Avoiding illusory effects in representational similarity analysis: What (not) to do with the diagonal

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Representational Similarity Analysis (RSA) is a key method in neuroimaging for comparing neural data, models, and behavior.
• RSA relies on Representational Dissimilarity Matrices (RDMs) to standardize comparisons across diverse data types.
• RDMs offer a flexible framework for understanding information processing in the brain.

Discussion:

• This study addresses two critical issues in RDM construction and comparison.
• The diagonal values of RDMs, representing within-condition reliability, are often not estimated, hindering overall RDM reliability assessment.
• Excluding diagonal values during RDM comparison is crucial to prevent spurious correlations, as demonstrated by simulation.

Key Insights:

• Calculating diagonal values is essential for a reliable RDM as a whole.
• Improper inclusion of diagonal values can lead to illusory correlations between unrelated datasets.
• A re-evaluation of existing studies, such as Coggan et al. (2016), suggests potential inflation of results due to diagonal value inclusion.

Outlook:

• Future RSA studies should prioritize the accurate estimation and appropriate handling of RDM diagonal values.
• Standardizing RDM construction and comparison protocols will enhance the robustness of neuroimaging findings.
• Further research can explore the precise impact of diagonal values on various RSA applications and interpretations.