Study Differentiates Schizophrenia With 88% Sensitivity Using 213 SPECT Scans

2 articles · Updated · PR Newswire · Jul 6

A NeuroImage: Reports study found whole-brain SPECT imaging paired with machine learning distinguished schizophrenia patients from healthy controls, with random forest reaching 88% sensitivity and logistic regression 87%.
The analysis used 213 brain scans—137 from diagnosed patients and 76 from controls—and applied spatially constrained independent component analysis to extract network-level features linked to altered brain function.
Logistic regression and random forest outperformed support vector machines, while specificity was lower at 68% and 61%, suggesting the approach is more effective at identifying cases than ruling them out.
Visual processing and cognitive control networks were among the strongest predictors, with the middle occipital gyrus, subthalamus and putamen highlighting schizophrenia's broad disruption across interconnected brain systems.
Researchers said larger, more balanced studies are needed before the method could support objective diagnosis, risk assessment or treatment monitoring in psychiatry.

Sources

PR Newswire11d ago

Study by TReNDS and Amen Clinics Uses SPECT Imaging and Machine Learning to Differentiate Schizophrenia

AOL11d ago

New Study Reveals Powerful New Method Can Distinguish Schizophrenia from Healthy Using Brain SPECT Imaging - AOL

This new brain scan tool can identify schizophrenia, but will it actually improve patient outcomes more than existing methods?

A clinic's own data built a new diagnostic AI. Is this a breakthrough or a major conflict of interest?

Landmark 2026 Study: SPECT and AI Distinguish Schizophrenia with High Accuracy, Paving the Way for Personalized Psychiatry

Overview

A landmark 2026 study published in NeuroImage: Reports marks a pivotal moment in schizophrenia diagnosis and understanding. By combining advanced brain imaging techniques with cutting-edge machine learning algorithms, researchers achieved high accuracy in distinguishing schizophrenia. The study analyzed intricate brain network abnormalities and developed diagnostic models with remarkable precision. This achievement reinforces the view of schizophrenia as a disorder involving widespread brain network alterations, not just localized issues. The findings open new possibilities for more objective, accurate, and personalized approaches to diagnosis and treatment, representing a significant advancement in neuroscience and psychiatry.

...