Mental health

With AI, our words could be a window into our mental health

Share this post:

Written by: Anne Nicholson, Brand Strategy Lead IBM A/NZ

The global cost of mental health conditions is projected to surge to US$6.0 trillion by 2030. In Australia its estimated that 45% of people will experience a mental health issue in their life time. Approx. 1 million adults in any given year have depression – over 2 million have anxiety. Worldwide, up to one in 100 people will experience schizophrenia. Up to one person in fifty will develop bipolar disorder. It is fair to say that mental health is one of the greatest challenges facing Australia and New Zealand today.

Accurate diagnosis

One of the difficulties facing Doctors and sufferers is accurate diagnosis and identifying those who are at risk. Could computational biology, analytics and machine learning, offer solutions in the future that could quickly and simply analyse language and predict the onset of these diseases?  Could the results be earlier intervention, better allocation of resources or better treatment planning?

Guillermo Cecchi

Guillermo Cecchi

Research Staff Member in Biometaphorical Computing at IBM Research, Guillermo Cecchi, thinks so. “As a neuroscientist, I want to understand the brain. Beyond just the physical structures of neurons and the synapses, but how it works. How is it that we think? How is it that two pounds of protein and water can produce this amazing, complex organ that literally drives humanity?” he says. “Ultimately, behavior is what the brain is for. We, in the scientific and medical community, are studying behaviour with the same types of computational approaches that we use to study the physical attributes and workings of the brain.

Characteristics of schizophrenia

One of the characteristics of schizophrenia is disorganized thought, which can present itself in disjointed patterns in speech.  Several studies of at-risk youths have shown around 79 percent of doctors were able to accurately predict those young people who would develop psychosis based on tracking that person’s speech patterns in interviews. However, in a study with Columbia University psychiatrists and IBM, we were able to predict, with 100 percent accuracy, who among a population of at-risk adolescents would develop their first episode of psychosis within two years. In other research with Pfizer, using only about 1 minute of speech from Parkinson’s patients, we were able to better track, predict and monitor the disease. Results are nearly 80 percent accuracy.

“In five years, we hope to advance the study of using words as windows into our mental health,” says Guillermo. “Some days I think I am more philosopher than scientist, but I am often reminded that those roles, like those of the neurons and behavior in the brain, are two halves of the same function. The field of neuroscience is moving quickly – we know so much, but still have much more to uncover.”

More Mental health stories

Fractals – a way to see infinity

Edited by: Anne Nicholson, Brand and Content A/NZ Ever wonder why a river, looks like a tree, looks like a leaf? IBMer Benoit Mandelbrot did. His quirky way of looking at the world led to the discovery of fractal geometry. Fractals have made it possible to mathematically explore the kinds of rough irregularities that exist […]

Continue reading

Guided by a sense of purpose

Author: David La Rose, Managing Director at IBM Australia and New Zealand I felt particularly proud to be an IBMer when I sat down to watch the ABC’s Catalyst Program talking about the advances we are making in robotics and the impact it’s having on the lives of so many Aussies. On the show, Dr […]

Continue reading

Data in Disasters: Coding our way to improved disaster resilience

Author: Iven Mareels, Lab Director of IBM Research – Australia Australia and New Zealand are among the most disaster-prone countries in the world, each year experiencing the wrath of bushfires, cyclones, floods and earthquakes. More than 9 million Australians have been impacted by a natural disaster or extreme weather event in the past 30 years, […]

Continue reading