Constructing a Personalized Cross-Day EEG-Based Emotion-Classification Model Using Transfer Learning

Using Social Media Data to Explore the Impact of Green Spaces on Physical Activity and Sentiment in China:A Systematic Literature Review

Social media data offers invaluable insights into the crucial role that green spaces play in promoting physical activity and sentiment. This study systematically reviewed the scientific evidence exploring the influence of green spaces on physical activity and sentimental states based on social media data. Overall, 8,842 studies were identified through searching PubMed, Web of Science, Scopus, and CNKI databases up to December 30, 2023. Forty studies met the inclusion criteria for this systematic review. The analysis of the included literature revealed that the physical attributes of green spaces-such as accessibility, area of green spaces, natural landscape features and quality, types of greenery, visibility of green spaces, and the Normalized Difference Vegetation Index are correlated with the expression of positive sentiments among residents or significantly enhance residents' physical activity. The research unveiled the profound impact of multidimensional factors on the green space experience, including gender, age, geographic location, individual needs and preferences, seasonal variations, and socioeconomic backgrounds. Additionally, thermal adaptation and the frequency of park visits serve as mediating factors in the relationship between green spaces and sentiment or physical activity. Although this study provides important insights into the associations of green spaces on physical activity and sentiment, it faces limitations such as reliance on a single data source and the lack of geographic universality. In conclusion, green spaces significantly enhance physical activity and improve sentimental states across populations. Analysis of social media data facilitates a nuanced understanding of public behaviors and attitudes toward green spaces. Using advanced data analysis techniques is essential for a deeper comprehension of the dynamic interplay and underlying mechanisms that shape the impact of green spaces on physical activity and sentiment. Future research should prioritize the integration data from various platforms and the implementation of longitudinal studies.

Motor task-to-task transfer learning for motor imagery brain-computer interfaces

Motor imagery (MI) is one of the popular control paradigms in the non-invasive brain-computer interface (BCI) field. MI-BCI generally requires users to conduct the imagination of movement (e.g., left or right hand) to collect training data for generating a classification model during the calibration phase. However, this calibration phase is generally time-consuming and tedious, as users conduct the imagination of hand movement several times without being given feedback for an extended period. This obstacle makes MI-BCI non user-friendly and hinders its use. On the other hand, motor execution (ME) and motor observation (MO) are relatively easier tasks, yield lower fatigue than MI, and share similar neural mechanisms to MI. However, few studies have integrated these three tasks into BCIs. In this study, we propose a new task-to-task transfer learning approach of 3-motor tasks (ME, MO, and MI) for building a better user-friendly MI-BCI. For this study, 28 subjects participated in 3-motor tasks experiment, and electroencephalography (EEG) was acquired. User opinions regarding the 3-motor tasks were also collected through questionnaire survey. The 3-motor tasks showed a power decrease in the alpha rhythm, known as event-related desynchronization, but with slight differences in the temporal patterns. In the classification analysis, the cross-validated accuracy (within-task) was 67.05% for ME, 65.93% for MI, and 73.16% for MO on average. Consistently with the results, the subjects scored MI (3.16) as the most difficult task compared with MO (1.42) and ME (1.41), with p < 0.05. In the analysis of task-to-task transfer learning, where training and testing are performed using different task datasets, the ME-trained model yielded an accuracy of 65.93% (MI test), which is statistically similar to the within-task accuracy (p > 0.05). The MO-trained model achieved an accuracy of 60.82% (MI test). On the other hand, combining two datasets yielded interesting results. ME and 50% of the MI-trained model (50-shot) classified MI with a 69.21% accuracy, which outperformed the within-task accuracy (p < 0.05), and MO and 50% of the MI-trained model showed an accuracy of 66.75%. Of the low performers with an within-task accuracy of 70% or less, 90% (n = 21) of the subjects improved in training with ME, and 76.2% (n = 16) improved in training with MO on the MI test at 50-shot. These results demonstrate that task-to-task transfer learning is possible and could be a promising approach to building a user-friendly training protocol in MI-BCI.

Hybrid deep models for parallel feature extraction and enhanced emotion state classification

Emotions play a vital role in recognizing a person's thoughts and vary significantly with stress levels. Emotion and stress classification have gained considerable attention in robotics and artificial intelligence applications. While numerous methods based on machine learning techniques provide average classification performance, recent deep learning approaches offer enhanced results. This research presents a hybrid deep learning model that extracts features using AlexNet and DenseNet models, followed by feature fusion and dimensionality reduction via Principal Component Analysis (PCA). The reduced features are then classified using a multi-class Support Vector Machine (SVM) to categorize different types of emotions. The proposed model was evaluated using the DEAP and EEG Brainwave datasets, both well-suited for emotion analysis due to their comprehensive EEG signal recordings and diverse emotional stimuli. The DEAP dataset includes EEG signals from 32 participants who watched 40 one-minute music videos, while the EEG Brainwave dataset categorizes emotions into positive, negative, and neutral based on EEG recordings from participants exposed to six different film clips. The proposed model achieved an accuracy of 95.54% and 97.26% for valence and arousal categories in the DEAP dataset, respectively, and 98.42% for the EEG Brainwave dataset. These results significantly outperform existing methods, demonstrating the model's superior performance in terms of precision, recall, F1-score, specificity, and Mathew correlation coefficient. The integration of AlexNet and DenseNet, combined with PCA and multi-class SVM, makes this approach particularly effective for capturing the intricate patterns in EEG data, highlighting its potential for applications in human-computer interaction and mental health monitoring, marking a significant advancement over traditional methods.

Lightweight attention mechanisms for EEG emotion recognition for brain computer interface

BACKGROUND

In the realm of brain-computer interfaces (BCI), identifying emotions from electroencephalogram (EEG) data is a difficult endeavor because of the volume of data, the intricacy of the signals, and the several channels that make up the signals.

NEW METHODS

Using dual-stream structure scaling and multiple attention mechanisms (LDMGEEG), a lightweight network is provided to maximize the accuracy and performance of EEG-based emotion identification. Reducing the number of computational parameters while maintaining the current level of classification accuracy is the aim. This network employs a symmetric dual-stream architecture to assess separately time-domain and frequency-domain spatio-temporal maps constructed using differential entropy features of EEG signals as inputs.

RESULT

The experimental results show that after significantly lowering the number of parameters, the model achieved the best possible performance in the field, with a 95.18 % accuracy on the SEED dataset.

COMPARISON WITH EXISTING METHODS

Moreover, it reduced the number of parameters by 98 % when compared to existing models.

CONCLUSION

The proposed method distinct channel-time/frequency-space multiple attention and post-attention methods enhance the model's ability to aggregate features and result in lightweight performance.

Enhancing EEG-based cross-day mental workload classification using periodic component of power spectrum

Objective. The day-to-day variability of electroencephalogram (EEG) poses a significant challenge to decode human brain activity in EEG-based passive brain-computer interfaces (pBCIs). Conventionally, a time-consuming calibration process is required to collect data from users on a new day to ensure the performance of the machine learning-based decoding model, which hinders the application of pBCIs to monitor mental workload (MWL) states in real-world settings.Approach. This study investigated the day-to-day stability of the raw power spectral density (PSD) and their periodic and aperiodic components decomposed by the Fitting Oscillations and One-Over-F algorithm. In addition, we validated the feasibility of using periodic components to improve cross-day MWL classification performance.Main results. Compared to the raw PSD (69.9% ± 18.5%) and the aperiodic component (69.4% ± 19.2%), the periodic component had better day-to-day stability and significantly higher cross-day classification accuracy (84.2% ± 11.0%).Significance. These findings indicate that periodic components of EEG have the potential to be applied in decoding brain states for more robust pBCIs.

Exploring Spatio-Spectral Electroencephalogram Modulations of Imbuing Emotional Intent During Active Piano Playing

Imbuing emotional intent serves as a crucial modulator of music improvisation during active musical instrument playing. However, most improvisation-related neural endeavors have been gained without considering the emotional context. This study attempts to exploit reproducible spatio-spectral electroencephalogram (EEG) oscillations of emotional intent using a data-driven independent component analysis framework in an ecological multiday piano playing experiment. Through the four-day 32-ch EEG dataset of 10 professional players, we showed that EEG patterns were substantially affected by both intra- and inter-individual variability underlying the emotional intent of the dichotomized valence (positive vs. negative) and arousal (high vs. low) categories. Less than half (3-4) of the 10 participants analogously exhibited day-reproducible ( ≥ three days) spectral modulations at the right frontal beta in response to the valence contrast as well as the frontal central gamma and the superior parietal alpha to the arousal counterpart. In particular, the frontal engagement facilitates a better understanding of the frontal cortex (e.g., dorsolateral prefrontal cortex and anterior cingulate cortex) and its role in intervening emotional processes and expressing spectral signatures that are relatively resistant to natural EEG variability. Such ecologically vivid EEG findings may lead to better understanding of the development of a brain-computer music interface infrastructure capable of guiding the training, performance, and appreciation for emotional improvisatory status or actuating music interaction via emotional context.

MS-FRAN: A Novel Multi-Source Domain Adaptation Method for EEG-Based Emotion Recognition

Electroencephalogram (EEG)-based emotion recognition has gradually become a research hotspot. However, the large distribution differences of EEG signals across subjects make the current research stuck in a dilemma. To resolve this problem, in this article, we propose a novel and effective method, Multi-Source Feature Representation and Alignment Network (MS-FRAN). The effectiveness of proposed method mainly comes from three new modules: Wide Feature Extractor (WFE) for feature learning, Random Matching Operation (RMO) for model training, and Top- h ranked domain classifier selection (TOP) for emotion classification. MS-FRAN is not only effective in aligning the distributions of each pair of source and target domains, but also capable of reducing the distributional differences among the multiple source domains. Experimental results on the public benchmark datasets SEED and DEAP have demonstrated the advantage of our method over the related competitive approaches for cross-subject EEG-based emotion recognition.

Electroencephalogram-Driven Machine-Learning Scenario for Assessing Impulse Control Disorder Comorbidity in Parkinson's Disease Using a Low-Cost, Custom LEGO-Like Headset

Patients with Parkinson's disease (PD) may develop cognitive symptoms of impulse control disorders (ICDs) when chronically treated with dopamine agonist (DA) therapy for motor deficits. Motor and cognitive comorbidities critically increase the disability and mortality of the affected patients. This study proposes an electroencephalogram (EEG)-driven machine-learning scenario to automatically assess ICD comorbidity in PD. We employed a classic Go/NoGo task to appraise the capacity of cognitive and motoric inhibition with a low-cost, custom LEGO-like headset to record task-relevant EEG activity. Further, we optimized a support vector machine (SVM) and support vector regression (SVR) pipeline to learn discriminative EEG spectral signatures for the detection of ICD comorbidity and the estimation of ICD severity, respectively. With a dataset of 21 subjects with typical PD, 9 subjects with PD and ICD comorbidity (ICD), and 25 healthy controls (HC), the study results showed that the SVM pipeline differentiated subjects with ICD from subjects with PD with an accuracy of 66.3% and returned an around-chance accuracy of 53.3% for the classification of PD versus HC subjects without the comorbidity concern. Furthermore, the SVR pipeline yielded significantly higher severity scores for the ICD group than for the PD group and resembled the ICD vs. PD distinction according to the clinical questionnaire scores, which was barely replicated by random guessing. Without a commercial, high-precision EEG product, our demonstration may facilitate deploying a wearable computer-aided diagnosis system to assess the risk of DA-triggered cognitive comorbidity in patients with PD in their daily environment.

Movement Artifact Suppression in Wearable Low-Density and Dry EEG Recordings Using Active Electrodes and Artifact Subspace Reconstruction

Wearable low-density dry electroencephalogram (EEG) headsets facilitate multidisciplinary applications of brain-activity decoding and brain-triggered interaction for healthy people in real-world scenarios. However, movement artifacts pose a great challenge to their validity in users with naturalistic behaviors (i.e., without highly controlled settings in a laboratory). High-precision, high-density EEG instruments commonly embed an active electrode infrastructure and/or incorporate an auxiliary artifact subspace reconstruction (ASR) pipeline to handle movement artifact interferences. Existing endeavors motivate this study to explore the efficacy of both hardware and software solutions in low-density and dry EEG recordings against non-tethered settings, which are rarely found in the literature. Therefore, this study employed a LEGO-like electrode-holder assembly grid to coordinate three 3-channel system designs (with passive/active dry vs. passive wet electrodes). It also conducted a simultaneous EEG recording while performing an oddball task during treadmill walking, with speeds of 1 and 2 KPH. The quantitative metrics of pre-stimulus noise, signal-to-noise ratio, and inter-subject correlation from the collected event-related potentials of 18 subjects were assessed. Results indicate that while treating a passive-wet system as benchmark, only the active-electrode design more or less rectified movement artifacts for dry electrodes, whereas the ASR pipeline was substantially compromised by limited electrodes. These findings suggest that a lightweight, minimally obtrusive dry EEG headset should at least equip an active-electrode infrastructure to withstand realistic movement artifacts for potentially sustaining its validity and applicability in real-world scenarios.

Validation of Model-Basis Transfer Learning for a Personalized Electroencephalogram-Based Emotion-Classification Model. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2023;2023:1-4. [PMID: 38082699 DOI: 10.1109/embc40787.2023.10340188] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

The electroencephalogram (EEG)-based affective brain-computer interface (aBCI) has attracted extensive attention in multidisciplinary fields in the past decade. However, the inherent variability of emotional responses recorded in EEG signals increases the vulnerability of pre-trained machine-learning models and impedes the applicability of aBCIs with real-life settings. To overcome the shortcomings associated with the limited personal data in affective modeling, this study proposes a model-basis transfer learning (TL) approach and verifies its feasibility to construct a personalized model using less emotion-annotated data in a longitudinal eight-day dataset comprising data on 10 subjects. By performing daily reliability testing, the proposed TL approach outperformed the subject-dependent counterpart (using limited data only) by ~6% in binary valence classification after recycling a compact set of the eight most transferable models from other subjects. These empirical findings practically contribute to progress in applying TL in realistic aBCI applications.Clinical Relevance- The proposed model-basis TL approach overcomes the shortcoming of inherent variability in EEG signals, supporting realistic aBCI applications.

Comparison of event-related modulation index and traditional methods for evaluating phase-amplitude coupling using simulated brain signals

The investigation of brain oscillations and connectivity has become an important topic in the recent decade. There are several types of interactions between neuronal oscillations, and one of the most interesting among these interactions is phase-amplitude coupling (PAC). Several methods have been proposed to measure the strength of PAC, including the phase-locking value, circular-linear correlation, and modulation index. In the current study, we compared these traditional PAC methods with simulated electroencephalogram signals. Further, to assess the PAC value at each time point, we also compared two recently established methods, event-related phase-locking value and event-related circular-linear correlation, with our newly proposed event-related modulation index (ERMI). Results indicated that the ERMI has better temporal resolution and is more tolerant to noise than the other two event-related methods, suggesting the advantages of utilizing ERMI in evaluating the strength of PAC within a brain region.

Side-Aware Meta-Learning for Cross-Dataset Listener Diagnosis With Subjective Tinnitus

With the development of digital technology, machine learning has paved the way for the next generation of tinnitus diagnoses. Although machine learning has been widely applied in EEG-based tinnitus analysis, most current models are dataset-specific. Each dataset may be limited to a specific range of symptoms, overall disease severity, and demographic attributes; further, dataset formats may differ, impacting model performance. This paper proposes a side-aware meta-learning for cross-dataset tinnitus diagnosis, which can effectively classify tinnitus in subjects of divergent ages and genders from different data collection processes. Owing to the superiority of meta-learning, our method does not rely on large-scale datasets like conventional deep learning models. Moreover, we design a subject-specific training process to assist the model in fitting the data pattern of different patients or healthy people. Our method achieves a high accuracy of 73.8% in the cross-dataset classification. We conduct an extensive analysis to show the effectiveness of side information of ears in enhancing model performance and side-aware meta-learning in improving the quality of the learned features.

Not All Electrode Channels Are Needed: Knowledge Transfer From Only Stimulated Brain Regions for EEG Emotion Recognition

Emotion recognition from affective brain-computer interfaces (aBCI) has garnered a lot of attention in human-computer interactions. Electroencephalographic (EEG) signals collected and stored in one database have been mostly used due to their ability to detect brain activities in real time and their reliability. Nevertheless, large EEG individual differences occur amongst subjects making it impossible for models to share information across. New labeled data is collected and trained separately for new subjects which costs a lot of time. Also, during EEG data collection across databases, different stimulation is introduced to subjects. Audio-visual stimulation (AVS) is commonly used in studying the emotional responses of subjects. In this article, we propose a brain region aware domain adaptation (BRADA) algorithm to treat features from auditory and visual brain regions differently, which effectively tackle subject-to-subject variations and mitigate distribution mismatch across databases. BRADA is a new framework that works with the existing transfer learning method. We apply BRADA to both cross-subject and cross-database settings. The experimental results indicate that our proposed transfer learning method can improve valence-arousal emotion recognition tasks.

Objective assessment of impulse control disorder in patients with Parkinson's disease using a low-cost LEGO-like EEG headset: a feasibility study

Background

Patients with Parkinson’s disease (PD) can develop impulse control disorders (ICDs) while undergoing a pharmacological treatment for motor control dysfunctions with a dopamine agonist (DA). Conventional clinical interviews or questionnaires can be biased and may not accurately diagnose at the early stage. A wearable electroencephalogram (EEG)-sensing headset paired with an examination procedure can be a potential user-friendly method to explore ICD-related signatures that can detect its early signs and progression by reflecting brain activity.

Methods

A stereotypical Go/NoGo test that targets impulse inhibition was performed on 59 individuals, including healthy controls, patients with PD, and patients with PD diagnosed by ICDs. We conducted two Go/NoGo sessions before and after the DA-pharmacological treatment for the PD and ICD groups. A low-cost LEGO-like EEG headset was used to record concurrent EEG signals. Then, we used the event-related potential (ERP) analytical framework to explore ICD-related EEG abnormalities after DA treatment.

Results

After the DA treatment, only the ICD-diagnosed PD patients made more behavioral errors and tended to exhibit the deterioration for the NoGo N2 and P3 peak amplitudes at fronto-central electrodes in contrast to the HC and PD groups. Particularly, the extent of the diminished NoGo-N2 amplitude was prone to be modulated by the ICD scores at Fz with marginal statistical significance (r = − 0.34, p = 0.07).

Conclusions

The low-cost LEGO-like EEG headset successfully captured ERP waveforms and objectively assessed ICD in patients with PD undergoing DA treatment. This objective neuro-evidence could provide complementary information to conventional clinical scales used to diagnose ICD adverse effects.

Application of Transfer Learning in EEG Decoding Based on Brain-Computer Interfaces: A Review

The algorithms of electroencephalography (EEG) decoding are mainly based on machine learning in current research. One of the main assumptions of machine learning is that training and test data belong to the same feature space and are subject to the same probability distribution. However, this may be violated in EEG processing. Variations across sessions/subjects result in a deviation of the feature distribution of EEG signals in the same task, which reduces the accuracy of the decoding model for mental tasks. Recently, transfer learning (TL) has shown great potential in processing EEG signals across sessions/subjects. In this work, we reviewed 80 related published studies from 2010 to 2020 about TL application for EEG decoding. Herein, we report what kind of TL methods have been used (e.g., instance knowledge, feature representation knowledge, and model parameter knowledge), describe which types of EEG paradigms have been analyzed, and summarize the datasets that have been used to evaluate performance. Moreover, we discuss the state-of-the-art and future development of TL for EEG decoding. The results show that TL can significantly improve the performance of decoding models across subjects/sessions and can reduce the calibration time of brain-computer interface (BCI) systems. This review summarizes the current practical suggestions and performance outcomes in the hope that it will provide guidance and help for EEG research in the future.

An On-Chip Processor for Chronic Neurological Disorders Assistance Using Negative Affectivity Classification

Chronic neurological disorders (CND's) are lifelong diseases and cannot be eradicated, but their severe effects can be alleviated by early preemptive measures. CND's, such as Alzheimer's, Autism Spectrum Disorder (ASD), and Amyotrophic Lateral Sclerosis (ALS), are the chronic ailment of the central nervous system that causes the degradation of emotional and cognitive abilities. Long term continuous monitoring with neuro-feedback of human emotions for patients with CND's is crucial in mitigating its harmful effect. This paper presents hardware efficient and dedicated human emotion classification processor for CND's. Scalp EEG is used for the emotion's classification using the valence and arousal scales. A linear support vector machine classifier is used with power spectral density, logarithmic interhemispheric power spectral ratio, and the interhemispheric power spectral difference of eight EEG channel locations suitable for a wearable non-invasive classification system. A look-up-table based logarithmic division unit (LDU) is to represent the division features in machine learning (ML) applications. The implemented LDU minimizes the cost of integer division by 34% for ML applications. The implemented emotion's classification processor achieved an accuracy of 72.96% and 73.14%, respectively, for the valence and arousal classification on multiple publicly available datasets. The 2 x 3mm² processor is fabricated using a 0.18 μm 1P6M CMOS process with power and energy utilization of 2.04 mW and 16 μJ/classification, respectively, for 8-channel operation.