Considering factors affecting the connectome-based identification process: Comment on Waller et al.

Area of Science:

• Neuroscience
• Psychiatry
• Data Science

Background:

• Functional connectivity data is increasingly used to identify individuals within groups.
• Previous studies suggested low-resolution neuroimaging might hinder individual identification.
• The generalizability of these findings to precision psychiatry applications is debated.

Purpose of the Study:

• To re-evaluate the factors influencing individual identification using functional connectivity data.
• To determine if spatial and temporal resolution are primary drivers of identification accuracy.
• To investigate the impact of data quantity and in-scanner motion on individual identification.

Main Methods:

• Analysis of Human Connectome Project (HCP) dataset to assess identification rates based on data amount and motion.
• Utilized Consortium for Reliability and Reproducibility (CoRR) dataset to compare multiband (high resolution) and non-multiband (low resolution) imaging parameters.
• Statistical analysis to correlate identification rates with scan length, motion, and imaging acquisition type.

Main Results:

• Scan length and in-scanner motion significantly impact individual identification rates.
• Spatiotemporal resolution of neuroimaging acquisition (e.g., multiband vs. non-multiband) did not affect identification accuracy.
• Individual differences in brain connectomes are observable when controlling for motion and data quantity.

Conclusions:

• Data quantity and motion are the predominant factors influencing the reliability of individual identification from functional connectivity.
• Connectome-based identification is feasible and potentially valuable for precision psychiatry, contrary to previous suggestions.
• Future research should prioritize optimizing data acquisition parameters and motion correction techniques.