this again, but this time-consuming and less natural method is far from the speed and expressiveness of a real language.
Further research is needed
If a Person would be able to the newly developed method, approach would be this much closer to the ordinary language, although perhaps at the expense of perfect accuracy, say the researchers. The Problem with this method was that it was very carefully collected data from a healthy system are needed by the brain to the tongue. For many people, it is not possible to collect this data, and for other invasive survey method is not recommended by Physicians. The good news is that the results are at least a beginning, and there are many conditions in which the new System would theoretically work. The collection of this critical brain and voice recorder data could also be done as a preventive measure, if a stroke or an occurring Degeneration is feared. (as)