Arvid Kappas

Effects of Dynamic Aspects of Facial Expressions: A Review

A key feature of facial behavior is its dynamic quality. However, most previous research has been limited to the use of static images of prototypical expressive patterns. This article explores the role of facial dynamics in the perception of emotions, reviewing relevant empirical evidence demonstrating that dynamic information improves coherence in the identification of affect (particularly for degraded and subtle stimuli), leads to higher emotion judgments (i.e., intensity and arousal), and helps to differentiate between genuine and fake expressions. The findings underline that using static expressions not only poses problems of ecological validity, but also limits our understanding of what facial activity does. Implications for future research on facial activity, particularly for social neuroscience and affective computing, are discussed.

Sentiment in short strength detection informal text

A huge number of informal messages are posted every day in social network sites, blogs, and discussion forums. Emotions seem to be frequently important in these texts for expressing friendship, showing social support or as part of online arguments. Algorithms to identify sentiment and sentiment strength are needed to help understand the role of emotion in this informal communication and also to identify inappropriate or anomalous affective utterances, potentially associated with threatening behavior to the self or others. Nevertheless, existing sentiment detection algorithms tend to be commercially oriented, designed to identify opinions about products rather than user behaviors. This article partly fills this gap with a new algorithm, SentiStrength, to extract sentiment strength from informal English text, using new methods to exploit the de facto grammars and spelling styles of cyberspace. Applied to MySpace comments and with a lookup table of term sentiment strengths optimized by machine learning, SentiStrength is able to predict positive emotion with 60.6p accuracy and negative emotion with 72.8p accuracy, both based upon strength scales of 1–5. The former, but not the latter, is better than baseline and a wide range of general machine learning approaches.

Emotion and Regulation are One

Emotions are foremost self-regulating processes that permit rapid responses and adaptations to situations of personal concern. They have biological bases and are shaped ontogenetically via learning and experience. Many situations and events of personal concern are social in nature. Thus, social exchanges play an important role in learning about rules and norms that shape regulation processes. I argue that (a) emotions often are actively auto-regulating—the behavior implied by the emotional reaction bias to the eliciting event or situation modifies or terminates the situation; (b) certain emotion components are likely to habituate dynamically, modifying the emotional states; (c) emotions are typically intra- and interpersonal processes at the same time, and modulating forces at these different levels interact; (d) emotions are not just regulated—they regulate. Important conclusions of my arguments are that the scientific analysis of emotion should not exclude regulatory processes, and that effortful emotion regulation should be seen relative to a backdrop of auto-regulation and habituation, and not the ideal notion of a neutral baseline. For all practical purposes unregulated emotion is not a realistic concept.

An analysis of the encoding and decoding of spontaneous and posed smiles: The use of facial electromyography

Twenty subjects judged 80 video segments containing brief episodes of smiling behavior for expression intensity and happiness of the stimulus person. The video records were produced under instructions to (a) pose, (b) experience a happy feeling or (c) to both experience and show a happy feeling. An analysis of the integrated facial electromyogram (EMG), recorded over four muscle regions (zygomaticus major, depressor anguli oris, corrugator supercilii, andmasseter), showed that judgments of happiness and of intensity of expression could be predicted in a multiple regression analysis (multipleR = .64 for perceived happiness and .79 for perceived expression intensity). The perception of happiness was affected by EMG activity in regions other thanzygomaticus major. The use of parameters other than the mean of the integrated EMG, namely variance, skewness, kurtosis and properties of the amplitude distributions across time, provided accurate classification of the elicitation conditions (pose happiness versus experience happiness) in a discriminant analysis. For the discrimination of posed and felt smiles variables describing aspects of facial activity in the temporal domain were more useful than any of the other measures. It is suggested that facial EMG can be a useful tool in the analysis of both the encoding and decoding of expressive behavior. The results indicate the advantage of using multiple-site EMG recordings as well as of using amplitude and temporal characteristics of the facial EMG measures.

Towards Empathic Virtual and Robotic Tutors

Building on existing work on artificial tutors with human-like capabilities, we describe the EMOTE project approach to harnessing benefits of an artificial embodied tutor in a shared physical space. Embodied in robotic platforms or through virtual agents, EMOTE aims to capture some of the empathic and human elements characterising a traditional teacher. As such, empathy and engagement, abilities key to influencing student learning, are at the core of the EMOTE approach. We present non-verbal and adaptive dialogue challenges for such embodied tutors as a foundation for researchers investigating the potential for empathic tutors that will be accepted by students and teachers.

Predicting Emotional Responses to Long Informal Text

Most sentiment analysis approaches deal with binary or ordinal prediction of affective states (e.g., positive versus negative) on review-related content from the perspective of the author. The present work focuses on predicting the emotional responses of online communication in nonreview social media on a real-valued scale on the two affective dimensions of valence and arousal. For this, a new dataset is introduced, together with a detailed description of the process that was followed to create it. Important phenomena such as correlations between different affective dimensions and intercoder agreement are thoroughly discussed and analyzed. Various methodologies for automatically predicting those states are also presented and evaluated. The results show that the prediction of intricate emotional states is possible, obtaining at best a correlation of 0.89 for valence and 0.42 for arousal with the human assigned assessments.

Appraisals are direct, immediate, intuitive, and unwitting…and some are reflective…

In 1960, Magda Arnold defined the appraisal construct as being causal to emotion. Appraisal, according to her, refers to a direct, immediate, and intuitive process that does not initially require recognition of the object that is being appraised. It is based on phylogenetically ancient subcortical brain structures. In addition, Arnold proposed the existence of a related conscious process, also referred to as appraisal, that interacts with the direct appraisal and is responsible for a differentiation of emotional states. This theory was revolutionary and is still to be considered modern because it already comprised much of what is currently hypothesised to be causal for emotion. However, Lazarus, who initially believed that conscious aspects of appraisal are more important than implicit and unaware appraisals, coined the term cognitive appraisals and popularised this notion widely. In opposition to what he perceived as a prevailing emphasis on higher cognitive processes in emotion theory, Zajonc (1980) argued that emotion elicitation does not depend on conscious cognition. I argue that Arnolds theory is in fact completely consistent with Zajoncs view and data. The concept of appraisal should be discussed in relation to Arnolds original intention, because it provides not only the basis of an integrated view of multiple levels of emotional processing, encompassing views espoused by Zajonc and by Lazarus, but may guide current and future research on multiple levels of processing in the elicitation of emotions.

Social regulation of emotion: messy layers

Emotions are evolved systems of intra- and interpersonal processes that are regulatory in nature, dealing mostly with issues of personal or social concern. They regulate social interaction and in extension, the social sphere. In turn, processes in the social sphere regulate emotions of individuals and groups. In other words, intrapersonal processes project in the interpersonal space, and inversely, interpersonal experiences deeply influence intrapersonal processes. Thus, I argue that the concepts of emotion generation and regulation should not be artificially separated. Similarly, interpersonal emotions should not be reduced to interacting systems of intraindividual processes. Instead, we can consider emotions at different social levels, ranging from dyads to large scale e-communities. The interaction between these levels is complex and does not only involve influences from one level to the next. In this sense the levels of emotion/regulation are messy and a challenge for empirical study. In this article, I discuss the concepts of emotions and regulation at different intra- and interpersonal levels. I extend the concept of auto-regulation of emotions (Kappas, 2008, 2011a,b) to social processes. Furthermore, I argue for the necessity of including mediated communication, particularly in cyberspace in contemporary models of emotion/regulation. Lastly, I suggest the use of concepts from systems dynamics and complex systems to tackle the challenge of the “messy layers.”

Damping sentiment analysis in online communication: discussions, monologs and dialogs

Sentiment analysis programs are now sometimes used to detect patterns of sentiment use over time in online communication and to help automated systems interact better with users. Nevertheless, it seems that no previous published study has assessed whether the position of individual texts within on-going communication can be exploited to help detect their sentiments. This article assesses apparent sentiment anomalies in on-going communication --- texts assigned significantly different sentiment strength to the average of previous texts --- to see whether their classification can be improved. The results suggest that a damping procedure to reduce sudden large changes in sentiment can improve classification accuracy but that the optimal procedure will depend on the type of texts processed.

Face-to-face communication over the internet : emotions in a web of culture, language and technology

Social platforms such as MySpace, Facebook and Twitter have rekindled the initial excitement of cyberspace. Text based computer-mediated communication has been enriched with face-to-face communication such as Skype, as users move from desk tops to laptops with integrated cameras and related hardware. Age, gender and culture barriers seem to have crumbled and disappeared as the user base widens dramatically. Other than simple statistics relating to e-mail usage, chatrooms and blog subscriptions, we know surprisingly little about the rapid changes taking place. This book assembles leading researchers on non-verbal communication, emotion, cognition and computer science to summarize what we know about the processes relevant to face-to-face communication as it pertains to telecommunication, including video-conferencing. The authors take stock of what has been learned regarding how people communicate, in person or over distance, and set the foundations for solid research helping to understand the issues, implications and possibilities that lie ahead.