Task-Free Recovery and Spatial Characterization of a P3-Like Global Network from Resting-State EEG
Diagnosis of mental illness, testing of treatment effects, and design of prevention strategies all require brain-based biomarkers that can serve as effective targets of evaluation. The search for such markers often starts with a search for neural correlates from brain imaging studies with measures of functions and behavior of interest. Yet such an approach can produce erroneous results for correlations do not guarantee causation. Only when the markers map onto neurocomputationally-relevant parameters can such markers best serve the intended function. Here we take an alternative approach to begin with targeting the neuroanatomically and neurophysiologically well-defined neuromoduatory systems that are well positioned to serve the computational role of generating globally synchronized neural activity for the purpose of functional integration [1]. By applying second-order blind identification (SOBI) [2], a blind source separation algorithm (BSS), to five minutes of resting-state EEG data (n=13), we provide evidence to support our conclusion that neuroelectrical signals associated with synchronized global network activity can be extracted using the detailed temporal information in the on-going continuously recorded EEG, instead of event-related potentials (ERPs). We report reliable extraction of a SOBI component, which we refer to as the P3-like component, in every individual studied, replicating our earlier report on data from a single participant [3]. We show that individual differences in the neural networks underlying this P3-like component can be revealed in high dimensional space by a vector of hits-based measures [4] for each of the P3-like networks constituent structures. Given that resting-state EEG can be obtained with greater ease at natural non-hospital settings and at much lower cost in comparison with fMRI, and that mobile EEG systems have become increasingly available, the present work offers an enabling technology to support rapid and low-cost assessment of much larger and diverse populations of individuals, addressing several methodological limitations in our current investigation of brain function. Future opportunities and current limitations will be discussed.