Arousal Dimensions Research Articles

Mood classification through physiological parameters

Future smart agents, like robots, should produce personalized behaviours based on user emotions and moods to fit more in ordinary users’ activities. Besides, the emotions are also linked to human cognitive systems, thus their monitoring could be extremely useful in the case of neurodegenerative diseases such as dementia and Alzheimer. Literature works propose the use of music tracks and videos to stimulate emotions, and cameras to recorder the evoked reactions in human beings. However, these approaches may not be effective in everyday life, due to camera obstructions and different types of stimulation which can be related also with the interaction with other human beings. In this work, we investigate the Electrocardiogram, the ElectroDermal Activity and the Brain Activity signals as main informative channels, acquired through a wireless wearable sensor network. An experimental methodology was built to induce three different emotional states through social interaction. Collected data were classified with three supervised machine learning approaches with different kernels (Support Vector Machine, Decision Tree and k-nearest neighbour) considering the valence dimension and a combination of valence and arousal dimension evoked during the interaction. 34 healthy young participants were involved in the study and a total of 239 instances were analyzed. The supervised algorithms achieve an accuracy of 0.877 in the best case.

A Novel Approach for EEG Electrode Selection in Automated Emotion Recognition Based on Lagged Poincare’s Indices and sLORETA

The goal of this paper was to develop a novel method to track emotional processing in different brain regions using electroencephalogram (EEG) analysis. In addition, the role of EEG electrode selection and feature reduction in emotion recognition was investigated. To this end, the multi-channel EEG signals of 32 subjects available in DEAP dataset were studied. The best EEG electrode positions were selected based on lagged Poincare’s measures of EEG recordings and a source localization method (sLORETA). Three feature reduction algorithms, including random subset feature selection (RSFS), sequential floating forward selection (SFFS), and sequential forward selection (SFS) in combination with support vector machine (SVM), were evaluated to classify high/low valence and high/low arousal. The results showed that RSFS outperformed the other feature selection approaches. In addition, the positive impact of the EEG electrode selection on the classification performances has been confirmed. The most active EEG electrodes were FP1, C3, Cp1, P3, and Pz. Adopting RSFS and selected EEG electrodes, the mean subject-independent accuracies of 73.89 and 74.62% and subject-dependent accuracies of 98.97 and 98.94% were obtained for valence and arousal dimensions, respectively.

Electroencephalogram Emotion Recognition Based on Empirical Mode Decomposition and Optimal Feature Selection

Electroencephalogram (EEG) emotion recognition based on a hybrid feature extraction method in empirical mode decomposition domain combining with optimal feature selection based on sequence backward selection is proposed, which can reflect subtle information of multiscale components of unstable and nonlinear EEG signals and remove the reductant features to improve the performance of emotion recognition. The proposal is tested on DEAP dataset, in which the emotional states in the Valance dimension and Arousal dimension are classified by both ${K}$ -nearest neighbor and support vector machine, respectively. In the experiments, temporal windows of different length and three kinds of rhythms of EEG signal are taken into account for comparison, from which the results show that EEG signal with 1s temporal window achieves highest recognition accuracy of 86.46% in Valence dimension and 84.90% in Arousal dimension, respectively, which is superior to some state-of-the-art works. The proposed method would be applied to real-time emotion recognition in multimodal emotional communication-based humans–robots interaction system.

Are shared models always cultural models? A study of the cultural model of affect and emotion in Chuuk

This article investigates a theoretical tension in cultural models theory between their sharedness and their origin in social and culturally mediated experiences. To address this tension empirically, this article presents a mixed-methods analysis of the shared understanding of a typology of emotion in Chuuk Lagoon. Based on a review of contemporary theory, one would expect that while people in Chuuk would have distinct cultural models for specific culturally meaningful paradigmatic emotions, the shared model for the entire typology of emotion would be structured by two universal dimensions of core affect (arousal and hedonic quality). But this theoretical literature ignores the importance of an intersubjective dimension of prereflective experience. This dimension may be culturally muted in industrial, educated, and rich contexts where most prior research has been conducted but culturally emphasized in places like Chuuk. The mixed-methods analysis finds that an intersubjective dimension along with a hedonic quality dimension structures the shared model of a typology for affect and emotion in Chuuk, while the arousal dimension found elsewhere is muted. Thus, shared models of affect and emotion are cultural models both in terms of specific, culturally elaborated emotions and in terms of the cultural emphasis given to underlying affective and intersubjective dimensions.

Different Realism Designs of 2D Virtual Agents and i ts' Arousal Effect on Students’ Emotions in Learning

This research was aimed to analyse students’ emotions caused by the character’s realism design in the dimension of arousal and its’ effect on emotions in learning. Thus, 2D female virtual agents with four different realism appearance namely realistic, semi-realistic, stylized and cartoon-like were designed and developed. Thus, quasi experimental design was used to answer the research questions derived from this study. The data obtained from the experiment was analysed using ANOVA, post hoc and bootstrap mediation analysis. A number of 600 Electrical Engineering students were chosen from seven polytechnics in Malaysia as respondents in the experiment. From the research it was found that cartoon-like design obtained the lowest mean score arousal followed by stylized, realistic and semi-realistic designs. However, cartoon-like design scored highest mean score for emotions in learning followed by semi-realistic, realistic and stylized designs. Therefore, cartoon-like agent is the best agent among others in inducing positive emotions in learning. In addition, the mediation analysis shows that arousal has mediating effects on relation between different realism designs and emotions in learning. In conclusion, this study recommends cartoon-like agent as the best suitable design for 2D virtual agent followed by semi-realistic, realistic and stylized designs subsequently. Finally, the findings of this study can be a useful guideline for multimedia designers in determining the ideal 2D virtual agent appearance to elicit maximum impact of it in promoting positive emotions in multimedia learning environment.

情绪词加工过程及其情绪效应特点：ERP的证据

<p id="C2">Language plays an important role in the exchange of emotional information. The processing of emotional language has always been the focus of emotional research. Some main views and results on emotional words processing from ERP data are analyzed and summarized: including ERP components and relationship between valence and arousal, as well as emotional effect characteristics. ERP components related to emotional word processing encompass very early P1-N1 components, early EPN components, and late LPP components. Emotional valence and arousal dimensions have shown to function as two significant factors in emotional word processing. Their relationship, however, remains undetermined. Compared with emotional pictures, emotional word processing exerts a weaker emotional effect, showing a bias towards positive valence. Future studies are required to explore relevant factors influencing emotional effect of emotional word processing, to distinguish emotional-laden words from emotional-label words, and to analyze the processing of these two types of emotional words. In addition, it is also an important topic for a future research to elucidate the emotional word processing mechanism in natural context reading by using eye-tracking technique.

A single-response emotion word questionnaire for measuring product-related emotional associations inspired by a circumplex model of core affect: Method characterisation with an applied focus

Product-centric emotion research can deepen understanding of consumers’ product experiences and food choices. Emotion measurement has become widespread in product research, and questionnaires where consumers respond to emotion words are popular and manifold. The present research adds to this multiplicity by presenting a parsimonious single-response questionnaire that spans the dimensions of pleasure and arousal, and their combinations. These are represented in a circular layout with 12 axes radiating from a central point. Each of the arms represents an emotion domain that is exemplified by two emotion words, and consumers’ task is to select the word pair that best represents how they feel. The questionnaire is informed by a circumplex model of human core affect, and methodological aspects relevant for use in applied product research were investigated in 23 consumer studies (n = 104–270 participants per study) spanning a very broad range of products and categories. The first of four research questions (RQ) established that the questionnaire is suitable for use with tasted stimuli (12 studies, 1 of which with aroma stimuli), written stimuli (10 studies) and image stimuli (1 study) (RQ1). Suitability of the questionnaire with New Zealand consumers was confirmed, and extended to Chinese consumers who took part in 7 of the 23 studies (RQ2). Responses obtained using the questionnaire appeared to be repeatable (RQ3a), and based on the criterion of the regression vector (RV) coefficient being equal to or greater than 0.95, stability (RQ3b) could be achieved with as few as 30 consumers despite only obtaining a single response per consumer per stimulus. Finally, it was found that the degree of differences between stimuli in a study influenced discrimination, which was larger when stimuli were more different (RQ4). This intuitively made sense and mirrored past research. By capturing the valence × arousal space, this questionnaire spans more broadly than most emotion word questionnaires for applied and product-focused research, and a notable feature of the obtained responses was that all were relevant (to a larger or smaller extent) in each of the 23 studies. The results are specific to the tested questionnaire variant and future research is needed to determine its goodness-of-fit to the underpinning circumplex model. Comparisons with existing product emotion research questionnaires are yet to be performed, as is testing of questionnaire variants.

Anterior cingulate and medial prefrontal cortex response to systematically controlled tonal dissonance during passive music listening

Several studies have attempted to investigate how the brain codes emotional value when processing music of contrasting levels of dissonance; however, the lack of control over specific musical structural characteristics (i.e., dynamics, rhythm, melodic contour or instrumental timbre), which are known to affect perceived dissonance, rendered results difficult to interpret. To account for this, we used functional imaging with an optimized control of the musical structure to obtain a finer characterization of brain activity in response to tonal dissonance. Behavioral findings supported previous evidence for an association between increased dissonance and negative emotion. Results further demonstrated that the manipulation of tonal dissonance through systematically controlled changes in interval content elicited contrasting valence ratings but no significant effects on either arousal or potency. Neuroscientific findings showed an engagement of the left medial prefrontal cortex (mPFC) and the left rostral anterior cingulate cortex (ACC) while participants listened to dissonant compared to consonant music, converging with studies that have proposed a core role of these regions during conflict monitoring (detection and resolution), and in the appraisal of negative emotion and fear‐related information. Both the left and right primary auditory cortices showed stronger functional connectivity with the ACC during the dissonant portion of the task, implying a demand for greater information integration when processing negatively valenced musical stimuli. This study demonstrated that the systematic control of musical dissonance could be applied to isolate valence from the arousal dimension, facilitating a novel access to the neural representation of negative emotion.

An emotion-based typology of wine consumers

Product consumption is a sensory experience that can evoke a wide range of emotional responses; and accordingly, the emotional component of food consumption has been widely addressed and has appeared in marketing literature with increasing frequency.To date, there is abundant literature on emotions elicited by product consumption, but there is little research concerning emotions and wine. In this context, we develop an emotion-based segmentation based on the emotions that consumers experience when consuming wine, both considering the valence and arousal dimensions of emotions. In addition, we profile and characterize the emotion-based obtained segments.For this purpose, an emotion-based cluster analysis is conducted through a two-step cluster procedure, followed by a MANOVA test on data from 1269 wine consumers.Our findings show that the average wine consumer does not exist, and that wine consumers cannot be seen as a homogenous group. More precisely, four clusters emerge from our results experiencing different emotions in wine consumption: “emotionally unattached”, “negatives”, “contented circumspects” and “wine lovers”. Results suggest that “wine lovers” is the most attractive segment due to their strong wine emotional bond; being the “negatives” the most challenging segment for wine makers. One useful insight for wineries is that it may not be possible to satisfy all these segments with one single wine.

Multichannel Electroencephalography-based Emotion Recognition Using Machine Learning

In recent years, research in the field of human-computer interaction (HCI) has focused on strengthening machine functions in recognizing and understanding human emotions. Emotion recognition can be done in several ways, among others, through sounds, facial expressions, or a combination of both. The different sounds and facial expressions from different races and nations cause less accurate in the reading of emotions using these methods. Another method for recognizing emotions can be done by analysing the data from an electroencephalograph (EEG). The EEG signals from the human brain are the result of various activities carried out. One of them is emotion. The EEG signal used in this study came from the DEAP dataset. This dataset consists of 32 files, each of which contains 40 EEG recordings. The emotions from this dataset are classified based on the dimensions of arousal and valence. The signal was then decomposed into three different frequency groups (alpha, beta, and gamma) through band-pass filtering. After that, the principal component analysis (PCA) and resampling were carried out. The classification processes used a number of methods of machine learning. The result was known that the performance of K-star was the highest while naïve Bayes was the lowest. The accuracies of K-star in arousal and valence classification were 81.2, 82.6, respectively. The naive Bayes got 51.2 for the arousal, and 52.5 for the valence.

Analyzing Emotional Oscillatory Brain Network for Valence and Arousal-Based Emotion Recognition Using EEG Data

Emotion recognition is a highly noteworthy and challenging work in both cognitive science and affective computing. Currently, neurobiology studies have revealed the partially synchronous oscillating phenomenon within brain, which needs to be analyzed from oscillatory synchronization. This combination of oscillations and synchronism is worthy of further exploration to achieve inspiring learning of the emotion recognition models. In this paper, we propose a novel approach of valence and arousal-based emotion recognition using EEG data. First, we construct the emotional oscillatory brain network (EOBN) inspired by the partially synchronous oscillating phenomenon for emotional valence and arousal. And then, a coefficient of variation and Welch’s [Formula: see text]-test based feature selection method is used to identify the core pattern (cEOBN) within EOBN for different emotional dimensions. Finally, an emotional recognition model (ERM) is built by combining cEOBN-inspired information obtained in the above process and different classifiers. The proposed approach can combine oscillation and synchronization characteristics of multi-channel EEG signals for recognizing different emotional states under the valence and arousal dimensions. The cEOBN-based inspired information can effectively reduce the dimensionality of the data. The experimental results show that the previous method can be used to detect affective state at a reasonable level of accuracy.

On the essentialism of places: Between conservative and progressive meanings

Psychological essentialism is a widely studied concept within the social sciences, including the field of psychology. However, this concept has not been empirically explored when applied to perception of places. In this paper, we show that the essentialism vs. anti-essentialism philosophical controversy underlies the discussion carried out in human geography over what constitutes a place, understood as a meaningful location. In a series of three studies, we tested: (a) which places that conform to either essentialist or anti-essentialist criteria are judged by people as more meaningful and are preferred as living places, (b) what the structure of perceived place essentialism is, and (c) whether there exist affective markers of perceived place essentialism (emotions triggered by the essentialist vs. anti-essentialist features of places). In all three studies, pictures of respective locations were used. The studies showed that the places which conform to essentialist criteria were perceived to be more meaningful and were preferred to the anti-essentialist places. A two-dimensional essentialism vs. anti-essentialism structure described the perceived places the best. Affective reactions ordered along the dimensions of valence and arousal were systematically correlated to the perceived essentialism vs. anti-essentialism. The results confirm the claims made by cognitive psychologists that humans are psychological essentialists. The findings may contribute to the rising interest in the essentialism of artifacts.

The impact of recent and concurrent affective context on cognitive control: An ERP study of performance monitoring

The impact of affective context on performance-monitoring aspects of cognitive control is not clear, and findings in the literature are contradictory. The contradictory findings might be due to failing to consider the impact of recent and concurrent affective contexts, participant ratings of emotional images, and the distinct impact of emotional arousal and valence. The present study utilized multilevel modeling (MLM) on person-specific ratings of arousal and valence to determine how recent and concurrent affective context uniquely impacts cognitive control and to determine the unique arousal and valence contributions. Event-related brain potentials (ERPs) were recorded while 35 and 33 participants completed a recent and concurrent affective manipulation, respectively. The recent manipulation used a task wherein a flanker stimulus was preceded by emotional images, and the concurrent manipulation used a task wherein a flanker stimulus was overlaid on emotional images. Higher arousing images were related to larger error-related negativity (ERN) and error positivity (Pe) during the recent but not concurrent affective context manipulation. For valence, more pleasant images were related to attenuated ERN during the concurrent, but not recent, affective context manipulation. For Pe, amplitude decreased with more pleasant valence during the concurrent affective context manipulation, and increased with more pleasant ratings during the recent affective context manipulation. Taken together, recent and concurrent affective context demonstrated unique impacts on performance monitoring. These effects were dissociable along orthogonal dimensions of arousal and valence. The present study highlights the importance of considering arousal and valence in studies of affective context and cognitive control.

Discriminating Emotions in the Valence Dimension from Speech Using Timbre Features

The most used and well-known acoustic features of a speech signal, the Mel frequency cepstral coefficients (MFCC), cannot characterize emotions in speech sufficiently when a classification is performed to classify both discrete emotions (i.e., anger, happiness, sadness, and neutral) and emotions in valence dimension (positive and negative). The main reason for this is that some of the discrete emotions, such as anger and happiness, share similar acoustic features in the arousal dimension (high and low) but are different in the valence dimension. Timbre is a sound quality that can discriminate between two sounds even with the same pitch and loudness. In this paper, we analyzed timbre acoustic features to improve the classification performance of discrete emotions as well as emotions in the valence dimension. Sequential forward selection (SFS) was used to find the most relevant acoustic features among timbre acoustic features. The experiments were carried out on the Berlin Emotional Speech Database and the Interactive Emotional Dyadic Motion Capture Database. Support vector machine (SVM) and long short-term memory recurrent neural network (LSTM-RNN) were used to classify emotions. The significant classification performance improvements were achieved using a combination of baseline and the most relevant timbre acoustic features, which were found by applying SFS on a classification of emotions for the Berlin Emotional Speech Database. From extensive experiments, it was found that timbre acoustic features could characterize emotions sufficiently in a speech in the valence dimension.

How do you feel? Preverbal infants match negative emotions to events.

There is extensive disagreement as to whether preverbal infants have conceptual categories for different emotions (e.g., anger vs. disgust). In addition, few studies have examined whether infants have conceptual categories of emotions within the same dimension of valence and arousal (e.g., high arousal, negative emotions). The current experiments explore one aspect of infants' ability to form conceptual categories of emotions: event-emotion matching. Three experiments investigated whether infants match different negative emotions to specific events. In Experiment 1, 14- and 18-month-olds were randomly assigned to 1 of 3 negative emotion conditions (Anger, Fear, or Disgust). Infants were familiarized with an Emoter interacting with objects in an anger-eliciting event (Unmet Goal) and a disgust-eliciting event (New Food). After each event, the Emoter expressed an emotion that was either congruent or incongruent with the event. Infants matched unmet goals to the expression of anger. However, neither age matched the expression of disgust to an event involving exposure to new food. To probe whether this was a design artifact, a revised New Food event and a fear-congruent event (Strange Toy) were created for Experiment 2. Infants matched the expression of disgust to the new food event, but they did not match fear to an event involving an unfamiliar object. Experiment 3 replicated the disgust findings from Experiment 2 in a sample of 14-month-olds. However, the anger findings from Experiment 1 did not replicate. Taken together, these results suggest that preverbal infants are beginning to form specific matches between some negative emotional expressions and events. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Using the emotion circumplex to uncover sensory drivers of emotional associations to products: six case studies

While emotional associations to products are increasingly used alongside sensory product characterisation in consumer research, studies that link these product responses remain less common. Six studies (104–270 people per study) were conducted using diverse products (salted snacks, potato chips, yoghurt, cheese, snack bars and fruit; 5–10 samples per study). To indicate their emotional response during product tasting, consumers used the emotion circumplex which is a single-response questionnaire that spans the dimensions of valence (pleasure to displeasure) and arousal (activation to deactivation). Sensory characterisation was obtained using check-all-that-apply (CATA) questions. Through regression analysis, sensory drivers of emotional associations were established, with different linkages to valence and arousal being identified. Many study-specific linkages were found, but the main result pertained to generalised patterns in some linkages including positive associations between ‘sweet’ and “pleasure”, often combined with “pleasant activation”. For ‘sour’, positive linkages were established to “activation”, typically with, or without, “negative pleasure”. Low flavour intensity was positively linked to the emotion circumplex quadrant that connected “deactivation” to “displeasure.” In general, linkages had straightforward and meaningful interpretations, confirming suitability of the emotion circumplex for use in product-focused emotion research.

Heart sound signals can be used for emotion recognition

This article studies whether heart sound signals can be used for emotion recognition. First, we built a small emotion heart sound database, and simultaneously recorded the participants’ ECG for comparative analysis. Second, according to the characteristics of the heart sound signals, two emotion evaluation indicators were proposed: HRV of heart sounds (difference between successive heartbeats) and DSV of heart sounds (the ratio of diastolic to systolic duration variability). Then, we extracted linear and nonlinear features from two emotion evaluation indicators to recognize four kinds of emotions. Moreover, we used valence dimension, arousal dimension and valence-arousal synthesis as evaluation standards. The experimental results demonstrated that heart sound signals can be used for emotion recognition. It was more effective to achieve recognition results by combining the features of HRV and DSV of heart sounds. Finally, the average accuracy of four emotion recognitions on valence dimension, arousal dimension and valence-arousal synthesis was up to 96.875%, 88.5417% and 81.25%, respectively.

Erroneously Disgusted: fMRI Study Supports Disgust-Related Neural Reuse in Obsessive-Compulsive Disorder (OCD).

Objective: fMRI scans of patients with obsessive-compulsive disorder (OCD) consistently show a hyperactivity of the insular cortex, a region responsible for disgust-processing, when confronted with symptom-triggering stimuli. This asks for an investigation of the role of disgust and the insula in OCD patients.Methods: Seventeen inpatients with OCD and 17 healthy controls (HC) underwent fMRI scanning. Whole-brain contrasts were calculated for “Disgust vs. Neutral” for both groups, plus an analysis of variance (ANOVA) to assess the interaction between group and condition. Additionally, the emotional dimensions of valence and arousal, along with the ability to cope, were assessed by picture ratings.Results: The picture ratings confirmed the patients’ heightened sensitivity to disgust with higher values for arousal and inability to cope, but not for valence. fMRI scans revealed no hyperactivity of the insula in patients compared to controls for the condition “Disgust vs. Neutral,” indicating no basic hypersensitivity to disgusting stimuli. Increased activity in the precuneus in controls for this condition might correspond to the downregulation of arousal.Conclusions: The absent differences in neural activity of the insula in patients compared to controls for the disgust-condition, but heightened activity for symptom-provoking conditions, suggests that the illness is due to an erroneous recruitment of the insula cortex for OCD-stimuli. The finding is interpreted within the framework of the neural reuse hypothesis.

An analysis of emotion-exchange motifs in multiplex networks during emergency events

In this paper, we present an analysis of the emotion-exchange patterns that arise from Twitter messages sent during emergency events. To this end, we performed a systematic structural analysis of the multiplex communication network that we derived from a data-set including more than 1.9 million tweets that have been sent during five recent shootings and terror events. In order to study the local communication structures that emerge as Twitter users directly exchange emotional messages, we propose the concept of emotion-exchange motifs. Our findings suggest that emotion-exchange motifs which contain reciprocal edges (indicating online conversations) only emerge when users exchange messages that convey anger or fear, either in isolation or in any combination with another emotion. In contrast, the expression of sadness, disgust, surprise, as well as any positive emotion are rather characteristic for emotion-exchange motifs representing one-way communication patterns (instead of online conversations). Among other things, we also found that a higher structural similarity exists between pairs of network layers consisting of one high-arousal emotion and one low-arousal emotion, rather than pairs of network layers belonging to the same arousal dimension.

Emotionality of Turkish language and primary adaptation of affective English norms for Turkish

Emotional load assessment of the written words has gained considerable interest in psycholinguistics, semantics, and analysis of psychophysiological and electrophysiological correlates of emotional processing. Considering the lack of a publicly available database with affective ratings of contemporary verbal stimuli obtained from native Turkish speakers, we present the affective norms for two datasets of Turkish words carefully adapted from the Affective Norms for English Words (ANEW) database. The valence and arousal ratings are obtained from 61 college-aged participants for 127 highly arousing, emotionally-loaded words in the Adapted Turkish Affective List (ATAL). The ATAL ratings show a tendency of classifying fewer words as positive compared to the original list of stimuli, significantly higher arousal levels for positively rated Turkish stimuli compared to the negative and neutral words, and more congruence in arousal levels of positively exciting words. For the medium to high arousing 508 words in the Expanded Turkish Affective List (ETAL) that cover the whole 9-point spectrum of the valence dimension, 136 Turkish respondents from a wider age, education, and occupation background show higher excitability towards highly unpleasant words. Strong cross-linguistic correlations of + 0.968 between the valence ratings of ANEW and ATAL and + 0.878 for ANEW and ETAL demonstrate the ease of transferring and perceiving the valence levels across English and Turkish. The medium correlation of roughly + 0.450 between the English and Turkish arousal ratings account for lower excitation levels perceived by the native Turkish speakers and indicate the arousal dimension is similar to familiarity and originality in exhibiting more variations between different cultures. These findings demonstrate that this expanded database of partial affective normative ratings can be used as the ground truth for emotional and neurocognitive assessments, and that the presented methodology can be utilized for developing a comprehensive Turkish affective lexicon. The utilized word selection criteria also enable a cross-cultural analysis of adapted words in Turkish and other languages. Detailed normative ratings of this Turkish adaptation are included in the supplementary materials.