Unlike other medical specialties, psychiatry no clinical analysis that allows you to diagnose mental illness. To that end, the main tool of the psychiatrist is the interview with the patient. Thus, the specialist can determine whether a person who arrives at his office at risk of suffering a psychotic break-a temporary break with reality at some point in their lives.
But to establish the risk to effectively event occurs there is a long way. Not only because the psychotic event can take years to occur, but because it may be that eventually will never happen. That is, the doctor can not give assurance that it anticipates will be fulfilled.
Having prognostic indicators of mental illness is key to improving the clinical management of patients. In this context, the ability to predict with a high degree of certainty the occurrence of a psychotic break takes on special importance.
Much of the search for predictive indicators in psychiatry is aimed at discovering biochemical markers by a blood test, anticipate a mental health. But a team of researchers from the Faculty of Natural Sciences of the UBA (Exact UBA) is moving into a new predictive paradigm: instead of imagining a blood test, they focus on discourse analysis
.
“it is the first work of automatic discourse analysis computer that allows two and a half years anticipate a psychotic event,” said Diego Fernandez Slezak, CONICET researcher at the Laboratory of Artificial Intelligence Applied Exact UBA, one of the authors of the study published in the journal Schizophrenia
the study results are surprising. achieved 100% accuracy with their predictions. “I do not dare say that our technique has a predictive value of 100% accuracy. Because the sample is very small, “Slezak minimized.
The sample included 34 young people, from interviews with psychiatrists, had been diagnosed as” high risk of psychosis. ” Of these, two and a half years later, only five had developed a psychotic break. The same five people who had predicted the technique developed by the researchers.
Exact UBA scientists working from the assumption that what a person says or writes at certain times reflects his state of mind. “We asked if we could extract from a text, spoken or written, about what is going through the head of a person at the time he wrote or said that,” says Slezak.
On this idea for years, used a technique of automatic discourse analysis computer allowing them to “measure” and therefore “put numbers” different semantic and structural elements of interest present in a text. Then they develop and implement algorithms (computer programs) that teach the machine to find certain patterns within those numbers obtained.
The first of these studies, conducted in 2011, used these software developments to confirmed, quantitatively, a qualitative study by the American psychologist Julian Jaynes in the ’70s that, from the analysis of Western literature of all time, showed how evolved the phenomenon of insight into the human mind.
for this, the team Slezak “measured” the “amount of introspection” texts present in thousands of years, as the Bible, the Iliad and the Odyssey, and compared with the text produced . XX century
the result coincided with the point made by Jaynes decades: the reflective capacity of human beings is growing. “What begins in the early texts, with indications of the gods on the way that people should unthinkingly follow, it evolves over the centuries, to texts with personal reflections on what happens and what to do “illustrates.
This work caught the attention of a group of researchers from Columbia University in the United States who had desgrabados texts of interviews with people who consumed drugs, which are asked that bespoke a story of a loved one
the Americans asked to Slezak to implement computer technology on those texts to see if they could discriminate between four groups of individuals:. those who had taken ecstasy in high doses those who had done so at lower doses, those who had taken methamphetamine and those who had not ingested any drugs. Without knowing which group each of the texts, the Argentine developed an algorithm and applied to the analysis of these discourses.
“Among other things, we went to the Wikipedia article explaining the effects of ecstasy and we choose words from that content. For example: love, money and empathy. And we teach the computer, “says Slezak. “In this work we were able to discriminate against the members of each group with an accuracy of 85%” he reveals, and concludes. “This could help the psychiatrist to determine what drug a patient”
Source: nexciencia.exactas.uba.ar
No comments:
Post a Comment