EEG-Based BCI Emotion Recognition: A Survey

Torres, Pablo; Torres, Edgar A.; Hernández-Álvarez, Myriam; Yoo, Sang Guun

doi:10.3390/s20185083

Cited by 193 publications

(115 citation statements)

References 131 publications

Supporting

Mentioning

111

Contrasting

Unclassified

Order By: Relevance

“…Despite their varied typical cultural and brain activity profiles [74] , [90] , [105] , [112] , [164] , [182] , and novel relationship between task and resting-state brain activity [57] , emotions represent only positive or negative energy conditions [62] , [71] . The activity patterns elicited by different tasks from the Human Connectome Project can also be reconstructed from a minimal subset of functional harmonics, indicating the interrelated relationship between task and resting-state brain activities.…”

Section: Thermodynamic Considerations Of Brain Activitiesmentioning

confidence: 99%

The thermodynamics of cognition: A mathematical treatment

Deli¹,

Peters

Kisvárday

2021

Computational and Structural Biotechnology Journal

View full text Add to dashboard Cite

Graphical abstract The thermodynamics consequences of intelligent computation. The neural system uses substantial resources to maintain the resting state; therefore, the Carnot cycle can model the evoked cycle. The synaptic complexity during neural computation changes in parallel with the ability to respond to future challenges. Positive emotions (top) trigger the reversed Carnot cycle, which accumulates energy and entropy (the entropy at the end of the cycle is A′ where S (A′) > S (A)). Negative emotions initiate an exothermic Carnot cycle, which accumulates information and wastes energy (entropy at the end of the cycle is A′ and S (A′) < S (A)).

show abstract

Section: Thermodynamic Considerations Of Brain Activitiesmentioning

confidence: 99%

The thermodynamics of cognition: A mathematical treatment

Deli¹,

Peters

Kisvárday

2021

Computational and Structural Biotechnology Journal

View full text Add to dashboard Cite

show abstract

“…Focusing on affecting computing, which is an artificial intelligence area of study that recognizes, interprets, processes and simulates human affect computers, a survey of the pertinent scientific literature on affecting computing from 2015 to 2020 was presented in [ 17 ]. The paper presented trends and compared algorithm applications in new implementations from a computer science perspective.…”

Section: Dermatological Sensorsmentioning

confidence: 99%

Emotion and Stress Recognition Related Sensors and Machine Learning Technologies

Kyamakya

Machot

Mosa

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Electroencephalography (EEG) has gained considerable attention from researchers since it has advanced into a viable option for portable, affordable, and easy-to-use solutions for detecting emotions directly from the brain. However, the biggest challenge for existing EEG technologies is that the signal is very weak and is naturally subject to non-emotion related noise effects [ 2 ]. This noise can be internal, a natural effect of user variability, or external, caused by electrical signals from the rest of the body or the surrounding environment.…”

Section: Introductionmentioning

confidence: 99%

Learning Subject-Generalized Topographical EEG Embeddings Using Deep Variational Autoencoders and Domain-Adversarial Regularization

Hagad

Kimura

Fukui

et al. 2021

Sensors

View full text Add to dashboard Cite

Two of the biggest challenges in building models for detecting emotions from electroencephalography (EEG) devices are the relatively small amount of labeled samples and the strong variability of signal feature distributions between different subjects. In this study, we propose a context-generalized model that tackles the data constraints and subject variability simultaneously using a deep neural network architecture optimized for normally distributed subject-independent feature embeddings. Variational autoencoders (VAEs) at the input level allow the lower feature layers of the model to be trained on both labeled and unlabeled samples, maximizing the use of the limited data resources. Meanwhile, variational regularization encourages the model to learn Gaussian-distributed feature embeddings, resulting in robustness to small dataset imbalances. Subject-adversarial regularization applied to the bi-lateral features further enforces subject-independence on the final feature embedding used for emotion classification. The results from subject-independent performance experiments on the SEED and DEAP EEG-emotion datasets show that our model generalizes better across subjects than other state-of-the-art feature embeddings when paired with deep learning classifiers. Furthermore, qualitative analysis of the embedding space reveals that our proposed subject-invariant bi-lateral variational domain adversarial neural network (BiVDANN) architecture may improve the subject-independent performance by discovering normally distributed features.

show abstract

EEG-Based BCI Emotion Recognition: A Survey

Cited by 193 publications

References 131 publications

The thermodynamics of cognition: A mathematical treatment

The thermodynamics of cognition: A mathematical treatment

Emotion and Stress Recognition Related Sensors and Machine Learning Technologies

Learning Subject-Generalized Topographical EEG Embeddings Using Deep Variational Autoencoders and Domain-Adversarial Regularization

Contact Info

Product

Resources

About