EEG microstate-based classification using machine learning in depressed adolescents with and without non-suicidal...

Area of Science:

• Neuroscience
• Psychiatry
• Computational Neuroscience

Background:

• Non-suicidal self-injury (NSSI) is a significant concern in adolescents, especially those with major depressive disorder (MDD).
• Identifying neural correlates of NSSI in adolescents with MDD is critical for timely intervention and treatment.
• Distinct electroencephalography (EEG) microstate patterns may differentiate individuals with and without NSSI within this demographic.

Purpose of the Study:

• To investigate unique EEG microstate patterns in adolescents diagnosed with MDD and exhibiting NSSI.
• To explore the potential of EEG microstates as biomarkers for NSSI in adolescents with MDD.
• To correlate EEG microstate parameters with psychological assessments and assess their diagnostic utility.

Main Methods:

• Recruited 134 participants: 44 MDD with NSSI, 41 MDD without NSSI, and 49 healthy controls (HCs).
• Recorded and analyzed electroencephalography (EEG) data using microstate analysis.
• Employed correlation analyses and machine learning algorithms for group classification and pattern identification.

Main Results:

• Microstate B showed significantly higher global explained variance, duration, occurrence, and coverage in HCs compared to both MDD groups.
• Adolescents with MDD and NSSI exhibited a lower transition probability from microstate D to B compared to MDD without NSSI and HCs.
• Machine learning models achieved significant accuracy in classifying groups, with the highest accuracy (77.42%) distinguishing MDD with NSSI from HCs.

Conclusions:

• EEG microstate features show promise as objective biomarkers for identifying NSSI in adolescents with MDD.
• These findings contribute to understanding the neural underpinnings of NSSI in adolescent depression.
• The study highlights the potential of EEG microstates for early detection and targeted interventions in at-risk adolescents.