Masked features of task states found in individual brain networks

Abstract

AbstractCompleting complex tasks requires flexible integration of functions across brain regions. While studies have shown that functional networks are altered across tasks, recent work highlights that brain networks exhibit substantial individual differences. Here we asked whether individual differences are important for predicting brain network interactions across cognitive states. We trained classifiers to decode state using data from single person “precision” fMRI datasets across 5 diverse cognitive states. Classifiers were then tested on either independent sessions from the same person or new individuals. Classifiers were able to decode task states in both the same and new participants above chance. However, classification performance was significantly higher within a person, a pattern consistent across model types, datasets, tasks, and feature subsets. This suggests that individualized approaches can uncover robust features of brain states, including features obscured in cross-subject analyses. Individualized approaches have potential to deepen our understanding of brain interactions during complex cognition.Citation Diversity StatementRecently, the field of neuroscience has reported a bias in citation practices such that papers from minority groups are more often under-cited relative to the number of papers in the field1. The authors of this paper were proactive in consideration of selecting references that reflect diversity of the field in thought, contribution, and gender. Utilizing previously derived databases 1,2 we obtained the predicted gender of authors referenced in this manuscript. By this measure (and excluding self-citations to the first and last authors of our current paper), our references contain 13.87% woman(first)/woman(last), 23.3% man/woman, 23.3% woman/man, and 39.53% man/man. This method is limited in that a) names, pronouns, and social media profiles used to construct the databases may not, in every case, be indicative of gender identity and b) it cannot account for intersex, non-binary, or transgender people. Second, we obtained the predicted racial/ethnic category of the first and last author of each reference by databases that store the probability of a first and last name being carried by an author of color3. By this measure (and excluding self-citations), our references contain 10.83% author of color (first)/author of color(last), 10.64% white author/author of color, 23.55% author of color/white author, and 54.98% white author/white author. This method is limited in that a) names and Florida Voter Data to make the predictions may not be indicative of racial/ethnic identity, and b) it cannot account for Indigenous and mixed-race authors, or those who may face differential biases due to the ambiguous racialization or ethnicization of their names. We look forward to future work that could help us to better understand how to support equitable practices in science.