Emotion Assessed Through Physiology

Emotion output that can be assessed with physiological methods can be divided into two categories: the somatic changes and changes reflecting autonomic or central nervous system activity. The somatic changes most useful to emotion researchers concern muscle movements associated with emotional expression, particularly those somatic changes on the face.

Measures of somatic change. One useful measurement strategy is to have an observer rate how much emotion a target participant appears to be feeling, based on expressive cues. The observers might be "experts" on the target person's emotional experiences (e.g., a spouse or a therapist). One limitation is that observer reports are based on social attributions of a target's emotional state, and such attributions will be limited by the information available, biased by a target's impression management strategies, or even influenced by the raters' own level of emotion being rated (Marcus & Miller, 1999). As such, observer ratings of emotion are probably best used in combination with other measures. One way to limit attributions is to use trained observers. A standardized training system for observers is the Specific Affect Coding System (SPAFF; Coan & Gottman, in press; Gottman & Krokoff, 1989; Gottman & Levenson, 1992; for a brief review, see Gottman, 1993). This system separates expressed emotion into specific categories of positive and negative categories (e.g., interest, affection, humor, validation, excitement/joy, anger, belligerence, domineering, contempt, disgust, tension, sadness, whining, and defensiveness). SPAFF training involves recognizing and attending to important facial, gestural, and vocal markers of emotion. Significant benefits of observer ratings are that they can be unobtrusive, can be used in naturalistic settings, are inexpensive and fast, and can provide emotion measures from a few visible cues.

Some somatic coding systems are based on specific observable changes in facial muscles. One such system for coding emotion in the face is the Facial Action Coding System (FACS; Ekman & Friesen, 1975, 1978). The FACS consists of 46 anatomically based "action units" (or AUs), which refer to a specific observable change in the face. For example, AU 1 raises the inner brows, AU 9 wrinkles the nose, and AU 12 raises the outer lip corners. The system requires extensive training and certification for reliable use (cf. Ekman & Friesen, 1975, 1978). A drawback of FACS is the extensive amount of time needed to code expressions. FACS scoring requires about 1 hour of coding for each minute of videotape (depending on the density of facial action). Researchers are developing computer vision to undertake the tiresome task of facial action coding. One of the more advanced systems is that being developed at Carnegie Mellon University under the guidance of Jeffry Cohn (see Cohn, Zlochower, Lien, & Kanade, 1999), which is able to accurately code approximately half of the FACS action units in real time. Alternatively, Ekman and others have developed more global coding systems, which are based on fewer AUs, for coding facial action (e.g., EMFACS by Ekman & Friesen, see Fridland, Ekman, & Oster, 1986; MAX by Izard, 1979).

Somatic facial assessments may also be obtained using physiological measures of muscle contractions. The neural activation of the muscles produces action potentials that can be directly measured on the surface of the skin using electromyography (EMG) using two electrodes placed over the muscle of interest. The amount of electrical activity detected is directly proportional to the magnitude of contraction. Detailed descriptions of facial electromyographic technique may be found in Cacioppo, Petty, Losch, and Kim (1986). EMG is able to assess muscular contractions that are too small to produce visible changes (i.e., not FACS codable; Cacioppo et al., 1986). Such sensitivity has a disadvantage, however, in that electrical signals from sites other than the muscle of interest may also be detected during EMG assessments. Researchers interested in measuring emotions with facial EMG should have training in electrophysiological technique or collaborate with someone with such expertise.

Physiological measures of nervous system activity associated with emotion. Emotions are closely tied to tendencies to act in specific ways, and changes in the nervous system occur primarily to support these actions (Frijda, 1986; Lazarus, 1991). In terms of the autonomic nervous system (ANS), a few researchers hold the view that distinct emotions are associated with distinct ANS activity (e.g., Levenson, Ekman, & Friesen, 1990). Empirical support for specific autonomic patterns being associated with specific emotions has been obtained in several studies. However, the cumulative data on specific emotional "signatures" are mixed and therefore remain inconclusive (for reviews, see Cacioppo & Gardner, 1999; Cacioppo, Klein, Berntson, & Hatfield, 1993; Levenson, 1992; Zajonc & Mcintosh, 1992).

Diverse autonomic measures have been used to assess emotion, some more fruitfully than others. We will mention here only a couple of the more promising measures and advise the interested reader to consult Cacioppo, Tassinary, and Berntson (2000); Stern, Davis, and Ray (1992); or Hugdahl (1996) for more details. Electrodermal activity, especially skin conductance, is a widely accepted and reliable measure used in emotion research. Another category of measures is based on respiratory activity. Perhaps the largest category of measures is those based on cardiovascular activity. This last set includes measures such as heart rate, diastolic and systolic blood pressure, cardiac output, stroke volume, and total peripheral resistance. Readers interested in cardiac measures should consult Sherwood (1993) and Sherwood et al. (1990) for details on impedance cardiography. Other researchers assess the link between respiratory and cardiovascular activity (e.g., respiratory sinus arrhythmia or heart rate variability), which appears related to emotion state (Grossman, van Beek, & Wientjes, 1990; Porges, 1995). It should be noted that professional polygraphers typically employ a multimethod approach, using measures of skin conductance, respiration, and heart rate to infer the emotion of guilt.

Researchers have recently begun to refine central nervous system measures of emotion. Scalp-recorded brain electrical activity, or electroencephalogram (EEG), has been used successfully to distinguish pleasant and unpleasant emotion states (e.g., Schmidt & Trainor, 2001), as well as individual differences in affective style (for a review, see Davidson, 1993). Other more localized imaging measures of emotion-related changes in the brain are on the horizon as well, including functional MRI (for an overview, see Berthoz, Blair, Le Clec'h, & Martinot, 2002). The versatility of functional imaging methods for studying mechanisms of emotion is significant, given its superior spatial resolution (Mayberg & McGinnis, 2000). The temporal resolution is not as good as EEG measures, however.

Many practical issues emerge when contemplating the use of physiological measures. First, these measures are typically invasive. Some measures (e.g., being inserted into a large MRI magnet) might elicit emotions (e.g., panic) themselves. Less-invasive measures are pulse rate and skin conductance. Impedance cardiography uses metal bands that cir cle a participant's neck and chest in several locations. Attaching these requires participants to partially disrobe. Blood pressure assessment typically uses pressurized cuffs that, when inflated, draw attention and sometimes even cause pain. Physiological measures also usually restrict participants' mobility because of wires that connect them to amplifiers and recording devices. Bodily movement can also create artifacts in measurement. In general, the use of physiological assessments requires special efforts on the part of both the researcher and the participants, but may potentially pay off with a unique methodological perspective or window on the emotional state under investigation.

Was this article helpful?

0 0

Post a comment