It is known that the semantic richness of people’s conversational language is an indicator of psychosis. The thing is, it is pretty much impossible to hear the difference, the same way it is impossible to see something microscopic with the naked eye. That’s why scientists at Emory University and Harvard University developed a new machine-learning method to be able to quantify it more precisely.
Emory psychologist Phillip Wolff, senior author of the study, said:
It was previously known that subtle features of future psychosis are present in people’s language, but we’ve used machine learning to actually uncover hidden details about those features. Machine learning technology is advancing so rapidly that it’s giving us tools to data mine the human mind.
There are two language variables this automated analysis processes to predict whether an at-risk person will later develop psychosis with 93% accuracy. They are more frequent use of words associated with sound and speaking with low semantic density, or vagueness. Meaning, the frequent use of certain words associated with their sound are the “hidden clue in people’s language” predictive of the later emergence of psychosis. The journal npj Schizophrenia published the researchers’ findings.
Co-author Elaine Walker, an Emory professor of psychology and neuroscience who researches how schizophrenia and other psychotic disorders develop, said:
Our finding is novel and adds to the evidence showing the potential for using machine learning to identify linguistic abnormalities associated with mental illness.
It’s such a difficult thing to notice that trained clinicians had not perceived how people at risk for psychosis use more words associated with sound than the average, even though abnormal auditory perception is a pre-clinical symptom.
Neguine Rezaii, first author of the paper, said:
Trying to hear these subtleties in conversations with people is like trying to see microscopic germs with your eyes. The automated technique we’ve developed is a really sensitive tool to detect these hidden patterns. It’s like a microscope for warning signs of psychosis.
The onset of schizophrenia and other psychotic disorders normally occurs in the early 20s. However, there are warning signs (known as prodromal syndrome) beginning around age 17. Around 25 to 30% of youth who meet criteria for a prodromal syndrome will develop schizophrenia or another psychotic disorder.
Rezaii said:
In the clinical realm, we often lack precision. We need more quantified, objective ways to measure subtle variables, such as those hidden within language usage.
Currently, trained clinicians can predict psychosis with about 80% accuracy in those with a prodromal syndrome using structured interviews and cognitive tests. Machine-learning research is among the many ongoing efforts to fine-tune diagnostic methods, identify new variables, and improve the accuracy of predictions.
There is no cure for psychosis, but the researchers believe it can be prevented. Walker said:
If we can identify individuals who are at risk earlier and use preventive interventions, we might be able to reverse the deficits. There are good data showing that treatments like cognitive-behavioral therapy can delay onset, and perhaps even reduce the occurrence of psychosis.
The Study
- First, the researchers used machine learning to establish “norms” for conversational language.
- They fed a computer software program (known as Word2Vec) the online conversations of 30,000 users of Reddit (a social media platform where people have informal discussions about a range of topics).
- The program used an algorithm to change individual words to vectors, assigning each one a location in a semantic space based on its meaning – those with similar meanings are positioned closer together than those with far different meanings.
- The Wolff lab then developed a computer program to perform what the researchers dubbed “vector unpacking,” which is an analysis of the semantic density of word usage.
- Their work using Word2Vec has measured semantic coherence between sentences, while vector unpacking allowed them to quantify how much information was packed into each sentence. By combining these they generated a baseline of “normal” data.
- Next, the researchers applied the same techniques to diagnostic interviews of 40 participants that had been conducted by trained clinicians, as part of the multi-site North American Prodrome Longitudinal Study (NAPLS) – a 14-year project focused on young people at clinical high risk for psychosis.
- Lastly, they took the automated analyses of the participant samples and compared them to the normal baseline sample as well as the longitudinal data on whether the participants converted to psychosis.
The Results
- A higher than normal usage of words related to sound, combined with a higher rate of using words with similar meaning, meant that psychosis was likely on the horizon.
Rezaii and Wolff are now gathering larger data sets and testing the application of their methods on a variety of neuropsychiatric diseases, including dementia. “This research is interesting not just for its potential to reveal more about mental illness, but for understanding how the mind works — how it puts ideas together,” Wolff says.
