作者：
Wu, XunZheng, Wei-LongLi, ZiyiLu, Bao-LiangShanghai Jiao Tong Univ
Ctr Brain Comp & Machine Intelligence Dept Comp Sci & Engn 800 Dongchuan Rd Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ 55455
RuiJin Hosp Clin Neurosci Ctr Sch Med 197 Ruijin 2nd Rd Shanghai 200020 Peoples R China Shanghai Jiao Tong Univ
RuiJin Hosp RuiJin Mihoyo Lab Sch Med 197 Ruijin 2nd Rd Shanghai 200020 Peoples R China Shanghai Jiao Tong Univ
Key Lab Shanghai Educ Commiss Intelligent Interac 800 Dongchuan Rd Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ
Brain Sci & Technol Res Ctr 800 Dongchuan Rd Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ
Qing Yuan Res Inst 800 Dongchuan Rd Shanghai 200240 Peoples R China MIT
Dept Brain & Cognit Sci Cambridge MA 02319 USA
Objective. Previous studies on emotion recognition from electroencephalography (EEG) mainly rely on single-channel-based feature extraction methods, which ignore the functional connectivity between brain regions. Henc...
详细信息
Objective. Previous studies on emotion recognition from electroencephalography (EEG) mainly rely on single-channel-based feature extraction methods, which ignore the functional connectivity between brain regions. Hence, in this paper, we propose a novel emotion-relevant critical subnetwork selection algorithm and investigate three EEG functional connectivity network features: strength, clustering coefficient, and eigenvector centrality. Approach. After constructing the brain networks by the correlations between pairs of EEG signals, we calculated critical subnetworks through the average of brain network matrices with the same emotion label to eliminate the weak associations. Then, three network features were conveyed to a multimodal emotion recognition model using deep canonical correlation analysis along with eye movement features. The discrimination ability of the EEG connectivity features in emotion recognition is evaluated on three public datasets: SEED, SEED-V, and DEAP. Main results. The experimental results reveal that the strength feature outperforms the state-of-the-art features based on single-channel analysis. The classification accuracies of multimodal emotion recognition are 95.08 +/- 6.42% on the SEED dataset, 84.51 +/- 5.11% on the SEED-V dataset, and 85.34 +/- 2.90% 86.61 +/- 3.76% for arousal and valence on the DEAP dataset, respectively, which all achieved the best performance. In addition, the brain networks constructed with 18 channels achieve comparable performance with that of the 62-channel network and enable easier setups in real scenarios. Significance. The EEG functional connectivity networks combined with emotion-relevant critical subnetworks selection algorithm we proposed is a successful exploration to excavate the information between channels.
Traditional machine learning methods suffer from severe overfitting in EEG-based emotion reading. In this paper, we use hierarchical convolutional neural network (HCNN) to classify the positive, neutral, and negative ...
详细信息
Traditional machine learning methods suffer from severe overfitting in EEG-based emotion reading. In this paper, we use hierarchical convolutional neural network (HCNN) to classify the positive, neutral, and negative emotion states. We organize differential entropy features from different channels as two-dimensional maps to train the HCNNs. This approach maintains information in the spatial topology of electrodes. We use stacked autoencoder (SAE), SVM, and KNN as competing methods. HCNN yields the highest accuracy, and SAE is slightly inferior. Both of them show absolute advantage over traditional shallow models including SVM and KNN. We confirm that the high-frequency wave bands Beta and Gamma are the most suitable bands for emotion reading. We visualize the hidden layers of HCNNs to investigate the feature transformation flow along the hierarchical structure. Benefiting from the strong representational learning capacity in the two-dimensional space, HCNN is efficient in emotion recognition especially on Beta and Gamma waves.
affective models based on EEG signals have been proposed in recent years. However, most of these models require subject-specific training and generalize worse when they are applied to new subjects. This is mainly caus...
详细信息
ISBN:
(数字)9783030042219
ISBN:
(纸本)9783030042219;9783030042202
affective models based on EEG signals have been proposed in recent years. However, most of these models require subject-specific training and generalize worse when they are applied to new subjects. This is mainly caused by the individual differences across subjects. While, on the other hand, it is time-consuming and high cost to collect subject-specific training data for every new user. How to eliminate the individual differences in EEG signals for implementation of affective models is one of the challenges. In this paper, we apply Deep adaptation network (DAN) to solve this problem. The performance is evaluated on two publicly available EEG emotion recognition datasets, SEED and SEED-IV, in comparison with two baseline methods without domain adaptation and several other domain adaptation methods. The experimental results indicate that the performance of DAN is significantly superior to the existing methods.
Emotional preference of people from different ethnicity would alter multimedia implicit tagging remarkably. It can be speculated that the people from each ethnic group would prefer the folk music of their own ethnicit...
详细信息
Emotional preference of people from different ethnicity would alter multimedia implicit tagging remarkably. It can be speculated that the people from each ethnic group would prefer the folk music of their own ethnicity more than the others. An emotionally intelligent system based on electroencephalography (EEG) is proposed in this study to test this hypothesis. Four channels of EEG signals of 16 healthy subjects from different ethnic groups were recorded during 4 two-minute long excerpts of folk music. Six types of features extracted and a subset of them were selected based on minimum-RedundancyMaximum-Relevance (mRMR) algorithm. The top-ranked features were fed to the Support Vector Machine (SVM) classifier with Radial Basis Function (RBF) kernel with various similarity metrics. The performance of the proposed method was assessed in terms of F1-score and accuracy (ACC) using random sub-sampling cross validation scheme. The highest performance for the single SVM classifier was achieved by Dynamic Time Warping (DTW) based RBF kernel which was significantly higher than the chance level. These results approve that the tendency of people from each ethnic group to their ethnicity is significantly reflected in their EEG signals which can be automatically detected. (C) 2017 Elsevier Inc. All rights reserved.
Constructing a robust emotion-aware analytical framework using non-invasively recorded electroencephalogram (EEG) signals has gained intensive attentions nowadays. However, as deploying a laboratory-oriented proof-of-...
详细信息
Constructing a robust emotion-aware analytical framework using non-invasively recorded electroencephalogram (EEG) signals has gained intensive attentions nowadays. However, as deploying a laboratory-oriented proof-of-concept study toward real-world applications, researchers are now facing an ecological challenge that the EEG patterns recorded in real life substantially change across days (i.e., day-to-day variability), arguably making the pre-defined predictive model vulnerable to the given EEG signals of a separate day. The present work addressed how to mitigate the inter-day EEG variability of emotional responses with an attempt to facilitate cross-day emotion classification, which was less concerned in the literature. This study proposed a robust principal component analysis (RPCA)-based signal filtering strategy and validated its neurophysiological validity and machine-learning practicability on a binary emotion classification task (happiness vs. sadness) using a five-day EEG dataset of 12 subjects when participated in a music-listening task. The empirical results showed that the RPCA-decomposed sparse signals (RPCA-S) enabled filtering off the background EEG activity that contributed more to the inter-day variability, and predominately captured the EEG oscillations of emotional responses that behaved relatively consistent along days. Through applying a realistic add-day-in classification validation scheme, the RPCA-S progressively exploited more informative features (from 12.67 +/- 5.99 to 20.83 +/- 7.18) and improved the cross-day binary emotion-classification accuracy (from 58.31 +/- 12.33% to 64.03 +/- 8.40%) as trained the EEG signals from one to four recording days and tested against one unseen subsequent day. The original EEG features (prior to RPCA processing) neither achieved the cross-day classification (the accuracy was around chance level) nor replicated the encouraging improvement due to the inter -day EEG variability. This result demonstrated the e
Positive environmental emotion feedback is important to influence the brain and behaviors. By measuring emotional signals and providing affective neurofeedback, people can be better aware of their emotional state in r...
详细信息
ISBN:
(纸本)9781479999538
Positive environmental emotion feedback is important to influence the brain and behaviors. By measuring emotional signals and providing affective neurofeedback, people can be better aware of their emotional state in real time. However, such direct mapping does not necessarily motivate people's emotion regulation effort. We introduce two levels of emotion feedback: an augmentation level that indicates direct feedback mapping and an intervention level which means feedback output is dynamically adapted with the regulation process. For the purpose of emotion regulation, this research summarizes the framework of emotion feedback design by adding new components that involve feature wrapping, mapping to output representation and interactive interface representation. By this means, the concept of intelligent emotion feedback is illustrated that not only enhances emotion regulation motivation but also considers subject and trial variability based on individual calibration and learning. An affective brain-computer interface technique is used to design the prototype among alternatives. Experimental tests and model simulation are planned for further evaluation.
The present Ph.D. project explores possibilities to apply neurophysiological methods for affect detection during human-technology interaction (HTI). Portable neurophysiological methods such as electroencephalography (...
详细信息
ISBN:
(纸本)9781479999538
The present Ph.D. project explores possibilities to apply neurophysiological methods for affect detection during human-technology interaction (HTI). Portable neurophysiological methods such as electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) offer an objective, ecologically valid and rather convenient way to infer the user's affective state through the monitoring of brain activity. To identify neural signatures for positive and negative affective user reactions an empirical study is proposed. The experimental design of this study enables synchronous data acquisition for EEG, fNIRS and psychophysiological measurements while the user is interacting with an adaptive web-interface. During the interaction process positive and negative affective states are induced by system-generated adaptive actions which are either appropriate and helpful or inappropriate and impedimental. The findings of the empirical study shed light into the question whether EEG, fNIRS or a hybrid approach that combines the employed methods is most reliable for affect detection during HTI.
暂无评论