Connectome-based Models Predict Separable Components of Attention in Novel Individuals

• DOI: 10.1162/jocn_a_01197
• Publication Date: 2017-10-17T18:44:35Z
• Authors: Monica D. Rosenberg; Wei-Ting Hsu; Dustin Scheinost; R. Todd Constable; Marvin M. Chun
• Keywords: Human Connectome Project; Alertness; Attention network; Stimulus (psychology); Vigilance (psychology)

Although we typically talk about attention as a single process, it comprises multiple independent components. But what are these components, and how they represented in the functional organization of brain? To investigate whether long-studied components reflected brain's intrinsic organization, here apply connectome-based predictive modeling (CPM) to predict Posner Petersen's influential model attention: alerting (preparing maintaining alertness vigilance), orienting (directing stimulus), executive control (detecting resolving cognitive conflict) [Posner, M. I., & Petersen, S. E. The system human brain. Annual Review Neuroscience, 13, 25-42, 1990]. Participants performed Attention Network Task (ANT), which measures three factors, rested during fMRI scanning. CPMs tested with leave-one-subject-out cross-validation successfully predicted novel individual's overall ANT accuracy, RT variability, scores from connectivity observed performance. also generalized participants' their resting-state alone, demonstrating that patterns absence an explicit task contain signature ability prepare for upcoming stimulus. Suggesting significant variance performance is explained by sustained factor, CPM, defined prior work attentional abilities, task-based data variability data. Our results suggest that, whereas may be closely related attention, infrastructure supports distinct can measured at rest. In future, CPM applied elucidate additional relationships between brain networks them.