[ad_1]
A trio of researchers have developed a Experimental machine learning method that allows the AI to listen to the early whispers of a psychotic break that humans can not hear.
The team consists of Neguine Rezaii of Harvard Medical School and Emory School of Medicine, Elaine Walker and Philipp Wolff of the Department of Psychology at Emory University, sought to determine if language could be used as an indicator. an impending latent psychosis.
They have developed a method of automatic learning that looks for specific indicators associated with psychosis, especially schizophrenia. The team then spent two years observing the volunteers of the study, a significant portion of which ended up presenting a psychotic break (first experience of a totally psychotic episode).
The results of the study were incredible. The team not only determined that its tool could experimentally predict a psychotic break with greater precision than human precision, but also discovered a new indicator of impending psychotic disruption. An area that could have extensive implications in the field of psychology: the ability to predict whether a person exhibits early signs of auditory hallucinations.
The medical profession accepts that a significant number of people with mental illnesses associated with psychosis have auditory hallucinations. Unfortunately, these positive symptoms (that is, HIV-positive people experience unobserved neurotypic symptoms) are usually seen too late in the prodromal (early) phase of psychosis, such as schizophrenia, to be very helpful. Early detection of such hallucinations could change the landscape of mental illness treatment.
According to the research of the team:
Our results indicate that during the prodromal phase of psychosis, the emergence of psychosis was predicted by a language with low levels of semantic density and an increased tendency to speak voices and sounds. Once combined, these two indicators of psychosis made it possible to predict with great precision the future psychosis.
Semantic density is the easiest. Psychologists know that people with prodromal psychosis generally communicate differently than those with neurotypic behavior. Those with low levels of semantic density are people who speak with little substance or context unless invited, even sparingly.
This is also called alogia, and can be characterized by the following table:
Researchers built an algorithm to detect semantic density, using over 30,000 Reddit messages to write a database. They then compared this information to the interviews conducted with the study participants to determine where the individuals were in relation to the standards.
However, it is a little more difficult to discern whether people or sounds that do not exist are heard, especially when these people have only mild hallucinations or are not at all aware of changes in their mental state.
It's not as simple as just asking patients many questions about what they hear in their head. Often, because people who end up with psychotic fractures have a low semantic density, the indices of their auditory hallucinations are too weak to be detected by scientists.
Rezaii, the newspaper's lead author, told Carol Clark of Emory University:
Trying to hear these subtleties in conversations with people is like trying to see the microscopic germs with the eyes. The automated technique we have developed is a very sensitive tool for detecting these hidden tendencies. It's like a microscope to detect the warning signs of psychosis.
In order to determine what essentially constitutes a handful of needles randomly distributed over an infinite number of potential haystacks, the researchers created a technique called "vector decompression" that determines the amount of meaning contained in a given sentence .
Basically, the machine determined that trial participants who were going to have a psychotic break – called converters – during the two years of the experiment were more likely to use words associated with sounds and noises than those who do not.
According to research:
More specifically, the language of converters tends to emphasize the subject of auditory perception, a group comprising the words "voice", "hear", "sound", "loud" and "singing". other, the words "murmur", "total" and "barely".
It is interesting to note that many words included in these groups – like the word whisper – have never been used explicitly by converters, but have been implicated in the general sense of their sentences.
When the researchers combined the system that looks for a low semantic density and the system that searches for auditory hallucination indications, they were able to determine with 93% accuracy if an individual would have a conversion and would have a psychotic break. Compare this to advanced methods involving interviews, analyzes and cognitive tests that only achieve an accuracy of 80%.
It is still too early to say what the real implications for this machine learning method will be, but the prognosis is certainly optimistic. There is no cure for psychosis, but early detection could provide one of our main weapons in the fight against mental illness. There is optimism that cognitive-behavioral therapy and targeted treatment could have a much greater impact on patients' perspectives if they were started before the onset of a psychotic breakup.
Those who suffer from psychosis have always asked for help, we have never had the right tools to understand exactly what they say. This breakthrough demonstrates how machine learning can be powerful and could represent a crucial moment for medical science, AI research, and human history.
Apple has just recorded 7 models of laptops that we could see this fall
[ad_2]
Source link