Moving Beyond ERP Components: A Selective Review of Approaches to Integrate EEG and Behavior

Abstract

Relationships between neuroimaging measures and behavior provide important clues about brain function and cognition in healthy and clinical populations. While electroencephalography (EEG) provides a portable, low cost measure of brain dynamics, it has been somewhat underrepresented in the emerging field of model-based inference. We seek to address this gap in this article by highlighting the utility of linking EEG and behavior, with an emphasis on approaches for EEG analysis that move beyond focusing on peaks or “components” derived from averaging EEG responses across trials and subjects (generating the event-related potential, ERP). First, we review methods for deriving features from EEG in order to enhance the signal within single-trials. These methods include filtering based on user-defined features (i.e., frequency decomposition, time-frequency decomposition), filtering based on data-driven properties (i.e., blind source separation, BSS), and generating more abstract representations of data (e.g., using deep learning). We then review cognitive models which extract latent variables from experimental tasks, including the drift diffusion model (DDM) and reinforcement learning (RL) approaches. Next, we discuss ways to access associations among these measures, including statistical models, data-driven joint models and cognitive joint modeling using hierarchical Bayesian models (HBMs). We think that these methodological tools are likely to contribute to theoretical advancements, and will help inform our understandings of brain dynamics that contribute to moment-to-moment cognitive function.

Applying deep learning to single-trial EEG data provides evidence for complementary theories on action control

“…Moreover, if such COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0846-z ARTICLE COMMUNICATIONS BIOLOGY | (2020) 3:112 | https://doi.org/10.1038/s42003-020-0846-z | www.nature.com/commsbio classifications are possible using single-trial EEG data, this will represent an important step towards going beyond conventional ERP components and to functionally relate EEG features to behavioral performance. Most noteworthy, this would be done at the time scale of single trials, i.e., the neurophysiological processes that are directly associated with a given single response 31 .…”

“…The interrelation between behavior and associated neurophysiological dynamics is hardly strictly linear, albeit most analysis approaches in cognitive neuroscience rely on the assumption of linearity when applying (correlational) approaches to connect behavioral data and neurophysiological data. Generally, there is rarely a one-to-one relationship between EEG-derived neural signals and behavior 31 , although this is often suggested, or at least implied. Furthermore, the neurophysiologic data used for the formation of ERPs is inherently noisy.…”

“…Therefore, it is substantially harder to establish such functional connections at the single-subject level or the single-trial level [31][32][33] . This problem severely limits the degree and level at which neural signatures may be functionally related to human behavior 31 , or indicate cognitive processes involved in a specific situation (e.g., conflict processing). These shortcomings may be tackled with machine learning methods 31 .…”

“…This problem severely limits the degree and level at which neural signatures may be functionally related to human behavior 31 , or indicate cognitive processes involved in a specific situation (e.g., conflict processing). These shortcomings may be tackled with machine learning methods 31 . There are already first encouraging approaches, as more conventional machine learning approaches like support vector machines (SVMs) have been successfully applied in comparable contexts [34][35][36][37][38] .…”

Transcranial Near-Infrared Light: Dose-Dependent Effects on EEG Oscillations but not Cerebral Blood Flow

“…Electroencephalography (EEG) is an imaging technique that can track electrophysiological activity of the brain with a high temporal resolution. Previously, EEG has been used to evaluate neurophysiological brain activity both during resting 'task free' scans -to study patterns of spontaneous dynamics of brain activity (Rosazza & Minati, 2011) and during task performance to examine how brain activity is modulated by the demands of a task (Bridwell et al, 2018). Fluctuations in the EEG signal have been studied in several frequency bands, including delta (0.5-3.5 Hz), theta (4-8 Hz), alpha (8-12 Hz), beta (12-30 Hz) and gamma (>30 Hz).…”