Heterogenous brain activations across individuals localize to a common network

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Neuroimaging

Background:

• Task functional magnetic resonance imaging (fMRI) studies often face challenges with low reproducibility across individuals.
• Heterogeneity in brain activations is a key factor contributing to this reproducibility issue.

Purpose of the Study:

• To investigate if heterogeneous brain activations across individuals localize to a common, reproducible network.
• To explore the potential of Activation Network Mapping (ANM) in identifying individual cognitive networks and predicting behavior.

Main Methods:

• Assessed the reproducibility of discrete brain activation during working memory (WM) tasks across individuals.
• Applied the Activation Network Mapping (ANM) technique to identify individual WM brain networks.
• Utilized machine learning algorithms for prediction analyses based on identified WM networks.

Main Results:

• Discrete brain activation reproducibility during WM tasks was found to be low across individuals.
• Network-based reproducibility of WM using ANM was significantly higher.
• Individual WM networks identified via ANM successfully predicted WM behavioral performance, outperforming discrete brain activations.

Conclusions:

• Individual brain networks demonstrate higher reproducibility than discrete activations, offering a novel explanation for fMRI reproducibility challenges.
• Activation Network Mapping (ANM) is a promising technique for identifying individual cognitive networks.
• ANM has broad potential applications in understanding individual cognitive processes and predicting behavior.