Brain network informed subject community detection in early-onset schizophrenia

Area of Science:

• Neuroscience
• Psychiatry
• Computational Biology

Background:

• Early-onset schizophrenia (EOS) presents a critical window for understanding schizophrenia's pathophysiology.
• Existing classification systems may not fully capture underlying brain dysfunctions in EOS.

Purpose of the Study:

• To investigate intrinsic connectivity network (ICN) deficits in drug-naïve, first-episode EOS patients.
• To apply a data-driven approach to identify homogeneous subject communities based on ICN characteristics.
• To associate these communities and ICNs with clinical diagnosis and symptom patterns.

Main Methods:

• Examined ICNs in 26 drug-naïve, first-episode EOS patients and 25 matched controls.
• Utilized a fully data-driven approach to cluster subjects into communities based on ICNs.
• Associated identified communities and ICNs with clinical diagnosis and symptom patterns.
• Performed post-hoc functional connectivity modeling.

Main Results:

• A default mode ICN was statistically absent in EOS patients.
• A frontotemporal ICN differentiated EOS patients with predominantly negative symptoms.
• EOS patients with positive symptoms showed ICN patterns similar to controls.
• Frontotemporal circuit connectivity was modulated by positive and negative symptom severity.

Conclusions:

• This study introduces a novel subtype discovery approach for EOS based on brain networks.
• Complex relationships exist between brain networks and symptom patterns in EOS.
• Findings suggest distinct neurobiological underpinnings for different symptom profiles in early-onset schizophrenia.