Summary and Conclusions

The hypothesis that lateralization of asymmetries across subjects would change with the action was confirmed. Analysis of variance revealed a main effect of different Action Units on bilateral asymmetry and better unilateral control scores. Bilateral asymmetry, better unilateral control, and preferred side scores showed that asymmetries of some actions were lateralized, but those of other actions showed no tendency towards laterality. Rated easier scores did not show laterality for any individual action. According to bilateral asymmetry and better unilateral control measures, laterality of some deliberate actions was opposite. Blinking and winking actions were sometimes lateralized right as predicted, but more often showed little asymmetry. Several actions predicted to have left laterality had right laterality instead. Two actions had left laterality as predicted.

The four different measures of laterality reflected common variance as predicted, but also appeared to measure some different variables. All the significant correlations between measures were positive, but the number of significant correlations was only slightly over half the total. Some measures correlated better with each other than with other measures. As predicted, the measures usually showed the same pattern of laterality for a given action, but not always.

There were two sources of evidence supporting the hypothesis that asymmetries in the strength of actions are related to factors involving the control of the actions. First were the significant, positive correlations between the measure of bilateral asymmetry in the strength of contraction and the other measures which more clearly reflected control over actions. The hypothesis that the unilateral action on the better side would be accompanied by fewer unrequested actions than the unilateral on the poorer side was confirmed, supporting the interpretation that this measure reflects control over facial actions. Second, a more detailed analysis of the strength of unilateral actions showed that the onside (requested side) contraction was stronger for the better unilateral than the poorer action. Conversely, the offside (unrequested side) contraction was stronger for the poorer unilateral action.

The hypothesis that spontaneous and deliberate actions would differ in asymmetry was confirmed. Deliberate actions were generally more asymmetrical than spontaneous actions. Deliberate individual actions of orbicularis oculi (crowsfeet wrinkles) and zygomatic major (smiling) were more asymmetrical than the spontaneous actions of these muscles in response to a humorous comment. Deliberate individual actions of orbicularis oculi (squinting) and risorius (lip stretching) were more asymmetrical than the spontaneous actions of these muscles in response to a startling noise. AUs 6, 45 (blinking), and the eye closure measure did not show a difference between deliberate and startle conditions.

A second difference between the asymmetry of spontaneous and deliberate actions was in lateralization across subjects. As predicted, the asymmetries that were observed in spontaneous actions were generally not lateralized, unlike asymmetries of deliberate actions. Only asymmetries in the action of orbicularis oculi in the startle reaction were lateralized.

Actions from conditions that mixed spontaneous and deliberate types of movement showed asymmetry and laterality intermediate between actions from conditions that elicited a purer sample of these two types of actions. Actions in emotion and startle simulations showed less asymmetry than requests for deliberate individual actions, but more asymmetry than spontaneous actions, but few of the differences were significant. Lateralization of asymmetries of simulated actions showed the same tendencies as deliberate actions, but these tendencies were not as pronounced.

The hypothesis that actions from startle conditions where the subject knew that the noise would occur would be more asymmetrical than actions from the unanticipated startle was not confirmed. The few differences in asymmetry between the startle conditions can easily be attributed to chance.

The asymmetry of individually requested actions did not depend upon whether they were requested verbally or visually. There were no significant differences between these conditions.

The major implication of these results is to call into question every explanation of laterality of facial actions that has been proposed previously. Laterality is specific to particular deliberate facial actions, and the pattern of laterality across actions does not entirely fit any theoretical model. Models of hemispheric specialization for emotional processes, whether specialization by only the right hemisphere or by dual specialization for positive or approach emotions versus negative or avoidance emotions, cannot account for the asymmetries and laterality observed here. Models that attribute asymmetry to processes involving the directed control of actions and to cognitive, deliberative processes appear to be more compatible with the evidence from this study, but they cannot explain all the findings. It may be that asymmetry is produced by a complex interaction of different processes and variables or by factors that await future conceptualization.

The findings of this study present a considerable challenge to researchers using asymmetries in facial muscle action as indices of hemispheric specialization. The four measures used in this study measured common variance but also tapped different variables affecting asymmetry. Only ratings and preferred side scores appeared to measure the same variable and showed any consistency of laterality across actions, but this laterality was weak. Although measures of asymmetry in bilateral movements and better unilateral actions showed more laterality, they also indicated that some other factor besides specialization of one process in a single hemisphere influenced asymmetry. More disparate measures from those used in this study could reflect other aspects of asymmetry in facial function. Researchers who use only one measure of asymmetry need to show how well it reflects hemispheric specialization. Multiple measures of asymmetry in facial function that carefully specify what aspect of asymmetry is being assessed could help to know what is being measured. Researchers who study only one muscle will not be able to generalize easily to other muscles. Likewise, measurements of the average activity of a group of muscles (e.g., with EMG) may not be applicable to particular muscles. Several representative muscle actions should be measured individually, carefully separating out the influence of other muscle actions or extraneous factors such as permanent facial features that could affect the measurement.

The results of this study, in combination with those of previous studies, indicate that the phenomenon of asymmetry in facial actions is one of considerable complexity demanding rigorous attention to control of many variables. Characteristics of the subject should be controlled, including gender, age, and handedness. An important variable to consider is the type of movement studied, whether the actions are spontaneous or deliberate, emotional, reflex, etc. This study has investigated asymmetry of deliberate actions thoroughly, and much evidence indicated that the asymmetry of these deliberate actions differs from reflex startle actions and positive emotional actions. The effort needed now is to better understand spontaneous emotional expressions, especially negative expressions. Such a study would face the difficult task of eliciting uncontrolled expressions of specific emotions, but it might resolve questions about the differential role of the hemispheres in emotion.