How people speak may indicate their risk of psychosis
The internet is filled with fun, language-based quizzes. Which English? will tell you if you speak more like an American or Canadian (among other choices); The Great English Dialect Quiz will try to guess your country of origin; and How Y’all, Youse and You Guys Talk will try to pinpoint your dialect down to the city.
What if scientists could look at your language patterns and determine more than simply your home country or state? What if they could use your language patterns to predict your future health?
Using language to recognize schizophrenia
Researchers at Columbia University, the New York State Psychiatric Institute and the IBM T. J. Watson Research Center have crafted a computer algorithm that can determine whether or not a person will develop schizophrenia. The study was published in the scientific journal Schizophrenia.
In this study, the researchers interviewed 34 young adults who were at risk of developing schizophrenia. After 2 ½ years, five of these subjects had demonstrated psychotic symptoms, whereas the rest of the subjects remained healthy.
The researchers used a computer algorithm to evaluate how each subject spoke during his or her initial interview. The program looked for sentence length, word usage and the overall meaning behind the words. Individuals who later went on to develop psychosis tended to use shorter sentences than the other subjects. They also used certain words less frequently (e.g., “that,” “what” and “which”) and sometimes abruptly changed topics from one sentence to the next.
When the computer was asked to use these speech features to predict which subjects would ultimately develop psychosis, it functioned at 100 percent accuracy.
Since the number of participants was relatively small, the researchers are hoping to perform the study again with a larger pool of at-risk individuals to make sure their algorithm continues to function properly. If they succeed, clinicians all over the world might start using language analysis to diagnosis and predict schizophrenia.
The future of language-based diagnosis
Schizophrenia is not the only disorder that scientists may soon be able to identify via language analysis. In 2015, a group of researchers found that individuals in the early stage of Parkinson’s disease processed action words (e.g. “applaud” or “punch”) differently than healthy subjects. A study published in 2013 was able to distinguish language differences between healthy individuals and individuals with PTSD on Twitter. A 2004 study published in the journal Cognition and Emotion found that college students with depression were more likely to use negative words and first-person pronouns than non-depressed college students.
Will language soon become the primary way to identify brain disorders? Only time will tell. In the meantime, pay a little more attention to how you speak. Which words do you use? Which words do you avoid? And what does that say about you?
The Sovereign Health Group provides patients with treatment programs designed to address mental illness by treating it at its source — the brain. Prior to treatment, our patients receive a thorough assessment that can include psychiatric evaluations and neurofeedback to identify — and ultimately treat — any underlying medical conditions that may influence their mental well-being. For more information, please contact us at our 24/7 helpline.
About the author
Courtney Lopresti, M.S., is a senior staff writer for the Sovereign Health Group where she uses her scientific background to write online blogs and articles for a general audience. At the University of Pittsburgh, where she earned her Master’s in neuroscience, she used functional neuroimaging to study how the human cerebellum contributes to language processing. In her spare time, she writes fiction, reads Oliver Sacks and spends time with her two cats and bird. Courtney is currently located in Minneapolis. For more information and other inquiries about this article, contact the author at email@example.com.