Electroencephalogram (EEG) individual differences are a critical factor influencing EEG-based emotion recognition, yet they have not been thoroughly investigated, hindering the development of affective Brain-Computer Interfaces (aBCI). Facing the lack of EEG information decoding research, we conducted an interpretable study on EEG individual differences using five datasets (SEED, SEED-IV, SEED-V, RCLS, and MPED). We analyzed the impact of different EEG information (individual, session, emotion, and trial) through sample space visualization, aggregation phenomena quantification, and energy pattern analysis. By examining emotional difference feature distribution patterns, we identified the Cross-Session Consistency of Individual Emotional Patterns (CCIEP) and the Individual Emotional Pattern Difference (IEPD). These characteristics are the main factors impacting emotion recognition stability. To quantify emotional patterns, we proposed the Correction T-test (CT) weight extraction method. Leveraging individual emotional pattern and trial information, we developed the Weight-based Channel-model Matrix Framework (WCMF) to address limitations of traditional modeling approaches caused by IEPD. Finally, WCMF was validated on cross-dataset tasks through two practical scenario experiments. The results demonstrated that WCMF achieves more stable and superior performance compared to traditional methods. This study provides a deeper understanding of EEG individual differences and offers a robust framework to advance aBCI systems.

Currently, treatment options for patients with treatment-resistant depression (TRD) are extremely limited. Although the rapid antidepressant effects of nitrous oxide have been preliminarily validated, the underlying neurophysiological mechanisms remain unclear. In this study, we intended to explore impact on brain modules induced by nitrous oxide through a randomized controlled study. A total of 44 patients with TRD were recruited. Subjects were randomly allocated to nitrous oxide-intervention group (1-hour inhalation of 50% nitrous oxide/50% oxygen) or placebo-control group (1-hour inhalation of 50% oxygen/50% air). The eye-closed resting state EEG signals were recorded. Our findings reveal that nitrous oxide treatment significantly altered the segregation of interconnected brain functional modules. The result of General Linear Model (repeated measurements) demonstrates superior antidepressant efficacy of nitrous oxide compared to placebo, as evidenced by reductions in both the participation coefficient and connector hub after treatment. Furthermore, changes in modular metrics moderately correlated with reductions in depressive symptoms. These findings offer valuable insights into mechanism underlying the treatment of TRD using nitrous oxide from a modular perspective.

Many studies attempt to explore the underlying mechanisms of depression and distinguish between depression patients and normal controls (NC) using electroencephalography (EEG) techniques. With the advancement of deep learning methods, an increasing number of studies aim to design Computer-Aided Diagnosis (CAD) systems for mild depression (MD) to achieve early identification. However, few studies construct models based on abnormal brain topological structures specific to MD patients. In this study, we investigate the abnormal brain topological structures of individuals with MD based on EEG data obtained during an emotional face paradigm. Functional connectivity analysis reveals a higher proportion of inter-hemispheric connections in the MD group compared to intra-hemispheric connections. Additionally, intra-hemispheric connections are primarily observed within the frontal and parietal lobes of both groups. Hierarchical clustering analysis results indicate impairments in the frontal and parietal lobes in the MD group compared to the NC group. Based on these findings, we propose a novel feature called "cross-brain feature" and introduce a multi-cross-brain attention fusion mechanism to integrate information between brain regions. We train and test our models using 5-fold cross-validation. The results demonstrate that the classification model based on abnormal brain topological structures achieves the highest performance among the three state-of-the-art algorithms, with an accuracy of 80.1%, an area under the ROC curve (AUC) of 80%, and a sensitivity (SEN) of 86.3%. These findings suggest that combining abnormal brain topological structures derived from functional connectivity matrices with deep learning techniques can provide an effective objective approach to the early detection of depression.

Deep learning methods have been widely adopted in the field of computer-aided EEG diagnosis, one of the major topics to be investigated is the input format of EEG data. Related researches have reported the application of Fast Fourier Transform (FFT) to topology-preserving multi-spectral images generation on three frequency bands (i.e. theta, alpha and beta) jointed to preserve the spatial information. In our work, we proposed a new approach of using Empirical Mode Decomposition (EMD) instead of FFT algorithm to generate topology-preserving multi-spectral images on intrinsic mode functions (IMFs) jointed and only on the single IMF. Meanwhile, images generated on three frequency bands jointed and the single frequency band were also used for comparison. We then applied two convolutional neural network (CNN) structures to distinguish depression patients and normal subjects from the topology-preserving multi-spectral images. As a result, both convolutional networks obtain better accuracies on three IMFs jointed (about 5% better, ≈ 75% vs. ≈ 70%) than three frequency bands jointed. Analysis based on the single frequency band indicates that alpha band performs the best, with an accuracy of more than 78%, and among all classification results, the best classification accuracy obtained is 80.23% on IMF2. The results are encouraging, despite the limited size of our cohort, the use of EMD and our findings cast a new light on application of deep learning method to EEG-based depression recognition.

The detection of mild depression is conducive to the early intervention and treatment of depression. This study explored the fusion of electroencephalography (EEG) and pupil area signals to build an effective and convenient mild depression recognition model. We proposed Mutual Information Based Fusion Model (MIBFM), which innovatively used pupil area signals to select EEG electrodes based on mutual information. Then we extracted features from EEG and pupil area signals in different bands, and fused bimodal features using the denoising autoencoder. Experimental results showed that MIBFM could obtain the highest accuracy of 87.03%. And MIBFM exhibited better performance than other existing methods. Our findings validate the effectiveness of the use of pupil area as signals, which makes eye movement signals can be easily obtained using high resolution camera, and the EEG electrode selection scheme based on mutual information is also proved to be an applicable solution for data dimension reduction and multimodal complementary information screening. This study casts a new light for mild depression recognition using multimodal data of EEG and pupil area signals, and provides a theoretical basis for the development of portable and universal application systems.

Welcome to the second issue of the IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS (TCSS) of 2023. According to the latest update of CiteScoreTracker from Elsevier Scopus released on February 5, 2023, the CitesSore of TCSS has reached a historical high of 9.6. Many thanks to all for your great effort and support.

In recent years, depression recognition using physiological signals has achieved certain progress, but mild depression recognition is still in its infancy. Early detection can prevent the development of depression, and combining multiple modalities for analysis has been proved effective in the domain of mental disorders detection. Electroencephalogram (EEG) and eye movements (EM) are widely used to identify depression. However, the problem of using physiological signals to detect mental illness is that the generalization ability of the model is not strong, which is caused by individual differences. In view of the above problems, this paper proposes a hybrid fusion model based on deep belief network (DBN) and secondary classifier, called HFMBDSC, which first uses unsupervised DBN to fuse EEG features (linear, nonlinear features and network features) at the feature level. DBN transforms EEG features into another form to mitigate the effects of individual differences in EEG, and obtains DBN features that can more comprehensively represent EEG information. In the next step of decision level fusion, DBN features and EM features jointly make the final decision using the three classifiers that perform best when using single modality. The results show that DBN can improve the classification accuracy and effectively reduce the influence of individual differences in EEG features through visual analysis of feature space. The classification performance of HFMBDSC is significantly improved compared to the traditional single modality results. The highest accuracy rate of 89.54% is obtained under 10-fold cross-validation. These results suggest that mild depression recognition based on HFMBDSC is promising.

Welcome to the fourth issue of IEEE Transactions on Computational Social Systems (TCSS) this year. I am excited to share some great news.

Electroencephalography (EEG) have been proved to be effective in the field of depression recognition, however, the application of EEG-based mild depression detection is still in its infancy. Our work mainly focused on mild depression recognition of college students, based on high-density 128-channel EEG recordings from 24 mild depression individuals and 24 normal subjects using facial expression as experimental materials. In order to prevent individual performance differences on the convolution kernel, and to integrate time information and spatial information instead of simply combining, we proposed a new deep learning model with multiple convolution kernels and a Long Short-Term Memory (LSTM) strategy based on convolution. Batch normalization has been widely used and proved to be effective in some research areas, for example computer vision. Our findings show that for EEG data, batch normalization will reduce the accuracy due to the special data characteristics of EEG. It was found that the proposed model achieved an accuracy of 83.47% with the 8-fold cross-validation, and Batch Normalization will reduce the accuracy because it eliminated the difference between depression and normal. Our findings cast a new light to recognize mild depression accurately and quickly, it could be used as auxiliary tools to diagnose and predict mild depression in the future.

Previous studies have shown that functional changes in depression are a context-specific rather than generalized across stimuli. In our study, with the aim of investigating the differences in the functional networks, electroencephalogram (EEG) data were collected from 27 mild depression (MD) and 27 normal controls (NC) using the Cue - Target paradigm. The exact low resolution electromagnetic tomography (eLORETA) method is applied to estimate the three-dimensional distribution of the current density of the brain source, and lagged coherence(LC), lagged phase synchronization(LPS), lagged linear connectivity(LLC), lagged nonlinear connectivity(LNC) are used to calculate the functional connections between pairs of regions of interest. In the four frequency bands of delta, theta, alpha and beta, the clustering coefficient (CC) and characteristic path length (CPL) were calculated and statistical analysis was performed. Our research found that the electrophysiological activity of MD in Brodmann area (BA) 7 was stronger than that of NC in all frequency bands. When the cue is color block and the prompt is inconsistent with the target, the CC and CPL of MD and NC were significantly different in the delta, theta, and beta frequency band. This result showed that the functional network change of MD was most obvious when cue is an arrow and the cue is inconsistent with the target. In this condition, the CC and CPL of MD are greater than that of NC, indicating that these network characteristics might be used as biological indicators to identify MD.

Major depression disorder (MDD) has been proved to have difficulty in emotional conflict processing. The objective of this study is to investigate the different attention deployment patterns in emotional conflict processing between MDDs and Healthy Controls (HCs) using eye tracking data. A face-word Stroop task was used, and 49 MDDs and 50 healthy controls (HCs) were recruited in the experiment. Finally, our results indicated that MDDs demonstrated lower accuracy (ACC) compared with HCs during the process of emotional conflict. Moreover, we found attention bias in the process of attention maintenance but not vigilance, and it may be one of the possible reasons for different ability of emotional conflict processing between MDDs and HCs.

Depression, influencing millions of people, has become a major disease in the past decade. However, the assessment methods of diagnosing depression almost exclusively rely on patient-reported or clinical judgments of symptom severity, which are associated with subjective biases and intensive labor. Some bio-signals such as EEG and eye movements are used for automatic detection but their accuracies are not accurate enough for the real application, further improvements are needed. This research proposes a content based ensemble method (CBEM) to promote the depression detection accuracy, generating data subsets by the content of the experiment, then using the majority vote of subsets to determine the subjects' label. The validation of the method is testified by two different experiments which included free viewing eye tracking and task-state EEG and these two experiments have 36, 40 subjects respectively. In these two experiments CBEM gains accuracies of 82.5% and 92.73% respectively. The results show that CBEM outperform traditional classification methods. Our findings provide an effective solution for promoting the accuracy of depression identification, and give an objective and quantitative evaluation of depression, which in the future could be used for the auxiliary diagnosis of depression.