Machine Learning Tool Improves Diagnosis of Functional Seizures

Source: Epilepsia

Summary

This study looked at how a new computer-based tool, called the Functional/Dissociative Seizures Likelihood Score (FSLS), can help doctors diagnose functional or dissociative seizures (FDS). The researchers involved 163 readers, including doctors and mental health clinicians, who reviewed 1,142 cases of patients with different types of seizures. They wanted to see if using the FSLS would improve the accuracy of diagnosing FDS compared to traditional methods.

The key findings showed that many readers performed better than random guessing when using the FSLS. About 66% of the readers improved their accuracy by 12% to 15% when they used the FSLS, especially those who had some training in seizures. However, not all readers benefited equally; some groups did not see any improvement in their diagnostic performance.

This research is important because it suggests that using machine learning tools like the FSLS can help doctors make better decisions when diagnosing complex seizure disorders. However, the study also has limits, as not all clinicians improved, and having more training did not always guarantee better results. This indicates that while the FSLS can be helpful, it should be used in specific situations to be most effective.

Original source