Uncovering individual differences in fine-scale dynamics of functional connectivity

Area of Science:

• Neuroscience
• Brain Imaging
• Computational Neuroscience

Background:

• Functional connectivity (FC) profiles offer subject-specific brain features conserved over time, crucial for understanding brain-behavior relationships.
• Prior research predominantly analyzed spatial aspects of FC fingerprints from entire imaging sessions.
• Identifying dynamic, time-varying aspects of FC is essential for a comprehensive understanding of brain function.

Purpose of the Study:

• To develop and validate a method for temporally filtering FC data.
• To identify specific moments in time that enhance the distinctiveness of FC fingerprints.
• To explore how temporal dynamics contribute to subject identifiability in brain imaging.

Main Methods:

• Utilized a decomposition of FC into edge time series (eTS).
• Systematically analyzed functional magnetic resonance imaging (fMRI) frames to define features enhancing identifiability.
• Employed multiple fingerprinting and similarity metrics across diverse datasets.
• Optimized features data-drivenly to maximize fingerprinting metrics.

Main Results:

• Temporally filtered FC features enhance identifiability across various metrics and datasets.
• Key metrics vary characteristically with eTS cofluctuation amplitude, frame similarity, transition velocity, and functional system expression.
• Data-driven feature optimization isolated distinct spatial patterns of system expression at specific time points.
• Utilizing only 10% of the data yielded stronger fingerprints compared to using the full dataset.

Conclusions:

• FC fingerprints are differentially expressed across time, indicating dynamic brain activity.
• Considering both spatial and temporal characteristics simultaneously allows for the identification of multiple distinct FC fingerprints.
• This temporal filtering approach offers a more sensitive method for capturing individual brain characteristics.