Novel TMS-derived metrics enable machine learning classification of major depressive disorder

Area of Science:

• Neuroscience
• Biomarker Discovery

Background:

• Major depressive disorder (MDD) lacks validated biomarkers for early detection and personalized treatment.
• Transcranial magnetic stimulation (TMS) is a clinical tool with potential for identifying neurophysiological biomarkers.

Purpose of the Study:

• To evaluate two novel TMS-derived cortical excitability metrics, delta and rho, as potential biomarkers for distinguishing MDD patients from healthy controls.
• To assess the diagnostic accuracy of these metrics, alone and in combination with motor-evoked potentials (MEPs).

Main Methods:

• Recorded motor-evoked potentials (MEPs) from the abductor pollicis brevis muscle during TMS of the right primary motor cortex.
• Calculated delta and rho metrics from MEP amplitudes in 26 unmedicated MDD patients and 17 healthy controls.
• Utilized a Gradient Boosting classifier to predict diagnostic status based on MEPs, delta, rho, or their combination.

Main Results:

• Raw MEPs alone were not predictive of diagnostic status.
• The delta and rho metrics significantly improved classification accuracy.
• Combining MEPs with delta and rho achieved 83.3% accuracy and 82.3% balanced accuracy in distinguishing MDD patients from controls.

Conclusions:

• The delta and rho metrics effectively capture neurophysiological alterations associated with MDD.
• These findings support the potential of delta and rho as candidate biomarkers for the early detection and personalized treatment of major depressive disorder.