مکاتبات هیجانی خودارزیابی با حالت چهره
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|37579||1999||10 صفحه PDF||سفارش دهید||5272 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Psychiatry Research, Volume 86, Issue 2, 31 May 1999, Pages 175–184
Abstract Emotional processing abilities are difficult to measure psychometrically. Ultimately their quantification has to rely on `subjective' judgment thereby leaving open the problem of response biases. Assessments of autonomic arousal similarly provide a mere unspecified measurement of a specific emotion. A standardized mood induction procedure capable of obtaining reliable happy and sad mood changes in healthy subjects was used to demonstrate the effectiveness of this procedure. We performed a two-part experiment using a rater-based analysis of facial expressions. This entailed analyzing the emotion portrayed in the faces. The faces of 24 healthy subjects were videotaped during the mood induction procedure of happiness and sadness, respectively. A group of 20 raters naive to the experimental task and conditions rated the facial expressions on six basic emotions. Results showed that ratings corresponded with the facial expressions, which were reflecting the mood of the task condition. Subjects' facial expressions together with self-ratings demonstrate the successful applicability of this standardized mood induction procedure for eliciting happy and sad mood.
Introduction The ability to experience, express and discriminate emotions is fundamental for living organisms. However, emotional processing abilities are difficult to measure psychometrically compelling their quantification to rely on `subjective' judgment. There exists a large number of possibilities for registering a subject's experienced affective state. Verbal descriptions of an emotion, rating scales, standardized checklists, and questionnaires are but a few possibilities. Because emotion research does not at present propose any preferential measuring technique, and mood experience is inherently a subjective event that depends on self-reports for its verification, self-ratings appear to be most feasible in assessing emotional experience despite susceptibility to demand characteristics and other response biases (Berkowitz and Troccoli, 1986). There likewise exist a number of mood induction procedures, ranging from recall of personal events to presentation of emotional material. The Velten procedure (Velten, 1968), for example, which relies on self-referent emotional statements, is disadvantaged by its reduced applicability: 30–50% of the subjects fail to show the intended mood induction effect (Gerrards-Hesse et al., 1994). Emotions can also be induced by the presentation of visual material, particularly pictures containing emotional content. Lang et al. (1988)introduced the International Affective Picture System which consists of colored slides of various kinds of emotional and neutral stimuli. This method of mood induction appears to have been evaluated most thoroughly with respect to its effectiveness; each slide has been rated on the dimensions of valence, arousal, and dominance by many raters of different cultural background. In addition, extensive data exist from studies with simultaneous measures of physiological and behavioral indices (Lang et al., 1990 and Lang et al., 1993; Bradley et al., 1993). Film clips are often used to induce changes in mood and attempts are made to find a standardized set (Philippot, 1993; Gross and Levenson, 1995) capable of producing predictable emotions. However, film clips often elicit emotional blends and their social relevance is equivocal, e.g. individuals respond differentially to different films, especially across cultures. This material is often not comparable with respect to number of actors, brightness of visual stimuli, degree of verbal communication, music, loudness, etc. Confounding the issue is the difficulty film clips pose when applied in neuroimaging studies. Taken together, these problems may produce uncontrollable activations in various cerebral regions. Sokolowski (1992)summed up the problems and aims in experimental mood induction procedures as follows: first, the induced emotion should be of sufficient intensity and authenticity to subjects. Only then are studies on emotion able to investigate consequences of emotions on interesting dependent variables, i.e. cognition and behavior. Second, subjects should be neither conscious of their mood changes nor of the purpose of the study so as to prevent them from controlling their behavior and cognitions or from developing certain expectations. In addition, guidelines in experimental research for methods, fulfillment of criteria for economics, repeatability, comparability, and ethics ought to be considered. The lack of sufficiently reliable, valid, and comparable mood induction methods comprising controlled stimulus material led us to develop our mood induction procedure. This procedure included ecologically valid and socially relevant emotional stimuli, which are controlled, comparable (Schneider et al., 1994a), and simultaneously applicable in neuroimaging. Straight angle monochromatic photographs of happy and sad facial expressions varying in intensity made up the stimuli (for details, see Erwin et al., 1992). This method demonstrated small intraindividual variability and high retest reliability. In general, making subjective ratings more objective requires the inclusion of a number of criteria. Evidence for the validity and objectivity of this standardized mood induction procedure comes from its already successful employment in several neuroimaging studies. Characteristic valence-specific regional cerebral and autonomic effects have been demonstrated measuring regional brain activity with a 133Xenon clearance method, H2O15 Positron Emission Tomography (PET), and functional Magnetic Resonance Imaging (fMRI; Schneider et al., 1994b, Schneider et al., 1995a, Schneider et al., 1997 and Schneider et al., 1998). Its successful employment has also been demonstrated in schizophrenic patients (Schneider et al., 1995b and Schneider et al., 1998). The aim of this study was to help in validating our standardized mood induction procedure. We set out to demonstrate the effectiveness of mood induction at a behavioral level taking into account its already established success at a subjective and autonomic level. Mood changes should, with this procedure, become visible in facial expressions of subjects and for raters (judging the facial expressions) who are naive to the task and who are not explicitly asked to look for happiness and sadness in faces. Raters independently judged the facial expressions according to different emotional dimensions. An indication of successful happy and sad mood induction should be reflected in the facial expressions. Our intent was not to establish a new method for analyzing facial expressions but rather to use a procedure of facial ratings merely to show the validity of our mood induction procedure. Controversy exists as to whether gender plays an influential role in emotional experiences or whether divergent results of males and females are attributable to normative social sex roles (Grossman and Wood, 1993). Some studies support the findings of neurobiological-based gender differences in emotion processing (Gur et al., 1995; George et al., 1996; Kring and Gordon, 1998). Women have been found to be more expressive than men (McConatha et al., 1997; Asthana and Mandal, 1998; Kring and Gordon, 1998). The finding that facial expressions vary between men and women (Alford, 1983; Frisch, 1995) may seem to reflect a neurobiological gender difference. To investigate the possibility of gender influences on subjective ratings, facial expressions, and ratings of facial expressions, gender was introduced as an interesting variable.
نتیجه گیری انگلیسی
Results 3.1. Emotion discrimination/age discrimination Subjects solved 95.42% (±5.96) of the emotion discrimination items. Sensitivity and specificity for happy were 0.99 (±0.04) and 1.00 (±0.00), for sad 1.00 (±0.01) and 0.92 (±0.10), respectively. Positive bias was 0.2% (±1.02), and negative bias 8.3% (±9.52). Percent correct scores of males and females revealed no significant difference (t=1.08; P=0.3). Performance relative to age discrimination task was 0.60 (±0.19). Males and females revealed no difference in this task (t=0.27; P=0.79). 3.2. Mood induction The three-factor ANOVA (gender, mood induction, rating scale), with mood induction and rating scale as repeated measures factors, demonstrated a significant interaction between mood induction and rating scale (F=20.46; d.f.=1,22; P=0.0002), indicating that subjects' self-ratings of emotion changed during each condition ( Table 1). Hence, more positive affect during the happy mood induction compared to the sad mood induction was shown. Likewise, more negative affect was present during sad and less negative affect during happy mood induction ( Fig. 1). Similar results were obtained for the ESR scores showing a significant interaction between mood induction×rating scale (F=76.85; d.f.=1,22; P=0.0001). No gender differences could be observed with respect to mood induction effect in either rating scale. The mood induction effect calculated for the ESR was a strong one, with an average value of 3.00 (±1.69). No gender differences (t=1.22; P=0.24) were found. Table 1. Self-ratings Happiness Sadness Control Baseline prior to mood induction PANAS Positive 29.17 (±5.56) 25.13 (±6.65) 27.00 (±6.37) 29.87 (±5.49) (17–41) (15–39) (17–40) (15–39) Negative 12.79 (±3.64) 17.33 (±6.38) 12.88 (±3.72) 14.45 (±5.21) (10–23) (10–33) (10–26) (10–30) ESR Happiness 2.92 (±0.93) 1.46 (±0.66) 2.04 (±0.81) 2.58 (±0.8) (1–5) (1–3) (1–4) (1–4) Sadness 1.46 (±0.78) 3.00 (±1.02) 1.33 (±0.76) 1.41 (±0.15) (1–4) (1–5) (1–4) (1–4) Anger 1.25 (±0.68) 1.67 (±0.92) 1.42 (±0.83) 1.50 (±0.22) (1–3) (1–4) (1–4) (1–5) Disgust 1.04 (±0.20) 1.33 (±0.87) 1.00 (±0.00) 1.12 (±0.09) (1–2) (1–5) (1–1) (1–3) Fear 1.04 (±0.20) 1.38 (±0.77) 1.21 (±0.59) 1.41 (±0.15) (1–2) (1–4) (1–3) (1–4) Surprise 2.21 (±0.93) 1.75 (±0.90) 2.04 (±1.04) 2.29 (±0.22) (1–4) (1–4) (1–4) (1–4) Mean values and standard deviations of emotional self-ratings (PANAS, ESR) followed by the range of 24 healthy subjects in the separate conditions. Table options Comparison of the two ratings: means and standard errors of self-reports of the ... Fig. 1. Comparison of the two ratings: means and standard errors of self-reports of the emotion experienced subjectively (ESR) and ratings of the facial expressions (EFR) for 24 subjects. Figure options Subjects had a mean duration of 3.97 min (±2.47) for viewing all slides per condition. A one-way analysis of variance (ANOVA) with gender as a factor and condition as a repeated measures factor showed a significant main effect for condition (F=12.23; d.f.=2,44; P=0.001). The required time for the separate conditions varied significantly: Subjects needed 3.82 (±1.62) min on average for happy mood induction, 5.39 (±3.50) min for sad, and 2.69 (±0.31) min for gender differentiation. No gender differences could be observed. Possible different effects of the stimulus material accounting for the differences in viewing times between the happy and sad condition were controlled. Happy and sad facial expressions did not significantly vary in the different ratings (five-point unipolar intensity scale) for five basic emotions, which were applied while constructing the task. The difference was in the intended emotion (high ratings for happiness in the slides for the happy mood induction, high ratings for sadness in the slides for the sad mood induction). Sad facial expressions of the mood induction task had a mean sadness rating of 1.85±0.63 (fear 0.80±0.39, happiness 0.15±0.41, surprise 0.36±0.12, anger 0.55±0.31, disgust 0.53±0.25). Happy facial expressions had a mean happiness rating of 2.41±0.64 (fear 0.12±0.13, sadness 0.07±0.09, surprise 0.50±0.17, anger 0.05±0.07, disgust 0.06±0.09). Therefore we assume that the sad faces were neither more surprising nor more puzzling to subjects than the happy faces. For the gender identification task, subjects had a mean of 97.76±3.11% correct responses and made 1.79±2.48 errors on average. 3.3. Analysis of facial expression Comparison of mean ratings for all emotional dimensions showed the highest ratings for sadness and happiness in the corresponding mood induction conditions, and lower ratings for the other emotions (Table 2). An equivalent three-factor repeated measures analysis of variance for ratings of happiness and sadness was performed and showed two significant main effects for mood induction (F=15.2; d.f.=1,22; P=0.0008), and rating scale (F=11.5; d.f.=1,22; P=0.0001), and a significant interaction between mood induction×rating scale (F=54.80; d.f.=1,22; P=0.0001, Fig. 1). No gender differences were found, thus indicating no differential ratings of male and female raters. Table 2. Emotional face ratings (EFR) Happiness Sadness Control EFR Happiness 2.78 (±0.95) 1.20 (±0.38) 1.74 (±0.73) (2.7–4) (1–2.6) (1–3.6) Sadness 1.15 (±0.21) 1.91 (±0.71) 1.48 (±0.57) (1–1.6) (1–3.4) (1–3.4) Anger 1.14 (±0.21) 1.30 (±0.32) 1.30 (±0.40) (1–1.8) (1–2.2) (1–2.8) Disgust 1.10 (±0.21) 1.14 (±0.25) 1.19 (±0.22) (1.1–1.8) (1–2) (1–1.8) Fear 1.06 (±0.12) 1.16 (±0.18) 1.25 (±0.27) (1.05–1.4) (1–1.6) (1–1.8) Surprise 1.54 (±0.53) 1.24 (±0.31) 1.63 (±0.68) (1.0–2.6) (1–2) (1–3.0) Expressiveness 2.71 (±0.81) 2.08 (±0.62) 2.33 (±0.67) (1.2–4.2) (1–3.6) (1.2–4.0) Judgments of the facial expressions of 24 subjects with respect to seven emotional dimensions averaged followed by range for the 20 raters (Mean±S.D.). Table options The resulting ratings confirmed the effectiveness of this mood induction method. The ratings of the facial expressions for sadness during the sad mood induction were higher than the ratings for happiness. Similarly, ratings for sadness were less than ratings for happiness during the happy mood induction. The raters were able to infer the emotional state of the subjects' facial expressions taped during mood induction. The facial expressions corresponded with the emotional state demanded by the task condition. Negative affect during sad mood induction was judged by raters as less sad compared to subjects' self-ratings for sadness (Fig. 1). Ratings for happiness and sadness were significantly different in the sad vs. neutral condition (thappy=3.14, P=0.005; tsad=−2.81, P=0.01). The same went for the comparison of the happy and neutral condition (thappy=5.93, P=0.0001; tsad=−2.53, P=0.02).