Relatively newer functional connectome-based predictive modeling approaches have made some progress in generating insights at the single individual level ( Bzdok et al., 2020 Shen et al., 2017), but several methodological issues need to be resolved before their clinical application becomes a reality ( Dadi et al., 2019). One of the main reasons for the lack of fMRI-based clinical translation is that the traditional neuroimaging analyses (e.g., GLM or functional connectivity) tend to measure group-averaged (or central) tendencies, largely due to the low signal-to-noise ratio of the blood oxygenation level–dependent (BOLD) signal ( Welvaert & Rosseel, 2013, 2014). However, unlike structural imaging, which has become standard in clinical practice, the clinical use of functional imaging (e.g., fMRI) has been limited to presurgical planning and functional mapping ( Mitchell et al., 2013 Tie et al., 2014).
Modern noninvasive brain imaging technologies such as structural and functional magnetic resonance imaging promise to not only provide a better understanding of the neural basis of behavior but also to fundamentally transform how we diagnose and treat mental health disorders ( Saggar & Uddin, 2019). Looking forward, we hope our framework will help researchers push the boundaries of psychiatric neuroimaging toward generating insights at the single-participant level across consortium-size datasets. Our new NeuMapper framework was developed and validated using multiple fMRI datasets where participants engaged in continuous multitask experiments that mimic “ongoing” cognition. We also introduce new meta-analytic approaches to better anchor Mapper-generated representations to neuroanatomy and behavior. Here, we present a novel computational framework for Mapper-designed specifically for neuroimaging data-that removes limitations and reduces computational costs associated with dimensionality reduction and parameter exploration. Despite promising results, initial applications of Mapper to neuroimaging data were constrained by (1) the need for dimensionality reduction and (2) lack of a biologically grounded heuristic for efficiently exploring the vast parameter space.
Recently, the Mapper approach from topological data analysis (TDA) has been successfully applied on noninvasive human neuroimaging data to characterize the entire dynamical landscape of whole-brain configurations at the individual level without requiring any spatiotemporal averaging at the outset.
For better translational outcomes, researchers and clinicians alike demand novel tools to distill complex neuroimaging data into simple yet behaviorally relevant representations at the single-participant level.