Multidomain Convolution Neural Network Models for Improved Event-Related Potential Classification.

Two convolution neural network (CNN) models are introduced to accurately classify event-related potentials (ERPs) by fusing frequency, time, and spatial domain information acquired from the continuous wavelet transform (CWT) of the ERPs recorded from multiple spatially distributed channels. The multidomain models fuse the multichannel Z-scalograms and the V-scalograms, which are generated from the standard CWT scalogram by zeroing-out and by discarding the inaccurate artifact coefficients that are outside the cone of influence (COI), respectively. In the first multidomain model, the input to the CNN is generated by fusing the Z-scalograms of the multichannel ERPs into a frequency-time-spatial cuboid. The input to the CNN in the second multidomain model is formed by fusing the frequency-time vectors of the V-scalograms of the multichannel ERPs into a frequency-time-spatial matrix. Experiments are designed to demonstrate (a) customized classification of ERPs, where the multidomain models are trained and tested with the ERPs of individual subjects for brain-computer interface (BCI)-type applications, and (b) group-based ERP classification, where the models are trained on the ERPs from a group of subjects and tested on single subjects not included in the training set for applications such as brain disorder classification. Results show that both multidomain models yield high classification accuracies for single trials and small-average ERPs with a small subset of top-ranked channels, and the multidomain fusion models consistently outperform the best unichannel classifiers.

Neural markers of emotion regulation difficulties moderate effects of COVID-19 stressors on adolescent depression.

BACKGROUND: Stressful events, such as those imposed by the COVID-19 pandemic, are associated with depression risk, raising questions about processes that make some people more susceptible to the effects of stress on mental health than others. Emotion regulation may be a key process, but methods for objectively measuring emotion regulation abilities in youth are limited. We leveraged event-related potential (ERP) measures and a longitudinal study of adolescents oversampled for depression and depression risk to examine emotion regulation difficulties as prospective predictors of depressive symptoms in response to pandemic-related stress. METHODS: Before the pandemic, adolescents with (n = 28) and without (n = 34) clinical depression (N = 62 total) completed an explicit emotion regulation task while ERP data were recorded and measures of depressive symptoms. Adolescents were re-contacted during the pandemic to report on COVID-19 related stressful events and depressive symptoms (n = 48). RESULTS: Adolescents who had never experienced a depressive episode showed an increase in depressive symptoms during the pandemic, but adolescents who were clinically depressed before the pandemic did not exhibit significant changes in symptoms. Neural markers of emotion regulation abilities interacted with pandemic-related stressful events to predict depressive symptoms during the pandemic, such that stressors predicted increases in depressive symptoms only for adolescents with greater difficulty modulating responses to negative images before the pandemic. CONCLUSIONS: Results provide insight into adolescent mental health during the COVID-19 pandemic and highlight the role of emotion regulatory brain function in risk and resilience for depression.

A Preliminary Investigation of ERP Components of Attentional Bias in Anxious Adults using Temporospatial Principal Component Analysis.

Threat-related attention bias is thought to contribute to the development and maintenance of anxiety disorders. Dot-probe studies using event-related potentials (ERPs) have indicated that several early ERP components are modulated by threatening and emotional stimuli in anxious populations, suggesting enhanced allocation of attention to threat and emotion at earlier stages of processing. However, ERP components selected for examination and analysis in these studies vary widely and remain inconsistent. The present study used temporospatial principal component analysis (PCA) to systematically identify ERP components elicited to face pair cues and probes in a dot-probe task in anxious adults. Cue-locked components sensitive to emotion included an early occipital C1 component enhanced for happy versus angry face pair cues and an early parieto-occipital P1 component enhanced for happy versus angry face pair cues. Probe-locked components sensitive to congruency included a parieto-occipital P2 component enhanced for incongruent probes (probes replacing neutral faces) versus congruent probes (probes replacing emotional faces). Split-half correlations indicated that the mean value around the PCA-derived peaks were reliably measured in the ERP waveforms. These results highlight promising neurophysiological markers for attentional bias research that can be extended to designs comparing anxious and healthy comparison groups. Results from a secondary exploratory PCA analysis investigating the effects of emotional face position and analyses on behavioral reaction time data are also presented.

Motivational context and neurocomputation of stop expectation moderate early attention responses supporting proactive inhibitory control.

Alterations in attention to cues signaling the need for inhibitory control play a significant role in a wide range of psychopathology. However, the degree to which motivational and attentional factors shape the neurocomputations of proactive inhibitory control remains poorly understood. The present study investigated how variation in monetary incentive valence and stake modulate the neurocomputational signatures of proactive inhibitory control. Adults (N = 46) completed a Stop-Signal Task (SST) with concurrent EEG recording under four conditions associated with stop performance feedback: low and high punishment (following unsuccessful stops) and low and high reward (following successful stops). A Bayesian learning model was used to infer individual's probabilistic expectations of the need to stop on each trial: P(stop). Linear mixed effects models were used to examine whether interactions between motivational valence, stake, and P(stop) parameters predicted P1 and N1 attention-related event-related potentials (ERPs) time-locked to the go-onset stimulus. We found that P1 amplitudes increased at higher levels of P(stop) in punished but not rewarded conditions, although P1 amplitude differences between punished and rewarded blocks were maximal on trials when the need to inhibit was least expected. N1 amplitudes were positively related to P(stop) in the high punishment condition (low N1 amplitude), but negatively related to P(stop) in the high reward condition (high N1 amplitude). Critically, high P(stop)-related N1 amplitude to the go-stimulus predicted behavioral stop success during the high reward block, providing evidence for the role of motivationally relevant context and inhibitory control expectations in modulating the proactive allocation of attentional resources that affect inhibitory control. These findings provide novel insights into the neurocomputational mechanisms underlying proactive inhibitory control under valence-dependent motivational contexts, setting the stage for developing motivation-based interventions that boost inhibitory control.

Exploiting the Cone of Influence for Improving the Performance of Wavelet Transform-Based Models for ERP/EEG Classification.

Features extracted from the wavelet transform coefficient matrix are widely used in the design of machine learning models to classify event-related potential (ERP) and electroencephalography (EEG) signals in a wide range of brain activity research and clinical studies. This novel study is aimed at dramatically improving the performance of such wavelet-based classifiers by exploiting information offered by the cone of influence (COI) of the continuous wavelet transform (CWT). The COI is a boundary that is superimposed on the wavelet scalogram to delineate the coefficients that are accurate from those that are inaccurate due to edge effects. The features derived from the inaccurate coefficients are, therefore, unreliable. In this study, it is hypothesized that the classifier performance would improve if unreliable features, which are outside the COI, are zeroed out, and the performance would improve even further if those features are cropped out completely. The entire, zeroed out, and cropped scalograms are referred to as the "same" (S)-scalogram, "zeroed out" (Z)-scalogram, and the "valid" (V)-scalogram, respectively. The strategy to validate the hypotheses is to formulate three classification approaches in which the feature vectors are extracted from the (a) S-scalogram in the standard manner, (b) Z-scalogram, and (c) V-scalogram. A subsampling strategy is developed to generate small-sample ERP ensembles to enable customized classifier design for single subjects, and a strategy is developed to select a subset of channels from multiple ERP channels. The three scalogram approaches are implemented using support vector machines, random forests, k-nearest neighbor, multilayer perceptron neural networks, and deep learning convolution neural networks. In order to validate the performance hypotheses, experiments are designed to classify the multi-channel ERPs of five subjects engaged in distinguishing between synonymous and non-synonymous word pairs. The results confirm that the classifiers using the Z-scalogram features outperform those using the S-scalogram features, and the classifiers using the V-scalogram features outperform those using the Z-scalogram features. Most importantly, the relative improvement of the V-scalogram classifiers over the standard S-scalogram classifiers is dramatic. Additionally, enabling the design of customized classifiers for individual subjects is an important contribution to ERP/EEG-based studies and diagnoses of patient-specific disorders.

Mindfulness-Based Cognitive Therapy: A Preliminary Examination of the (Event-Related) Potential for Modifying Threat-Related Attentional Bias in Anxiety.

Objectives: Mindfulness-based cognitive therapy (MBCT) can reduce anxiety and depression symptoms in adults with anxiety disorders, and changes in threat-related attentional bias may be a key mechanism driving the intervention's effects on anxiety symptoms. Event-related potentials (ERPs) can illuminate the physiological mechanism through which MBCT targets threat bias and reduces symptoms of anxiety. This preliminary study examined whether P1 ERP threat-related attentional bias markers in anxious adults change from pre- to post-MBCT delivered in-person or virtually (via Zoom) and investigated the relationship between P1 threat-related attentional bias markers and treatment response. Methods: Pre- and post-MBCT, participants with moderate to high levels of anxiety (N = 50) completed a dot-probe task with simultaneous EEG recording. Analyses focused on pre- and post-MBCT P1 amplitudes elicited by angry-neutral and happy-neutral face pair cues, probes, and reaction times in the dot-probe task and anxiety and depression symptoms. Results: Pre- to post-MBCT, there was a significant reduction in P1-Probe amplitudes (d = .23), anxiety (d = .41) and depression (d = .80) symptoms, and reaction times (d = .10). Larger P1-Angry Cue amplitudes, indexing hypervigilance to angry faces, were associated with higher levels of anxiety both pre- and post-MBCT (d = .20). Post-MBCT, anxiety symptoms were lower in the in-person versus virtual group (d = .80). Conclusions: MBCT may increase processing efficiency and decreases anxiety and depression symptoms in anxious adults. However, changes in threat bias specifically were generally not supported. Replication with a comparison group is needed to clarify whether changes were MBCT-specific. Clinical Trials Registration: NCT03571386, June 18, 2018. Supplementary Information: The online version contains supplementary material available at 10.1007/s12671-022-01910-x.

CINET: A Brain-Inspired Deep Learning Context-Integrating Neural Network Model for Resolving Ambiguous Stimuli.

The brain uses contextual information to uniquely resolve the interpretation of ambiguous stimuli. This paper introduces a deep learning neural network classification model that emulates this ability by integrating weighted bidirectional context into the classification process. The model, referred to as the CINET, is implemented using a convolution neural network (CNN), which is shown to be ideal for combining target and context stimuli and for extracting coupled target-context features. The CINET parameters can be manipulated to simulate congruent and incongruent context environments and to manipulate target-context stimuli relationships. The formulation of the CINET is quite general; consequently, it is not restricted to stimuli in any particular sensory modality nor to the dimensionality of the stimuli. A broad range of experiments is designed to demonstrate the effectiveness of the CINET in resolving ambiguous visual stimuli and in improving the classification of non-ambiguous visual stimuli in various contextual environments. The fact that the performance improves through the inclusion of context can be exploited to design robust brain-inspired machine learning algorithms. It is interesting to note that the CINET is a classification model that is inspired by a combination of brain's ability to integrate contextual information and the CNN, which is inspired by the hierarchical processing of information in the visual cortex.

Multimodal Object Classification Models Inspired by Multisensory Integration in the Brain.

Two multimodal classification models aimed at enhancing object classification through the integration of semantically congruent unimodal stimuli are introduced. The feature-integrating model, inspired by multisensory integration in the subcortical superior colliculus, combines unimodal features which are subsequently classified by a multimodal classifier. The decision-integrating model, inspired by integration in primary cortical areas, classifies unimodal stimuli independently using unimodal classifiers and classifies the combined decisions using a multimodal classifier. The multimodal classifier models are implemented using multilayer perceptrons and multivariate statistical classifiers. Experiments involving the classification of noisy and attenuated auditory and visual representations of ten digits are designed to demonstrate the properties of the multimodal classifiers and to compare the performances of multimodal and unimodal classifiers. The experimental results show that the multimodal classification systems exhibit an important aspect of the "inverse effectiveness principle" by yielding significantly higher classification accuracies when compared with those of the unimodal classifiers. Furthermore, the flexibility offered by the generalized models enables the simulations and evaluations of various combinations of multimodal stimuli and classifiers under varying uncertainty conditions.

The neural chronometry of threat-related attentional bias: Event-related potential (ERP) evidence for early and late stages of selective attentional processing.

Rapid and accurate detection of threat is adaptive. Yet, threat-related attentional biases, including hypervigilance, avoidance, and attentional disengagement delays, may contribute to the etiology and maintenance of anxiety disorders. Behavioral measures of attentional bias generally indicate that threat demands more attentional resources; however, indices exploring differential allocation of attention using reaction time fail to clarify the time course by which attention is deployed under threatening circumstances in healthy and anxious populations. In this review, we conduct an interpretive synthesis of 28 attentional bias studies focusing on event-related potentials (ERPs) as a primary outcome to inform an ERP model of the neural chronometry of attentional bias in healthy and anxious populations. The model posits that both healthy and anxious populations display modulations of early ERP components, including the P1, N170, P2, and N2pc, in response to threatening and emotional stimuli, suggesting that both typical and abnormal patterns of attentional bias are characterized by enhanced allocation of attention to threat and emotion at earlier stages of processing. Compared to anxious populations, healthy populations more clearly demonstrate modulations of later components, such as the P3, indexing conscious and evaluative processing of threat and emotion and disengagement difficulties at later stages of processing. Findings from the interpretive synthesis, existing bias models, and extant neural literature on attentional systems are then integrated to inform a conceptual model of the processes and substrates underlying threat appraisal and resource allocation in healthy and anxious populations. To conclude, we discuss therapeutic interventions for attentional bias and future directions.

Measuring cognitive outcomes in mindfulness-based intervention research: a reflection on confounding factors and methodological limitations.

One potential pathway by which mindfulness-based meditation improves health outcomes is through changes in cognitive functioning. Here, we summarize and comment upon three systematic reviews conducted over the last seven years that have had the goal of identifying the impact of mindfulness on cognitive outcomes. In our analysis, we identified a number of methodological limitations and potential confounding factors that interfere with and limit our ability to interpret the results. In order to gain a granular view of the relationship between mindfulness training and cognition, we report on the following: 1) What do we know? How does mindfulness affect cognition? 2) variable criteria that define an MBI; 2) limitations of assays used to measure cognition; and 3) methodological quality of an MBI trial and reporting of findings. Finally, we offer constructive means for interpretation and recommendations for moving the field of mindfulness research forward regarding effects on cognition.

Adult Age Differences in the Use of Conceptual Combination as an Associative Encoding Strategy.

It is well-established that aging impairs memory for associations more than it does memory for single items. Aging also impacts processes involved in online language comprehension, including the ability to form integrated, message-level representations. These changes in comprehension processes could impact older adults' associative memory performance, perhaps by reducing or altering the effectiveness of encoding strategies that encourage semantic integration. The present study examined age differences in the use of a strategy termed conceptual combination, which involves integrating two words (e.g., "winter" and "salad") into a single concept ("a salad for winter"). We recorded ERPs while participants studied unrelated noun pairs using a strategy that either did or did not encourage conceptual combination. We also varied the concreteness of the first noun in each pair in order to measure compositional concreteness effects, or ERP differences at the second noun due to the concreteness of the first noun. At the first nouns, older adults showed word-level concreteness effects that were similar to those of younger adults. However, compositional concreteness effects were diminished in older adults, consistent with reduced semantic integration. Older adults' associative memory performance was better for word pairs studied during the conceptual combination task versus the non-combinatory encoding task; however, the magnitude of the age-related associative memory deficit did not differ between tasks. Finally, analyses of both memory accuracy and trial-by-trial ratings of perceived combination success suggested that older adults had disproportionate difficulty applying the conceptual combination strategy to word pairs that began with abstract nouns. Overall, these results indicate that changes to integrative language processing that occur with age are not independent of - and may sometimes exacerbate - age-related memory decline.