Multivariate machine learning distinguishes cross-network dynamic functional connectivity patterns in state and trait...

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Psychiatry

Background:

• Chronic pain involves dynamic fluctuations in intensity.
• Brain communication is also dynamic, but how it's disrupted in chronic pain is unclear.
• Understanding brain dynamics across different pain timescales (state vs. trait) is crucial.

Purpose of the Study:

• To investigate the relationship between dynamic and static brain communication patterns and chronic pain characteristics.
• To differentiate how brain dynamics relate to short-term pain states versus long-term pain traits.
• To explore these relationships in patients with chronic neuropathic pain (NP) and non-NP.

Main Methods:

• Utilized network-based multivariate machine learning models.
• Analyzed resting-state dynamic functional connectivity (dFC) and static functional connectivity.
• Focused on default mode, salience, and executive control networks.

Main Results:

• Dynamic functional connectivity (dFC) measures, particularly cross-network interactions, were more strongly associated with chronic pain than static functional connectivity.
• These associations were more prominent for trait pain (long-term condition) than state pain (day-to-day variation).
• Specific dFC patterns with executive control and limbic networks correlated with pain severity in neuropathic pain patients.

Conclusions:

• Dynamic functional connectivity (dFC) captures the enduring nature of chronic pain (trait pain), not just transient pain states.
• Altered dFC is a significant biomarker in chronic neuropathic pain, reflecting complexities beyond mere pain intensity.
• Brain network dynamics offer a novel perspective on the pathophysiology of chronic pain.