Predicting psychosis before symptoms appear
summary: Researchers have developed a machine learning tool that accurately identifies individuals at risk of developing psychosis through MRI brain scans. This innovative approach, which achieved an 85% accuracy rate in training and 73% using new data, provides a promising avenue for early intervention in psychosis, potentially improving treatment outcomes.
The study included more than 2,000 participants from 21 global sites, highlighting the potential of the tool in diverse clinical settings. By detecting structural differences in the brain before the onset of psychosis, this tool represents a major advance in psychiatric care, with the goal of improving prediction and prevention strategies.
- A machine learning classifier can distinguish between individuals at high risk of developing psychosis and those not at risk with high accuracy, using MRI brain scans.
- Early identification of psychosis risk through MRI scans can lead to more effective interventions and reduce the impact on individuals’ lives.
- The research underscores the need for further development to ensure the classifier can be applied across different datasets and clinical settings.
source: University of Tokyo
The onset of psychosis can be predicted before it happens, using a machine learning tool that can classify MRI brain scans into those who are healthy and those at risk of having a psychotic episode.
An international consortium, including researchers from the University of Tokyo, used the classifier to compare scans of more than 2,000 people from 21 global locations. Nearly half of the participants were identified as being clinically at risk for psychosis.
Using training data, the classifier was 85% accurate in distinguishing between people who were not at risk and those who later experienced clear psychotic symptoms.
Using new data, the accuracy was 73%. This tool could be useful in future clinical settings, as while most people with psychosis make a full recovery, early intervention usually leads to better outcomes with less negative impact on people’s lives.
Anyone may experience a psychotic episode, which usually includes delusions, hallucinations, or disorganized thinking. There is no single cause, but it can be caused by illness or injury, trauma, drug or alcohol use, medication, or genetic predisposition.
Although psychosis can be frightening or disturbing, it is treatable and most people recover. Because the most common age for a first seizure is during adolescence or early adulthood, when the brain and body are going through a lot of changes, it can be difficult to identify young people who need help.
“At most, only 30% of clinically at-risk individuals experience obvious psychotic symptoms, while the remaining 70% do not,” explained Associate Professor Shinsuke Koike of the University of Tokyo Graduate School of Arts and Sciences.
“Therefore, clinicians need to help identify those who will experience psychotic symptoms using not only subclinical signs, such as changes in thinking, behavior and emotions, but also some biological markers.”
A consortium of researchers has worked together to create a machine learning tool that uses brain MRI scans to identify people at risk of psychosis before it starts. Previous studies using brain MRI have indicated that structural differences occur in the brain after the onset of psychosis.
However, this is said to be the first time that differences have been identified in the brains of those who are at very high risk but have not yet suffered from psychosis.
The team from 21 different institutions in 15 different countries brought together a large and diverse group of participating adolescents and young adults.
According to Koike, MRI research in psychotic disorders can be difficult because differences in brain development and in MRI machines make it difficult to obtain highly accurate and comparable results. Also, for young people, it may be difficult to distinguish between changes caused by typical development and those caused by mental illness.
“Different MRI models have different parameters that also affect the results,” Koike explained.
“Just as with cameras, various tools and imaging specifications create different images of the same scene, in this case the participant’s brain. However, we were able to correct for these differences and create a classifier that is well-tuned to predict the onset of psychosis.”
Participants were divided into three groups of people at high clinical risk: those who later developed psychosis; Those who did not develop psychosis. and people with uncertain follow-up status (1,165 people in total for all three groups), and a fourth group of healthy controls for comparison (1,029 people). Using the scans, the researchers trained a machine learning algorithm to identify patterns in the participants’ brain anatomy.
From these four groups, the researchers used the algorithm to classify participants into two main groups of interest: healthy controls and those at high risk who later developed clear psychotic symptoms.
In training, the tool was 85% accurate in classifying outcomes, while in final testing using new data it was 73% accurate in predicting which participants were most at risk of developing psychosis.
Based on the findings, the team believes that performing brain MRI scans in people identified as being at clinically high risk may be useful for predicting the onset of psychosis in the future.
“We still have to test whether the classifier will work well with new sets of data. Since some of the software we used is best for a static data set, we need to build a classifier that can robustly classify MRIs from new sites and machines, which is a challenge that It’s being faced by a national brain science project in Japan, called Brain/MINDS Beyond.“Now we’re taking over,” Koike said.
“If we can do this successfully, we can create more powerful classifiers for new datasets, which can then be applied to real-world and routine clinical settings.”
Financing: This research was supported in part by AMED (Grant Nos. JP18dm0307001, JP18dm0307004 and JP19dm0207069), JST Moonshot R&D (JPMJMS2021), JSPS KAKENHI (JP23H03877 and JP21H02851), Takeda Science Foundation and Senshin Medical Research Foundation. This study was also supported by the International Neurointelligence Research Center (WPI-IRCN), University of Tokyo.
About this psychosis research news
author: Joseph Kreischer
source: University of Tokyo
communication: Joseph Kreischer – University of Tokyo
picture: Image credited to Neuroscience News
Original search: Open access.
“Using brain structural neuroimaging measures to predict the onset of psychosis in clinically high-risk individuals” by Shinsuke Koike et al. Molecular psychiatry
Using structural brain neuroimaging measures to predict the onset of psychosis in individuals at high clinical risk
Machine learning approaches using structural magnetic resonance imaging (sMRI) can be useful for disease classification, although their ability to predict psychosis is largely unknown.
We created a model with CHR individuals who later developed psychosis (CHR-PS+) from healthy controls (HCs) that could discriminate between each other.
We also assessed whether we could distinguish CHR-PS+ individuals from those who did not later develop psychosis (CHR-PS-) and those with uncertain follow-up status (CHR-UNK). T1-weighted structural brain MRI of 1165 CHR individuals (CHR-PS+, n= 144; Human Rights Committee-PS-, n= 793; And the Center for Human Rights-UNK, n= 228), and 1029 HCs were obtained from 21 sites.
We used ComBat to harmonize measures of subcortical volume, cortical thickness, and surface area data and correct for nonlinear effects of age and sex using a general additive model. Human Rights-PS+ (n= 120) and HC (n= 799) Data from 20 sites served as a training dataset, which we used to build a classifier.
The remaining samples from the external validation datasets were used to evaluate classifier performance (test, independent confirmation, and independent ensemble (CHR-PS- and CHR-UNK) datasets). The accuracy of the classifier on the training and independent confirmatory datasets was 85% and 73%, respectively.
Measures of regional cortical surface area—including those in the right superior frontal cortex, right superior temporal cortex, and bilateral insular cortex—contributed strongly to the CHR-PS+ classification of HC. CHR-PS- and CHR-UNK individuals were more likely to be classified as HC than CHR-PS+ (classification rate to HC: CHR-PS+, 30%; CHR-PS-, 73%; CHR-UNK, 80%). .
We used multisite magnetic resonance imaging (sMRI) to train a classifier to predict the onset of psychosis in CHR individuals, and it showed promising prediction of CHR-PS+ in an independent sample.
The results suggest that when considering adolescent brain development, baseline MRI scans for CHR individuals may be useful in determining their prognosis.
Future prospective studies are needed on whether the classifier can actually be useful in clinical settings.