scholarly journals Convolutional Neural Networks for Decoding of Covert Attention Focus and Saliency Maps for EEG Feature Visualization

2019 ◽  
Author(s):  
Amr Farahat ◽  
Christoph Reichert ◽  
Catherine M. Sweeney-Reed ◽  
Hermann Hinrichs
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Cognitive Task ◽  
Event Related Potentials ◽  
Eeg Signal ◽  
Linear Discriminant ◽  
Saliency Maps ◽  
Significant Difference ◽  
Related Potentials ◽  
The Impact

ABSTRACTObjectiveConvolutional neural networks (CNNs) have proven successful as function approximators and have therefore been used for classification problems including electroencephalography (EEG) signal decoding for brain-computer interfaces (BCI). Artificial neural networks, however, are considered black boxes, because they usually have thousands of parameters, making interpretation of their internal processes challenging. Here we systematically evaluate the use of CNNs for EEG signal decoding and investigate a method for visualizing the CNN model decision process.ApproachWe developed a CNN model to decode the covert focus of attention from EEG event-related potentials during object selection. We compared the CNN and the commonly used linear discriminant analysis (LDA) classifier performance, applied to datasets with different dimensionality, and analyzed transfer learning capacity. Moreover, we validated the impact of single model components by systematically altering the model. Furthermore, we investigated the use of saliency maps as a tool for visualizing the spatial and temporal features driving the model output.Main resultsThe CNN model and the LDA classifier achieved comparable accuracy on the lower-dimensional dataset, but CNN exceeded LDA performance significantly on the higher-dimensional dataset (without hypothesis-driven preprocessing), achieving an average decoding accuracy of 90.7% (chance level = 8.3%). Parallel convolutions, tanh or ELU activation functions, and dropout regularization proved valuable for model performance, whereas the sequential convolutions, ReLU activation function, and batch normalization components, reduced accuracy or yielded no significant difference. Saliency maps revealed meaningful features, displaying the typical spatial distribution and latency of the P300 component expected during this task.SignificanceFollowing systematic evaluation, we provide recommendations for when and how to use CNN models in EEG decoding. Moreover, we propose a new approach for investigating the neural correlates of a cognitive task by training CNN models on raw high-dimensional EEG data and utilizing saliency maps for relevant feature extraction.

Neuroelectric Correlates of Hemispheric Asymmetry: Spectral Discrimination and Stimulus Competition

2003 ◽  
Vol 14 (08) ◽  
pp. 434-443
Author(s):  
Ralf R. Greenwald ◽  
James Jerger
Keyword(s):  
Auditory Processing ◽  
Hemispheric Asymmetry ◽  
Right Hemisphere ◽  
Cognitive Task ◽  
Hemispheric Specialization ◽  
Event Related Potentials ◽  
Task Demands ◽  
Significant Difference ◽  
Related Potentials ◽  
The Right

In an effort to explore further the role of the right hemisphere in auditory processing, this study utilized brain event-related potentials (ERPs) to investigate hemispheric asymmetry for the processing of complex spectral tones. Subjects participated in two pitch discrimination tasks, one diotic, the other dichotic. ERP components were recorded from 28 electrodes on the scalp and analyzed via individual/group average area measurements. Results showed that ERPs recorded in response to the dichotic target pairs exhibited a larger P3 area when the target tone was presented to the left ear, while the N1 area showed no significant difference. ERPs recorded in the diotic condition showed a larger P3 area and smaller N1 area compared to the dichotic conditions. Finally, all experimental tasks showed that topographic hemispheric activation patterns were asymmetric to the right hemisphere. Findings support the notion that ERP topographic asymmetries may be dependent on specific cognitive task demands (e.g., diotic vs. dichotic modes of presentation). In addition, the data suggest that the P3 component may better reflect interaural advantages for complex tones than the N1 component and may, therefore, be a more sensitive indicator of hemispheric specialization.

scholarly journals Expectancy of Stress-Reducing Aromatherapy Effect and Performance on a Stress-Sensitive Cognitive Task

2015 ◽  
Vol 2015 ◽  
pp. 1-10 ◽  
Author(s):  
Irina Chamine ◽  
Barry S. Oken
Keyword(s):  
Task Performance ◽  
Stress Reduction ◽  
Cognitive Task ◽  
Reaction Times ◽  
Event Related Potentials ◽  
Specific Effects ◽  
N200 Amplitude ◽  
Related Potentials ◽  
And Performance

Objective. Stress-reducing therapies help maintain cognitive performance during stress. Aromatherapy is popular for stress reduction, but its effectiveness and mechanism are unclear. This study examined stress-reducing effects of aromatherapy on cognitive function using the go/no-go (GNG) task performance and event related potentials (ERP) components sensitive to stress. The study also assessed the importance of expectancy in aromatherapy actions.Methods. 81 adults were randomized to 3 aroma groups (active experimental, detectable, and undetectable placebo) and 2 prime subgroups (prime suggesting stress-reducing aroma effects or no-prime). GNG performance, ERPs, subjective expected aroma effects, and stress ratings were assessed at baseline and poststress.Results. No specific aroma effects on stress or cognition were observed. However, regardless of experienced aroma, people receiving a prime displayed faster poststress median reaction times than those receiving no prime. A significant interaction for N200 amplitude indicated divergent ERP patterns between baseline and poststress for go and no-go stimuli depending on the prime subgroup. Furthermore, trends for beneficial prime effects were shown on poststress no-go N200/P300 latencies and N200 amplitude.Conclusion. While there were no aroma-specific effects on stress or cognition, these results highlight the role of expectancy for poststress response inhibition and attention.

ASSESSING THE IMPACT OF CONGRUENT AND INCONGRUENT STIMULUS IN STROOP TASK, USING EVENT-RELATED POTENTIALS (ERP) IN PATIENTS WITH DEPRESSION

2018 ◽  
Vol 30 (05) ◽  
pp. 1850034
Author(s):  
Yeganeh Shahsavar ◽  
Majid Ghoshuni
Keyword(s):  
Stroop Task ◽  
Time Window ◽  
Event Related Potentials ◽  
Peak Time ◽  
Eeg Signals ◽  
Incongruent Stimulus ◽  
Related Potentials ◽  
Recording Channel ◽  
The Mean ◽  
The Impact

The main goal of this event-related potentials (ERPs) study was to assess the effects of stimulations in Stroop task in brain activities of patients with different degrees of depression. Eighteen patients (10 males, with the mean age [Formula: see text]) were asked to fill out Beck’s depression questionnaire. Electroencephalographic (EEG) signals of subjects were recorded in three channels (Pz, Cz, and Fz) during Stroop test. This test entailed 360 stimulations, which included 120 congruent, 120 incongruent, and 120 neutral stimulations. To analyze the data, 18 time features in each type of stimulus were extracted from the ERP components and the optimal features were selected. The correlation between the subjects’ scores in Beck’s depression questionnaires and the extracted time features in each recording channel was calculated in order to select the best features. Total area, and peak-to-peak time window in the Cz channel in both the congruent and incongruent stimulus showed significant correlation with Beck scores, with [Formula: see text], [Formula: see text] and [Formula: see text], [Formula: see text], respectively. Consequently, given the correlation between time features and the subjects’ Beck scores with different degrees of depression, it can be interpreted that in case of growth in degrees of depression, stimulations involving congruent images would produce more challenging interferences for the patients compared to incongruent stimulations which can be more effective in diagnosing the level of disorder.

scholarly journals The Role of Leadership Level in College Students’ Decision-Making: Evidence From Event-Related Potential Analysis

2021 ◽  
Vol 12 ◽  
Author(s):  
Yuwei Yang ◽  
Shunshun Du ◽  
Hui He ◽  
Chengming Wang ◽  
Xueke Shan ◽  
...  
Keyword(s):  
College Students ◽  
Decision Making ◽  
Event Related Potentials ◽  
Event Related Potential ◽  
Economic Decision Making ◽  
Significant Difference ◽  
Leadership Group ◽  
Related Potentials ◽  
Acceptance Rates ◽  
Economic Transaction

Although risk decision-making plays an important role in leadership practice, the distinction in behavior between humans with differing levels of leadership, as well as the underlying neurocognitive mechanisms involved, remain unclear. In this study, the Ultimatum Game (UG) was utilized in concert with electroencephalograms (EEG) to investigate the temporal course of cognitive and emotional processes involved in economic decision-making between high and low leadership level college students. Behavioral results from this study found that the acceptance rates in an economic transaction, when the partner was a computer under unfair/sub unfair condition, were significantly higher than in transactions with real human partners for the low leadership group, while there was no significant difference in acceptance rates for the high leadership group. Results from Event-Related Potentials (ERP) analysis further indicated that there was a larger P3 amplitude in the low leadership group than in the high leadership group. We concluded that the difference between high and low leadership groups was at least partly due to their different emotional management abilities.

scholarly journals Expect the unexpected: Processing of sensory information in the anticipation of predictable and unpredictable threats

2019 ◽  
Vol 11 (1) ◽  
pp. 80-115
Author(s):  
Eva Koderman
Keyword(s):  
Sensory Information ◽  
Event Related Potentials ◽  
Perceived Threat ◽  
Subject Design ◽  
Unpleasant Picture ◽  
Before And After ◽  
Attentional Engagement ◽  
Related Potentials ◽  
The Impact ◽  
Threat Of Shock

Abstract Anxiety is characterized by a sustained state of heightened vigilance due to uncertain danger, producing increased attention to a perceived threat in one's environment. To further examine this exploited the temporal resolution afforded by event-related potentials to investigate the impact of predictability of threat on early perceptual activity. We recruited 28 participants and utilized a within-subject design to examine hypervigilance in anticipation of shock, unpleasant picture and unpleasant sound during a task with unpredictable, predictable and no threat. We investigated if habituation to stimuli was present by asking the participants to rate unpleasantness and intensity of the stimuli before and after the experiment. We observed hypervigilance in the unpredictable threat of shock. Habituation was observed for the visual stimuli. The present study suggests that unpredictability enhances attentional engagement with neutral somatosensory stimuli when the threat is of the same modality, meaning we observed the presence of hypervigilance which is a characteristic of anxiety.

scholarly journals Error-Monitoring is Modulated by School Pedagogy

2019 ◽  
Author(s):  
Solange Denervaud ◽  
Jean-François Knebel ◽  
Emeline Mullier ◽  
Patric Hagmann ◽  
Micah M. Murray
Keyword(s):  
Fluid Intelligence ◽  
Recognition Task ◽  
Event Related Potentials ◽  
Anterior Cingulate ◽  
Error Monitoring ◽  
Traditional Curriculum ◽  
Source Estimation ◽  
Related Potentials ◽  
School Pedagogy ◽  
The Impact

Within an inherently dynamic environment, unexpected outcomes are part of daily life. Performance monitoring allows us to detect these events and adjust behavior accordingly. The necessity of such an optimal functioning has made error-monitoring a prominent topic of research over the last decades. Event-related potentials (ERPs) have differentiated between two brain components involved in error-monitoring: the error-related negativity (ERN) and error-related positivity (Pe) that are thought to reflect detection vs. emotional/motivational processing of errors, respectively. Both ERN and Pe depend on the protracted maturation of the frontal cortices and anterior cingulate through adolescence. To our knowledge, the impact of schooling pedagogy on error-monitoring and its brain mechanisms remains unknown and was the focus of the present study. Swiss schoolchildren completed a continuous recognition task while 64-channel EEG was recorded and later analyzed within an electrical neuroimaging framework. They were enrolled either in a Montessori curriculum (N=13), consisting of self-directed learning through trial-and-error activities with sensory materials, or a traditional curriculum (N=14), focused on externally driven activities mainly based on reward feedback. The two groups were controlled for age, gender, socio-economic status, parental educational style, and scores of fluid intelligence. The ERN was significantly enhanced in Montessori schoolchildren (driven by a larger response to errors), with source estimation differences localized to the cuneus and precuneus. In contrast, the Pe was enhanced in traditional schoolchildren (driven by a larger response to correct trials), with source estimation differences localized to the ventral anterior cingulate. Receiver operating characteristic (ROC) analysis demonstrated that the ERN and Pe could reliably classify if a child was following a Montessori or traditional curriculum. Brain activity subserving error-monitoring is modulated differently according to school pedagogy.

scholarly journals Modeling Nonlinear Transfer Functions from Speech Envelopes to Encephalography with Neural Networks

2019 ◽  
Vol 11 (4) ◽  
pp. 1 ◽  
Author(s):  
Tobias de Taillez ◽  
Florian Denk ◽  
Bojana Mirkovic ◽  
Birger Kollmeier ◽  
Bernd T. Meyer
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Linear Models ◽  
Transfer Functions ◽  
Event Related Potentials ◽  
Stimulus Presentation ◽  
Related Potentials ◽  
Linear And Nonlinear ◽  
Artificial Neural

Diferent linear models have been proposed to establish a link between an auditory stimulus and the neurophysiological response obtained through electroencephalography (EEG). We investigate if non-linear mappings can be modeled with deep neural networks trained on continuous speech envelopes and EEG data obtained in an auditory attention two-speaker scenario. An artificial neural network was trained to predict the EEG response related to the attended and unattended speech envelopes. After training, the properties of the DNN-based model are analyzed by measuring the transfer function between input envelopes and predicted EEG signals by using click-like stimuli and frequency sweeps as input patterns. Using sweep responses allows to separate the linear and nonlinear response components also with respect to attention. The responses from the model trained on normal speech resemble event-related potentials despite the fact that the DNN was not trained to reproduce such patterns. These responses are modulated by attention, since we obtain significantly lower amplitudes at latencies of 110 ms, 170 ms and 300 ms after stimulus presentation for unattended processing in contrast to the attended. The comparison of linear and nonlinear components indicates that the largest contribution arises from linear processing (75%), while the remaining 25% are attributed to nonlinear processes in the model. Further, a spectral analysis showed a stronger 5 Hz component in modeled EEG for attended in contrast to unattended predictions. The results indicate that the artificial neural network produces responses consistent with recent findings and presents a new approach for quantifying the model properties.

Sign in / Sign up

Export Citation Format

Share Document