PRESERVING HUMAN LARGE-SCALE BRAIN CONNECTIVITY FINGERPRINT IDENTIFIABILITY WITH RANDOM PROJECTIONS

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Personalized medicine faces challenges due to the complex nature of neurodegenerative and psychiatric disorders.
• Functional magnetic resonance imaging (fMRI) advances allow precise mapping of brain connectivity, forming the basis of brain connectomics.
• The clinical utility of functional connectomes is hindered by challenges, including computational demands.

Purpose of the Study:

• To address the computational challenges of functional connectomes for improved clinical utility.
• To propose a novel method for efficient functional connectome analysis.
• To maintain subject identifiability while reducing computational load.

Main Methods:

• A random projection method was developed to sample and retain a proportion of functional edges.
• The method aims to preserve subject identifiability within the functional connectome.
• Evaluation of the method's impact on biomarker integrity from whole-brain connectivity data.

Main Results:

• The random projection method significantly improves the computational efficiency of functional connectomes.
• Subject identifiability is largely preserved using the proposed sampling technique.
• The method demonstrates potential for enhancing the clinical applicability of neuroimaging biomarkers.

Conclusions:

• Random projection offers a computationally efficient approach to functional connectomics.
• This method facilitates the clinical utility of brain connectome data in personalized medicine.
• The approach maintains biomarker integrity, paving the way for more accessible neuroimaging analyses.