SUBJECTIVE MOOD STATE AND PERCEPTION OF EMOTION IN CHIMERIC FACES Lauren Julius Harris! and Peter J. Snyder eMichigan State University; 2Long Island Jewish Medical Center)
INTRODUCTION
Clinical and experimental research has established that the right cerebral hemisphere plays a leading role in the processing and mediation of visuospatial and other non-verbal functions. The research indicates that this lateralization pattern applies strongly in the case of right-handers. In left-handers, the lateralization pattern, although in the same direction, is typically weaker and more variable, possibly reflecting a relatively greater measure of bilaterality of functional organization in the form of secondary left-hemisphere commitment to visuospatial processing (e.g., Borod, Carper, Naeser et al., 1985). The clinical research also suggests that only in rare instances do left-handers show what amounts to true reversed dominance for visuospatial functions (e.g., Dronkers and Knight, 1988). Despite the evidence suggesting relative uniformity of lateralization for normal right-handers, there also is much evidence of individual differences in this same group on tests of lateralized functioning. For example, among righthanded college students, Levy, Heller, Banich and Burton (1983) found striking differences in both the strength and direction of the visual hemifield advantage for tachistoscopically-presented non-verbal stimuli as well as for stimuli presented for free-viewing. The latter were mirror-image pairs of "chimeric faces" , each with a smile on one side, a frown on the other. Asked to choose the "happier" chimeric face of each pair, most subjects consistently made their choice based on the emotional cue (the smile) located to their left (in their left visual hemifield, or L VH), but a significant minority just as consistently made their choice based on the emotional cue located to their right (right visual hemifield, or RVH). Believing it unlikely that such individual differences in a presumably modal lateralization group (right-handed adults) reflected actual differences in the direction of lateral specialization for visuospatial functions, Levy et al. proposed that they instead were reflections of differences in characteristic patterns of hemispheric arousal, or activation, regardless of the task. More specifically, they proposed that people vary in their characteristic patterns of arousal, with some showing more left-hemisphere arousal and others showing more right-hemisphere arousal, so that when that arousal tendency is commensurate with the hemispheric processing elicited by a particular task, it serves to enhance the overall asymmetry in the direction of the specialized hemisphere. Cortex, (1992) 28, 471-481
472
L.J. Harris and P.J. Snyder
When the arousal tendency opposes the hemisphere processing elicited by the task, it attenuates the overall asymmetry in the direction of the specialized hemisphere. Inasmuch as most people show a L VH bias on the Chimeric Faces Test, we therefore can conclude that the test elicits right-hemisphere specialized processing and that, superimposed on that bias, is a characteristic arousal bias that acts either to strengthen or to attenuate the bias (for further discussion, see Heller, 1990). If there are such differences in patterns of characteristic hemispheric arousal, what is their origin if not in actual differences in lateral cortical specialization? One possible basis may lie in individual differences in characteristic emotional, or mood, state insofar as such differences might reflect differences in direction of hemispheric arousal. One immediate complication is that the neuropsychological evidence is inconsistent on the question of hemispheric specialization for emotion. Some evidence, with both clinical and normal populations, points to the primacy of the right hemisphere for all emotional functions (e.g., Flor-Henry, 1979; Levy et aI., 1983; Ley and Bryden, 1979), whereas other evidence suggests that each hemisphere is specialized to process different types of emotion depending on valence, the left hemisphere for positive emotions, the right hemisphere for negative emotions (e.g., Davidson, 1984; Sackeim, Greenberg, Weiman et aI., 1982). Much of the evidence supporting a valence effect, however, comes from studies of the subject's own emotional expression or feeling (e.g., studies of emotional reactions in patients with either left- or righttemporal lobe epileptogenic foci, cf. Bear and Fedio, 1977; review by Pritchard, 1983), whereas the evidence supporting the position of right-hemisphere primacy for all emotions more often comes from studies of the perception of emotion (e.g., recognition of emotion in tachistoscopically-projected faces; for reviews, see Benson and Zaidel, 1986; Heilman and Satz, 1983; Heller, 1990). One way to reconcile the two bodies of evidence is to suppose that the right hemisphere mediates the perception, or recognition, of emotion, whereas each hemisphere takes somewhat different responsibilities for the expression of emotion (cf. Heller, 1990). If, then, we credit the hypothesis linking valence of subject mood with a particular cerebral hemisphere, it raises the question whether subject mood state will affect performance on a perceptual task such as the Chimeric Faces Test. Some support for the general hypothesis linking performance with mood has been reported in a clinical population of depressed patients (Jaeger, Borod and Peselow, 1987). Like normal controls, the patients showed a left hemifield bias but to a weaker extent, consistent with independent evidence from studies of unipolar depressive patients that show diminished right (relative to left) electroencephalogram amplitude (D'Elia and Perris, 1973) and amplitude of skin conductance for an orienting response (Gruzelier and Venables, 1974). Accordingly, we would not go so far as to suppose that normal persons whose mood states are negative would show a complete reversal of the usual L VH bias on Chimeric Faces. The reason is that this supposition implies that in most normal subjects, the typical mood state is negative. We already know, however, that most normal persons show LVH, not RVH, biases on Chimeric Faces. Therefore, among normal subjects, at most we might raise the possibility that the strength of LVH bias will be slightly en-
Mood state and chimeric faces
473
hanced among those whose mood state is more positive. On the other hand, it is possible that hemispheric arousal style for visuospatial tests is either independent of, or only weakly related to, mood state (within the normal range) or is robust enough to transcend changes in mood state. In the current study, we sought further evidence on this question by comparing the performance of right-handed college students on the Chimeric Faces Test and the Profile of Mood States (POMS) questionnaire (McNair, Lorr and Droppleman, 1971). In previous free-viewing studies with chimeric faces, subjects have been asked only to choose the happier of the two faces in each face pair. On the possibility that laterality effects are related to valence of the target emotion even in a perceptual test, Carlson and Harris (1986) gave a free-viewing chimeric faces test to 80 college students, with each subject given the standard instruction to "choose the happier face" of the pair as well as the instruction to "choose the sadder face". The chimeric faces were constructed from photographs of faces published by Ekman and Friesen (1975) (constructed in the same way and from the same faces as were used in the current study). The results failed to disclose any significant differences between the "choose the happier face" and "choose the sadder face" instructions. Instead, the two instruction conditions yielded equally strong L VH effects, thus suggesting that the two instruction conditions are equivalent. In the current study, we nevertheless decided to seek more information about the effect of instruction on the possibility that any differences related to instruction type might be moderated by subject mood, an uncontrolled variable in previous research. For example, we wondered whether subjects whose mood was either positive or negative might be more sensitive to the instruction condition consistent with the valence of that state and therefore whether they would show a stronger visual hemifield bias in that direction. MATERIALS AND METHOD
Subjects The subjects were 126 college students (95 women, mean age = 18.8 yrs., std. dev. = 1.04; 31 men, mean age = 19.2 yrs., std. dev. =0.87) enrolled in introductory psychology classes at Michigan State University. All subjects participated for course credit. All subjects were righthanded as indexed by self-description and by their answers to a 9-item handedness questionnaire modeled after Annett (1967). All 9 items pertained to hand-use for unimanual tasks and were answered using a 5-point scale from 1 (exclusive left-hand use) to 5 (exclusive right-hand use). The possible range of scores therefore was from 9 (exclusive left-handedness) to 45 (exclusive right-handedness). The mean score for the entire sample was 42.7, std. dev. = 2 .34, range 35 to 45. Of the entire sample, only 12 of the 126 subjects scored below 40, with 58 scoring between 44 and 45. The questionnaire scores thus strongly confirmed the subjects' description of themselves as right-handed. The one possible exceptional case was of a 19-yearold woman who reported that her handedness had been changed, although she gave no details of the circumstances and, in contrast to other converted left-handers (see review by Harris, 1990), reported exclusive or near exclusive right-hand use on all 9 handedness questionnaire items (score = 43 of 45). In addition, familial sinistrality (FS) was determined by questionnaire. Only information about the hand preference of parents and grandparents was solicited because determination of FS on the basis of reports of hand use by siblings, aunts, and uncles risks confounding FS
474
L.l. Harris and P.l. Snyder
Fig. 1 - Sample pair of faces from the Chimeric Faces Test, as adapted from original photographs in Ekman and Friesen (1975). Reproduced with permission of the author.
Mood state and chimeric faces
475
with family size (Bishop, 1983). A subject with either one left-handed biological parent or with two left-handed grandparents was classified as FS + . Of the 95 women, 13 were FS + ; of 31 men, 5 were FS + . On the handedness questionnaire, there was no sign of any difference between the FS + and FS - subjects (for example, 10 of the 18 FS + subjects achieved the maximum score of 45).
Chimeric Faces Test The chimeric faces were constructed from photographs of faces published by Ekman and Friesen (1975). We used the photographs of only those four models (two men, two women) for whom both "happy" and "sad" expressions were modeled. All faces were front-view. Each of the two photographs for each model was divided along the midline axis and then recombined into a composite, or chimeric, face with the happy expression on one side and the sad expression on the other. Each chimera was paired with its mirror image, with the resulting pair of chimeric faces arranged vertically on one page (see Figure 1). To control for the position of the chimeric faces comprising each pair, the positions were counterbalanced so that on half the trials, the face with the target emotion to the viewer's left (thus in the viewer's left visual hemifield) was the top face on half the trials and was the bottom face on the remaining trials. Two series of 8 pairs (original series and replication) were bound into a booklet, which was stapled across the top.
Profile of Mood States Test (POMS) Mood level was assessed with the POMS questionnaire (Profile of Mood States; McNair, Lorr and Droppleman, 1971), a 65-item questionnaire designed to assess mood level in normal persons as well as psychiatric outpatients. The POMS represents the refinement of a total of 100 different adjective scales by means of repeated factor analyses. The test is clustered into six separate factors (Tension-Anxiety, Depression-Dejection, Anger-Hostility, Vigor-Activity, Fatigue-Inertia, and Confusion-Bewilderment). The subjects rate themselves, using a 5-point scale, on adjectives chosen to be components of one or another of these factors, according to whether the item fits them "not at all" (score of 0), "a little" (1), "moderately" (2), "quite a bit" (3), or "extremely" (4). The Tension-Anxiety factor includes such items as tense, shaky, on edge; Depression-Dejection, unhappy, sorry, sad; Anger-Hostility, angry, peeved, grouchy; Vigor-Activity, lively, active, energetic; Fatigue-Inertia, worn out, listless, fatigued; and Confusion-Bewilderment, confused, unable to concentrate, muddled. From the overall score for each factor, a total mood disturbance (TMD) composite score can be derived (weighting the direction of the vigor-activity factor score negatively and all other factors scores positively). Validation studies have revealed high internal consistency within mood dimensions and test-retest reliabilities ranging from r = .65 to r =.74 (Peterson and Headen, 1987). We elected to use the TMD score instead of individual factor scores because the single, global estimate of affective state provided by the TMD score comes closer to capturing the dimension of positive vs. negative mood than do any of the individual factor scores (cf. McNair et aI., 1971, p. 9). Depending on instructions to the subject, the POMS questionnaire can be used either as a measure of present mood state in the sense of mood state that is transient and responsive to environmental change, or as a measure of a more stable, long-term state that may reflect enduring personality traits. (It is noteworthy that the factor structure of responses to this test did not differ significantly between the two instructional conditions in a sample of 235 male college students; McNair et aI., 1971, p. 5). We elected to assess mood state in this latter sense. Following the standard procedure (McNair et aI., 1971), we therefore instructed the subjects as follows: "For each of the feelings listed below, rate yourself to best describe how you have been feeling during the past week, including today". Normative scores for 856 college students given these instructions, as provided in McNair et aI. (1971), served as the basis of comparison with our sample.
476
L.J. Harris and P.J. Snyder
Procedure
The subjects were tested in groups of 20-25. The tests were administered in the following order: handedness questionnaire, POMS questionnaire, Chimeric Faces Test. For the Chimeric Faces Test, half of the subjects, chosen at random, were assigned to the test order Happy-Sad (HS), the remaining subjects to the test order Sad-Happy (SH). In each condition, the subject began by reading the instructions on the test booklet cover page, "For the following 8 pairs of faces, look at each pair and decide which face is happier [sadder]". After completing the first series of 8 trials, subjects were asked to read the instructions for the second series and then to complete the next 8 trials. Testing time for the entire protocol took approximately 45 minutes.
RESULTS
Chimeric Faces
On the Chimeric Faces Test, visual inspection of the data indicated that the distributions of scores were nearly identical for males and females. Except where noted, we therefore elected to combine the data for further analyses. Figure 2 shows how the 126 subjects were distributed according to the number of left visual hemifield (L VH) responses made on the 16 trials of the Chimeric Faces Test (combining the scores for the 8-trial "Which face is happier?" and the 8-trial "Which face is sadder?" series), where a high score (X axis) represents judgments based on the emotional cue positioned in the viewer's LVH, and a low score represents judgments based on the emotional cue positioned in the viewer's RVH. As Figure 2 shows, most of the subjects based the majority of their judgments on the emotional cue positioned in the L VH, with 44 percent Percent of Subjects (N·126)
16~--------------------------------------------'
14~----------------------------------------~~
12~--------------------------------
10 8~--------------------------~~---7----~-----1
2F---~~~~--~-------------------------------1 oL--L~
o
__J-~__~~__~-L__L--L__L--L~~~~~
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Number of LVH Responses Fig. 2 - Distribution of 126 subjects according to the number of L VH responses when: 0= favors L VH on none of the 16 trials across the two 8-trial series (i.e., favors RVH on all 16 trials), and 16 =favors L VH on all 16 trials.
477
Mood state and chimeric faces
Chimeric Faces Test Percent of Subjects 25.-----------------------------------------~
o+----,-----.----,----,,----.----,-----,---~
o
1
2
3
4
5
6
7
Number of LVH Responses per Series -
'Happier'
8
--+- 'Sadder'
Fig. 3 - Distribution of 126 subjects according to the number of L VH responses in each series ("Happier" = "Which face is happier?"; "Sadder" = "Which face is sadder?"), when: 0 =favors L VH on none of the 8 trials across the 8-trial series (i.e., favors RVH on all 8 trials), and 8 = favors L VH on all 8 trials.
of the subjects having a score of 13-16 (L VH choice on at least 13 of the 16 trials). It also is noteworthy that about 10 percent of the subjects were equally consistent in the reverse direction (13-16 RVH choices, or 0-3 LVH choices). Figure 3 shows the distributions of scores for the two instruction conditions separately, where the "Happier" series represents performance on the 8 "Which face is happier" trials, and the "Sadder" series represents performance on the 8 "Which face is sadder" trials. As can be seen, the distributions are virtually identical. A comparison of the two distributions by the Kolmogorov-Smirnov goodness-of-fit test for related samples failed to disclose any difference (Z = .880, p = .421, 2-tailed). This finding is consistent with the positive, significant correlation found when we directly compared individual scores on the two series (Pearson r= .71, p< .001). Making judgments based on choosing the happier face appears to be equivalent to making judgments based on choosing the sadder face. POMS Questionnaire
The POMS questionnaire was scored to yield a global estimate of affective state (the total mood disturbance score, or TMD) for each subject. On this test, the scores can range from 0 to a maximum of 232, where the lower the score, the lower the estimate of negative affective state (McNair et aI., 1971). For the total sample, the mean score was 131 (std. dev. = 31.60; range 100 to 215). When converted to T-scores, the mean TMD score was 45.85 (std. dev. 6.03; range 30.67 to 65.83), which falls easily within the normal range for this age group
478
L.J. Harris and P.J. Snyder
(mean T-score=43.33 for 856 college students; McNair et aI., 1971, p. 20, Table 23) and was to have been expected in this sample of normal college students. Sex differences in both the normative group and in our sample were negligibly small. For men, the mean TMD score was 47.32, std. dev. =7.54; for women, mean TMD score = 45.33, std. dev. = 5.44. Relation between Performance on Chimeric Faces and POMS The relation between performance on Chimeric Faces Test (CFT) and POMS was assessed by correlational analysis. We initially conducted correlational analyses between the two dependent measures separately for each sex. Because the subjects' scores on the "happier" and "sadder" series of the Chimeric Faces Test (CFT) were significantly and positively correlated, for this analysis we therefore calculated each subject's combined score on both series of the CFT stimuli and then correlated the combined score for each subject with the subject's POMS score. The result showed the scores on the two measures to be unrelated for both males (Pearson r = .168) and females (Pearson r= - .154) as well as for the combined sample (Pearson r = - .04). To examine the possible relationship between performance on the CFT and any of the six separate subscales of the POMS test, separate correlational analyses were run for each subscale. All of the Pearson-r correlations, whether for males and females separately or for the sexes combined, were insignificant (range = -0.13 to +0.06). On the possibility that the range of scores for one or both tests was too narrow to permit an adequate correlational analysis, we also performed an "extreme groups" analysis by considering only those subjects whose combined score on the CFT was in the bottom quartile (0-3 LVH choices out of 16 total trials, N = 15, mean = 1.52, std. dev. = 1.14) or the top quartile (13-16 choices out of 16 trials, N = 56, mean = 15.0, std. dev. = 0.94). Then, for each of the resulting groups, the average TMD scores were calculated and compared by analysis of variance. The results failed to disclose any significant differences in TMD scores between the two extreme groups for the combined sample (bottom group, mean TMD T-score=46.3, std. dev.=8.71; top group, mean TMD T-score=45.3, std. dev. =4.72; F< 1.0), for either sex alone, or for any of the six subscales of the POMS separately (all F's < 1.0). CONCLUSIONS
The results for this sample of right-handed college students provide further evidence for the "left visual hemifield effect" on the free-viewing Chimeric Faces Test. Most of the subjects on most of the trials made their choices based on the left-side location of the target emotion (L VH), with 44 percent of the subjects being consistent on at least 13 of the 16 trials, and with 25 percent of the subjects being consistent on at least 15 of the 16 trials. What is also noteworthy is that a small minority of subjects made as many RVH choices as the majority made LVH choices. The results also corroborate the earlier finding
Mood state and chimeric faces
479
(Carlson and Harris, 1986) that the direction and strength of the hemifield bias in the free-viewing Chimeric Faces Test are unrelated to the valence of the emotion (positive or negative) that the subject is asked to judge in the target face. Given previous evidence (e.g., Levy et aI., 1983) attesting to the validity and reliability of the Chimeric Faces Test as a measure of characteristic hemispheric arousal style, and accepting the POMS questionnaire as a valid index of longterm, or enduring, mood state, the results, however, fail to disclose any relation between hemispheric arousal style and mood state, even when valence of the target emotion is taken into account. After collecting our data, we learned of the results of an unpublished pilot study by Banich (personal communication, 1990) with which our own findings appear to be consistent. Banich gave the Chimeric Faces Test to 12 college students whose mood state had been pharmacologically altered, during different testing sessions, by administration of amphetamines, alcohol, and the psychotropic tranquilizer haloperidol. All subjects received all conditions, with mood state measured by self-ratings on the POMS questionnaire. Instead of using the TMD score or any individual factor score, Banich used a subset of adjectives specifically chosen as likely indices of arousal and positive mood. Although the scores on this subset of POMS adjective indicated that the mood manipulation had been successful, mood state was not significantly related to performance on the Chimeric Faces Test. This finding leads us to believe that our own results would not have been different had we used individual factor scores instead of the TMD score. There may be several reasons for our failure to find any relationship between mood state, as measured by the POMS questionnaire, and hemispheric arousal style, as measured by the Chimeric Faces Test. First, there may indeed be no relationship between these two psychological dimensions. Second, if there is a relationship, it may have been masked by the relatively truncated range of the total mood disturbance (TMD) scores for our subjects, all of whom were healthy normal college students with scores closer to the low end of the TMD scale. This suggests that, for statistical reasons, the relationship will appear only in comparisons between normal and clinical populations. For example, it might appear between normal persons and patients with chronic major depressive disorders, or as a within-subject effect in patients who are alternating between the manic and depressive poles in bi-polar disorders. Banich's results, however, would appear to argue against this explanation inasmuch as she failed to find any relationship between mood state and performance on the Chimeric Faces Test even though her normal subjects presumably were exhibiting a range of mood scores more nearly equivalent to that shown by a clinical population. On the other hand, even if normal persons can be made temporarily 'manic' or 'depressed' at levels approximating those of some bipolar patients in manic or depressed phases, or of patients with unipolar depression, neuropharmacolog~ ically and neuropsychologically we may be comparing apples and oranges (cf. Jaeger, Borod and Peselow, 1987). In summary, the results provide further evidence that, for normal righthanded adults, the L VH bias on the Chimeric Faces Test is the modal bias, but they fail to support the hypothesis that hemispheric arousal style as measured by this test is related to long-term mood state as measured by the POMS ques-
480
L.J. Harris and P.J. Snyder
tionnaire. That performance on the Chimeric Faces Test also was unrelated to instruction condition, even taking individual differences in subject mood state into account, further supports the position that the right hemisphere has primacy in the recognition of emotion, irrespective of the valence of the emotion in the target stimulus. ABSTRACT
Levy, Heller, Banich, and Burton (1983) have shown that hemispheric activational, or arousal, style, as measured by direction and consistency of choice of the "happer face" on a free-viewing Chimeric Faces Test, is highly reliable and varies across individuals who otherwise presumably have similar cortical organization (e.g., right-handed college students). The current experiment asks whether such individual differences in hemispheric arousal style are moderated by long-term, in the sense of enduring, individual differences in mood, as measured by the Profile of Mood States (POMS) questionnaire. Like Levy et al. 's subjects, most of our subjects (126 right-handed college students) made most of their choices on the Chimeric Faces Test based on the emotional cue (the smile) positioned in the half of the face to the viewer's left, or left visual hemifield (L VH), whereas a small minority were equally consistent in making their choice based on the emotional cue positioned in the half of the face to the viewer's right, or right visual hemifield (RVH). Performance on the Chimeric Faces Test, however, proved to be unrelated to scores on the POMS questionnaire. This suggests that hemispheric arousal style, as indexed by the Chimeric Faces Test, is robust enough, for a population of normal right-handers, to transcend any differences in mood as measured by the POMS questionnaire.
Acknowledgement. We are indebted to Kimberly Burnett for her conscientious help in the collection and reduction of the data for this study. We also are grateful to Marie T. Banich, Wendy Heller, and three anonymous referees for their helpful comments on an earlier version of this paper. REFERENCES
ANNETT, M. The binomial distribution of right, mixed, and left handedness. Quarterly Journal oj Experimental Psychology, 19: 327-333, 1967. BANICH, M. Personal communication, 15 August, 1990. BEAR, D.M., and FEOlO, P. Quantitative analysis of interictal behavior in temporal lobe epilepsy. Archives oj Neurology, 34: 454-467, 1977. BENSON, F., and ZAIDEL, E. (Eds.), The Dual Brain: Hemispheric Specialization in Humans. New York: Guilford, 1985. BISHOP, D.V.M. How sinister is sinistrality? Journal oj the Royal College oj Physicians oj London, 17: 161-172, 1983. BOROD, J.C., CARPER, M., NAESER, M., and GOODGLASS, H. Left-handed and right-handed aphasics with left hemisphere lesions compared on nonverbal measures. Cortex, 21: 81-90, 1985. CARLSON, D.F., and HARRIS, L.J. Perception of happiness and sadness in free viewing of chimeric faces. Paper presented at the 14th Annual Meeting of the International Neuropsychological Society, 6-9 Feb., 1986, Denver, Colorado (Abstract in Journal oj Clinical and Experimental Neuropsychology, 1985, 7, 636). DAVIDSON, R.J. Affect, cognition, and hemispheric specialization. In C.E. Izard, J. Kagan and R. Zajonc (Eds.), Emotion, Cognition, and Behavior. New York: Cambridge University Press, 1984. DRONKERS, N.F., and KNIGHT, R.T. Right-sided neglect in a left-hander: Evidence for reversed hemispheric specialization of attention capacity. Neuropsychologia, 27: 729-735, 1988. D'ELIA, G., and PERRIS, C. Cerebral functional dominance and depression. Acta Psychiatrica Scandinavica, 49: 191-197, 1973. EKMAN, P., and FRIESEN, W. V. Unmasking the Face: A Guide to Recognizing EmotionsJrom Facial Cues. Englewood Cliffs, New Jersey: Prentice-Hall, 1975. FLOR-HENRY, P. On certain aspects of the localization of the cerebral systems regulating and determining emotion. Biological Psychiatry, 14: 677-698, 1979. GAINOTTI, G. Emotional behavior and hemispheric side of lesion. Cortex, 8: 41-55, 1972.
Mood state and chimeric faces
481
GRUZELlER, J ., and VENABLES, P. Bimodality and lateral asymmetry of skin conductance orienting activity in schizophrenics: Replication and evidence of lateral asymmetry in patients with depression and disorders of personality. Biological Psychiatry, 8: 55-73, 1974. HARRIS, L.l. Cultural influences on handedness: Historical and contemporary theory and evidence. In S. Coren (Ed.), Left-handedness: Behavioral Implications and Anomalies. Advances in Psychology Series, Vol. 67. Amsterdam: Elsevier Science Publishers B.V.lNorth Holland Book Series, 1990. HEILMAN, K.M., and SATZ, P . Neuropsychology of Human Emotion . New York: Guilford Press, 1983. HELLER, W. The neuropsychology of emotion: Developmental patterns and implications for psychopathology. In N.L. Stein, B.L. Leventhal and T. Trabasso (Eds.), Psychological and Biological Approaches to Emotion. Hillsdale, N.l .: Lawrence Erlbaum Associates, 1990. l AEGER, 1., BOROD, 1.C., and PESELOW, E . Depressed patients have atypical hemispace biases in the perception of emotional chimeric faces. Journal of Abnormal Psychology, 96: 321-324,1987. LEVY, 1., HELLER , W .,BANICH, M .T ., and BURTON, L.A . Are variations among right-handed individuals in perceptual asymmetries caused by characteristic arousal differences between hemispheres? Journal of Experimental Psychology: Human Perception and Performance, 9: 329-359, 1983. LEY, R.G ., and BRYDEN, M.P . Hemispheric differences in processing emotions and faces. Brain and Language, 7: 127-138, 1979. Mc NAIR, D.M., LORR, M., and DROPPLEMAN, L.F. EITS Manual for the Profile of Mood States. San Diego, Calif.: Educational and Industrial Testing Service, 1971. PETERSON, R.A., and HEADEN, S.W. Profile of mood states. in D .l . Keyser and R.C. Sweetland (Eds .), Test Critiques Compendium. Kansas City, Mo.: Test Corporation of America, 1987. PRITCHARD, P.B. Personality and emotional complications of epilepsy. In K.M. Heilman and P. Satz (Eds.), Neuropsychology of Human Emotion. New York: Guilford Press, 1983. SACKEIM, H.A., GREENBERG, M .S., WEIMAN, A.L., GUR, R.C., HUNGERBUHLER, J.P., and GESCHWIND, N. Hemispheric asymmetry in the emotion of positive and negative emotions: Neurological evidence. Archives of Neurology, 39: 210-218, 1982. Lauren Julius Harris, Department of Psychology. Michigan Stale Uni versity. East L ansing. Michigan 48824, U .S.A.
Peter J . S nyder, De partment of Psychiatry, Hillside Hospital-Resea rch, Long Island Jewish Medical Center, P.O. Box 38,Glen Oaks , N. Y. 11004, U.S.A.
INTRODUCTION
Clinical and experimental research has established that the right cerebral hemisphere plays a leading role in the processing and mediation of visuospatial and other non-verbal functions. The research indicates that this lateralization pattern applies strongly in the case of right-handers. In left-handers, the lateralization pattern, although in the same direction, is typically weaker and more variable, possibly reflecting a relatively greater measure of bilaterality of functional organization in the form of secondary left-hemisphere commitment to visuospatial processing (e.g., Borod, Carper, Naeser et al., 1985). The clinical research also suggests that only in rare instances do left-handers show what amounts to true reversed dominance for visuospatial functions (e.g., Dronkers and Knight, 1988). Despite the evidence suggesting relative uniformity of lateralization for normal right-handers, there also is much evidence of individual differences in this same group on tests of lateralized functioning. For example, among righthanded college students, Levy, Heller, Banich and Burton (1983) found striking differences in both the strength and direction of the visual hemifield advantage for tachistoscopically-presented non-verbal stimuli as well as for stimuli presented for free-viewing. The latter were mirror-image pairs of "chimeric faces" , each with a smile on one side, a frown on the other. Asked to choose the "happier" chimeric face of each pair, most subjects consistently made their choice based on the emotional cue (the smile) located to their left (in their left visual hemifield, or L VH), but a significant minority just as consistently made their choice based on the emotional cue located to their right (right visual hemifield, or RVH). Believing it unlikely that such individual differences in a presumably modal lateralization group (right-handed adults) reflected actual differences in the direction of lateral specialization for visuospatial functions, Levy et al. proposed that they instead were reflections of differences in characteristic patterns of hemispheric arousal, or activation, regardless of the task. More specifically, they proposed that people vary in their characteristic patterns of arousal, with some showing more left-hemisphere arousal and others showing more right-hemisphere arousal, so that when that arousal tendency is commensurate with the hemispheric processing elicited by a particular task, it serves to enhance the overall asymmetry in the direction of the specialized hemisphere. Cortex, (1992) 28, 471-481
472
L.J. Harris and P.J. Snyder
When the arousal tendency opposes the hemisphere processing elicited by the task, it attenuates the overall asymmetry in the direction of the specialized hemisphere. Inasmuch as most people show a L VH bias on the Chimeric Faces Test, we therefore can conclude that the test elicits right-hemisphere specialized processing and that, superimposed on that bias, is a characteristic arousal bias that acts either to strengthen or to attenuate the bias (for further discussion, see Heller, 1990). If there are such differences in patterns of characteristic hemispheric arousal, what is their origin if not in actual differences in lateral cortical specialization? One possible basis may lie in individual differences in characteristic emotional, or mood, state insofar as such differences might reflect differences in direction of hemispheric arousal. One immediate complication is that the neuropsychological evidence is inconsistent on the question of hemispheric specialization for emotion. Some evidence, with both clinical and normal populations, points to the primacy of the right hemisphere for all emotional functions (e.g., Flor-Henry, 1979; Levy et aI., 1983; Ley and Bryden, 1979), whereas other evidence suggests that each hemisphere is specialized to process different types of emotion depending on valence, the left hemisphere for positive emotions, the right hemisphere for negative emotions (e.g., Davidson, 1984; Sackeim, Greenberg, Weiman et aI., 1982). Much of the evidence supporting a valence effect, however, comes from studies of the subject's own emotional expression or feeling (e.g., studies of emotional reactions in patients with either left- or righttemporal lobe epileptogenic foci, cf. Bear and Fedio, 1977; review by Pritchard, 1983), whereas the evidence supporting the position of right-hemisphere primacy for all emotions more often comes from studies of the perception of emotion (e.g., recognition of emotion in tachistoscopically-projected faces; for reviews, see Benson and Zaidel, 1986; Heilman and Satz, 1983; Heller, 1990). One way to reconcile the two bodies of evidence is to suppose that the right hemisphere mediates the perception, or recognition, of emotion, whereas each hemisphere takes somewhat different responsibilities for the expression of emotion (cf. Heller, 1990). If, then, we credit the hypothesis linking valence of subject mood with a particular cerebral hemisphere, it raises the question whether subject mood state will affect performance on a perceptual task such as the Chimeric Faces Test. Some support for the general hypothesis linking performance with mood has been reported in a clinical population of depressed patients (Jaeger, Borod and Peselow, 1987). Like normal controls, the patients showed a left hemifield bias but to a weaker extent, consistent with independent evidence from studies of unipolar depressive patients that show diminished right (relative to left) electroencephalogram amplitude (D'Elia and Perris, 1973) and amplitude of skin conductance for an orienting response (Gruzelier and Venables, 1974). Accordingly, we would not go so far as to suppose that normal persons whose mood states are negative would show a complete reversal of the usual L VH bias on Chimeric Faces. The reason is that this supposition implies that in most normal subjects, the typical mood state is negative. We already know, however, that most normal persons show LVH, not RVH, biases on Chimeric Faces. Therefore, among normal subjects, at most we might raise the possibility that the strength of LVH bias will be slightly en-
Mood state and chimeric faces
473
hanced among those whose mood state is more positive. On the other hand, it is possible that hemispheric arousal style for visuospatial tests is either independent of, or only weakly related to, mood state (within the normal range) or is robust enough to transcend changes in mood state. In the current study, we sought further evidence on this question by comparing the performance of right-handed college students on the Chimeric Faces Test and the Profile of Mood States (POMS) questionnaire (McNair, Lorr and Droppleman, 1971). In previous free-viewing studies with chimeric faces, subjects have been asked only to choose the happier of the two faces in each face pair. On the possibility that laterality effects are related to valence of the target emotion even in a perceptual test, Carlson and Harris (1986) gave a free-viewing chimeric faces test to 80 college students, with each subject given the standard instruction to "choose the happier face" of the pair as well as the instruction to "choose the sadder face". The chimeric faces were constructed from photographs of faces published by Ekman and Friesen (1975) (constructed in the same way and from the same faces as were used in the current study). The results failed to disclose any significant differences between the "choose the happier face" and "choose the sadder face" instructions. Instead, the two instruction conditions yielded equally strong L VH effects, thus suggesting that the two instruction conditions are equivalent. In the current study, we nevertheless decided to seek more information about the effect of instruction on the possibility that any differences related to instruction type might be moderated by subject mood, an uncontrolled variable in previous research. For example, we wondered whether subjects whose mood was either positive or negative might be more sensitive to the instruction condition consistent with the valence of that state and therefore whether they would show a stronger visual hemifield bias in that direction. MATERIALS AND METHOD
Subjects The subjects were 126 college students (95 women, mean age = 18.8 yrs., std. dev. = 1.04; 31 men, mean age = 19.2 yrs., std. dev. =0.87) enrolled in introductory psychology classes at Michigan State University. All subjects participated for course credit. All subjects were righthanded as indexed by self-description and by their answers to a 9-item handedness questionnaire modeled after Annett (1967). All 9 items pertained to hand-use for unimanual tasks and were answered using a 5-point scale from 1 (exclusive left-hand use) to 5 (exclusive right-hand use). The possible range of scores therefore was from 9 (exclusive left-handedness) to 45 (exclusive right-handedness). The mean score for the entire sample was 42.7, std. dev. = 2 .34, range 35 to 45. Of the entire sample, only 12 of the 126 subjects scored below 40, with 58 scoring between 44 and 45. The questionnaire scores thus strongly confirmed the subjects' description of themselves as right-handed. The one possible exceptional case was of a 19-yearold woman who reported that her handedness had been changed, although she gave no details of the circumstances and, in contrast to other converted left-handers (see review by Harris, 1990), reported exclusive or near exclusive right-hand use on all 9 handedness questionnaire items (score = 43 of 45). In addition, familial sinistrality (FS) was determined by questionnaire. Only information about the hand preference of parents and grandparents was solicited because determination of FS on the basis of reports of hand use by siblings, aunts, and uncles risks confounding FS
474
L.l. Harris and P.l. Snyder
Fig. 1 - Sample pair of faces from the Chimeric Faces Test, as adapted from original photographs in Ekman and Friesen (1975). Reproduced with permission of the author.
Mood state and chimeric faces
475
with family size (Bishop, 1983). A subject with either one left-handed biological parent or with two left-handed grandparents was classified as FS + . Of the 95 women, 13 were FS + ; of 31 men, 5 were FS + . On the handedness questionnaire, there was no sign of any difference between the FS + and FS - subjects (for example, 10 of the 18 FS + subjects achieved the maximum score of 45).
Chimeric Faces Test The chimeric faces were constructed from photographs of faces published by Ekman and Friesen (1975). We used the photographs of only those four models (two men, two women) for whom both "happy" and "sad" expressions were modeled. All faces were front-view. Each of the two photographs for each model was divided along the midline axis and then recombined into a composite, or chimeric, face with the happy expression on one side and the sad expression on the other. Each chimera was paired with its mirror image, with the resulting pair of chimeric faces arranged vertically on one page (see Figure 1). To control for the position of the chimeric faces comprising each pair, the positions were counterbalanced so that on half the trials, the face with the target emotion to the viewer's left (thus in the viewer's left visual hemifield) was the top face on half the trials and was the bottom face on the remaining trials. Two series of 8 pairs (original series and replication) were bound into a booklet, which was stapled across the top.
Profile of Mood States Test (POMS) Mood level was assessed with the POMS questionnaire (Profile of Mood States; McNair, Lorr and Droppleman, 1971), a 65-item questionnaire designed to assess mood level in normal persons as well as psychiatric outpatients. The POMS represents the refinement of a total of 100 different adjective scales by means of repeated factor analyses. The test is clustered into six separate factors (Tension-Anxiety, Depression-Dejection, Anger-Hostility, Vigor-Activity, Fatigue-Inertia, and Confusion-Bewilderment). The subjects rate themselves, using a 5-point scale, on adjectives chosen to be components of one or another of these factors, according to whether the item fits them "not at all" (score of 0), "a little" (1), "moderately" (2), "quite a bit" (3), or "extremely" (4). The Tension-Anxiety factor includes such items as tense, shaky, on edge; Depression-Dejection, unhappy, sorry, sad; Anger-Hostility, angry, peeved, grouchy; Vigor-Activity, lively, active, energetic; Fatigue-Inertia, worn out, listless, fatigued; and Confusion-Bewilderment, confused, unable to concentrate, muddled. From the overall score for each factor, a total mood disturbance (TMD) composite score can be derived (weighting the direction of the vigor-activity factor score negatively and all other factors scores positively). Validation studies have revealed high internal consistency within mood dimensions and test-retest reliabilities ranging from r = .65 to r =.74 (Peterson and Headen, 1987). We elected to use the TMD score instead of individual factor scores because the single, global estimate of affective state provided by the TMD score comes closer to capturing the dimension of positive vs. negative mood than do any of the individual factor scores (cf. McNair et aI., 1971, p. 9). Depending on instructions to the subject, the POMS questionnaire can be used either as a measure of present mood state in the sense of mood state that is transient and responsive to environmental change, or as a measure of a more stable, long-term state that may reflect enduring personality traits. (It is noteworthy that the factor structure of responses to this test did not differ significantly between the two instructional conditions in a sample of 235 male college students; McNair et aI., 1971, p. 5). We elected to assess mood state in this latter sense. Following the standard procedure (McNair et aI., 1971), we therefore instructed the subjects as follows: "For each of the feelings listed below, rate yourself to best describe how you have been feeling during the past week, including today". Normative scores for 856 college students given these instructions, as provided in McNair et aI. (1971), served as the basis of comparison with our sample.
476
L.J. Harris and P.J. Snyder
Procedure
The subjects were tested in groups of 20-25. The tests were administered in the following order: handedness questionnaire, POMS questionnaire, Chimeric Faces Test. For the Chimeric Faces Test, half of the subjects, chosen at random, were assigned to the test order Happy-Sad (HS), the remaining subjects to the test order Sad-Happy (SH). In each condition, the subject began by reading the instructions on the test booklet cover page, "For the following 8 pairs of faces, look at each pair and decide which face is happier [sadder]". After completing the first series of 8 trials, subjects were asked to read the instructions for the second series and then to complete the next 8 trials. Testing time for the entire protocol took approximately 45 minutes.
RESULTS
Chimeric Faces
On the Chimeric Faces Test, visual inspection of the data indicated that the distributions of scores were nearly identical for males and females. Except where noted, we therefore elected to combine the data for further analyses. Figure 2 shows how the 126 subjects were distributed according to the number of left visual hemifield (L VH) responses made on the 16 trials of the Chimeric Faces Test (combining the scores for the 8-trial "Which face is happier?" and the 8-trial "Which face is sadder?" series), where a high score (X axis) represents judgments based on the emotional cue positioned in the viewer's LVH, and a low score represents judgments based on the emotional cue positioned in the viewer's RVH. As Figure 2 shows, most of the subjects based the majority of their judgments on the emotional cue positioned in the L VH, with 44 percent Percent of Subjects (N·126)
16~--------------------------------------------'
14~----------------------------------------~~
12~--------------------------------
10 8~--------------------------~~---7----~-----1
2F---~~~~--~-------------------------------1 oL--L~
o
__J-~__~~__~-L__L--L__L--L~~~~~
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Number of LVH Responses Fig. 2 - Distribution of 126 subjects according to the number of L VH responses when: 0= favors L VH on none of the 16 trials across the two 8-trial series (i.e., favors RVH on all 16 trials), and 16 =favors L VH on all 16 trials.
477
Mood state and chimeric faces
Chimeric Faces Test Percent of Subjects 25.-----------------------------------------~
o+----,-----.----,----,,----.----,-----,---~
o
1
2
3
4
5
6
7
Number of LVH Responses per Series -
'Happier'
8
--+- 'Sadder'
Fig. 3 - Distribution of 126 subjects according to the number of L VH responses in each series ("Happier" = "Which face is happier?"; "Sadder" = "Which face is sadder?"), when: 0 =favors L VH on none of the 8 trials across the 8-trial series (i.e., favors RVH on all 8 trials), and 8 = favors L VH on all 8 trials.
of the subjects having a score of 13-16 (L VH choice on at least 13 of the 16 trials). It also is noteworthy that about 10 percent of the subjects were equally consistent in the reverse direction (13-16 RVH choices, or 0-3 LVH choices). Figure 3 shows the distributions of scores for the two instruction conditions separately, where the "Happier" series represents performance on the 8 "Which face is happier" trials, and the "Sadder" series represents performance on the 8 "Which face is sadder" trials. As can be seen, the distributions are virtually identical. A comparison of the two distributions by the Kolmogorov-Smirnov goodness-of-fit test for related samples failed to disclose any difference (Z = .880, p = .421, 2-tailed). This finding is consistent with the positive, significant correlation found when we directly compared individual scores on the two series (Pearson r= .71, p< .001). Making judgments based on choosing the happier face appears to be equivalent to making judgments based on choosing the sadder face. POMS Questionnaire
The POMS questionnaire was scored to yield a global estimate of affective state (the total mood disturbance score, or TMD) for each subject. On this test, the scores can range from 0 to a maximum of 232, where the lower the score, the lower the estimate of negative affective state (McNair et aI., 1971). For the total sample, the mean score was 131 (std. dev. = 31.60; range 100 to 215). When converted to T-scores, the mean TMD score was 45.85 (std. dev. 6.03; range 30.67 to 65.83), which falls easily within the normal range for this age group
478
L.J. Harris and P.J. Snyder
(mean T-score=43.33 for 856 college students; McNair et aI., 1971, p. 20, Table 23) and was to have been expected in this sample of normal college students. Sex differences in both the normative group and in our sample were negligibly small. For men, the mean TMD score was 47.32, std. dev. =7.54; for women, mean TMD score = 45.33, std. dev. = 5.44. Relation between Performance on Chimeric Faces and POMS The relation between performance on Chimeric Faces Test (CFT) and POMS was assessed by correlational analysis. We initially conducted correlational analyses between the two dependent measures separately for each sex. Because the subjects' scores on the "happier" and "sadder" series of the Chimeric Faces Test (CFT) were significantly and positively correlated, for this analysis we therefore calculated each subject's combined score on both series of the CFT stimuli and then correlated the combined score for each subject with the subject's POMS score. The result showed the scores on the two measures to be unrelated for both males (Pearson r = .168) and females (Pearson r= - .154) as well as for the combined sample (Pearson r = - .04). To examine the possible relationship between performance on the CFT and any of the six separate subscales of the POMS test, separate correlational analyses were run for each subscale. All of the Pearson-r correlations, whether for males and females separately or for the sexes combined, were insignificant (range = -0.13 to +0.06). On the possibility that the range of scores for one or both tests was too narrow to permit an adequate correlational analysis, we also performed an "extreme groups" analysis by considering only those subjects whose combined score on the CFT was in the bottom quartile (0-3 LVH choices out of 16 total trials, N = 15, mean = 1.52, std. dev. = 1.14) or the top quartile (13-16 choices out of 16 trials, N = 56, mean = 15.0, std. dev. = 0.94). Then, for each of the resulting groups, the average TMD scores were calculated and compared by analysis of variance. The results failed to disclose any significant differences in TMD scores between the two extreme groups for the combined sample (bottom group, mean TMD T-score=46.3, std. dev.=8.71; top group, mean TMD T-score=45.3, std. dev. =4.72; F< 1.0), for either sex alone, or for any of the six subscales of the POMS separately (all F's < 1.0). CONCLUSIONS
The results for this sample of right-handed college students provide further evidence for the "left visual hemifield effect" on the free-viewing Chimeric Faces Test. Most of the subjects on most of the trials made their choices based on the left-side location of the target emotion (L VH), with 44 percent of the subjects being consistent on at least 13 of the 16 trials, and with 25 percent of the subjects being consistent on at least 15 of the 16 trials. What is also noteworthy is that a small minority of subjects made as many RVH choices as the majority made LVH choices. The results also corroborate the earlier finding
Mood state and chimeric faces
479
(Carlson and Harris, 1986) that the direction and strength of the hemifield bias in the free-viewing Chimeric Faces Test are unrelated to the valence of the emotion (positive or negative) that the subject is asked to judge in the target face. Given previous evidence (e.g., Levy et aI., 1983) attesting to the validity and reliability of the Chimeric Faces Test as a measure of characteristic hemispheric arousal style, and accepting the POMS questionnaire as a valid index of longterm, or enduring, mood state, the results, however, fail to disclose any relation between hemispheric arousal style and mood state, even when valence of the target emotion is taken into account. After collecting our data, we learned of the results of an unpublished pilot study by Banich (personal communication, 1990) with which our own findings appear to be consistent. Banich gave the Chimeric Faces Test to 12 college students whose mood state had been pharmacologically altered, during different testing sessions, by administration of amphetamines, alcohol, and the psychotropic tranquilizer haloperidol. All subjects received all conditions, with mood state measured by self-ratings on the POMS questionnaire. Instead of using the TMD score or any individual factor score, Banich used a subset of adjectives specifically chosen as likely indices of arousal and positive mood. Although the scores on this subset of POMS adjective indicated that the mood manipulation had been successful, mood state was not significantly related to performance on the Chimeric Faces Test. This finding leads us to believe that our own results would not have been different had we used individual factor scores instead of the TMD score. There may be several reasons for our failure to find any relationship between mood state, as measured by the POMS questionnaire, and hemispheric arousal style, as measured by the Chimeric Faces Test. First, there may indeed be no relationship between these two psychological dimensions. Second, if there is a relationship, it may have been masked by the relatively truncated range of the total mood disturbance (TMD) scores for our subjects, all of whom were healthy normal college students with scores closer to the low end of the TMD scale. This suggests that, for statistical reasons, the relationship will appear only in comparisons between normal and clinical populations. For example, it might appear between normal persons and patients with chronic major depressive disorders, or as a within-subject effect in patients who are alternating between the manic and depressive poles in bi-polar disorders. Banich's results, however, would appear to argue against this explanation inasmuch as she failed to find any relationship between mood state and performance on the Chimeric Faces Test even though her normal subjects presumably were exhibiting a range of mood scores more nearly equivalent to that shown by a clinical population. On the other hand, even if normal persons can be made temporarily 'manic' or 'depressed' at levels approximating those of some bipolar patients in manic or depressed phases, or of patients with unipolar depression, neuropharmacolog~ ically and neuropsychologically we may be comparing apples and oranges (cf. Jaeger, Borod and Peselow, 1987). In summary, the results provide further evidence that, for normal righthanded adults, the L VH bias on the Chimeric Faces Test is the modal bias, but they fail to support the hypothesis that hemispheric arousal style as measured by this test is related to long-term mood state as measured by the POMS ques-
480
L.J. Harris and P.J. Snyder
tionnaire. That performance on the Chimeric Faces Test also was unrelated to instruction condition, even taking individual differences in subject mood state into account, further supports the position that the right hemisphere has primacy in the recognition of emotion, irrespective of the valence of the emotion in the target stimulus. ABSTRACT
Levy, Heller, Banich, and Burton (1983) have shown that hemispheric activational, or arousal, style, as measured by direction and consistency of choice of the "happer face" on a free-viewing Chimeric Faces Test, is highly reliable and varies across individuals who otherwise presumably have similar cortical organization (e.g., right-handed college students). The current experiment asks whether such individual differences in hemispheric arousal style are moderated by long-term, in the sense of enduring, individual differences in mood, as measured by the Profile of Mood States (POMS) questionnaire. Like Levy et al. 's subjects, most of our subjects (126 right-handed college students) made most of their choices on the Chimeric Faces Test based on the emotional cue (the smile) positioned in the half of the face to the viewer's left, or left visual hemifield (L VH), whereas a small minority were equally consistent in making their choice based on the emotional cue positioned in the half of the face to the viewer's right, or right visual hemifield (RVH). Performance on the Chimeric Faces Test, however, proved to be unrelated to scores on the POMS questionnaire. This suggests that hemispheric arousal style, as indexed by the Chimeric Faces Test, is robust enough, for a population of normal right-handers, to transcend any differences in mood as measured by the POMS questionnaire.
Acknowledgement. We are indebted to Kimberly Burnett for her conscientious help in the collection and reduction of the data for this study. We also are grateful to Marie T. Banich, Wendy Heller, and three anonymous referees for their helpful comments on an earlier version of this paper. REFERENCES
ANNETT, M. The binomial distribution of right, mixed, and left handedness. Quarterly Journal oj Experimental Psychology, 19: 327-333, 1967. BANICH, M. Personal communication, 15 August, 1990. BEAR, D.M., and FEOlO, P. Quantitative analysis of interictal behavior in temporal lobe epilepsy. Archives oj Neurology, 34: 454-467, 1977. BENSON, F., and ZAIDEL, E. (Eds.), The Dual Brain: Hemispheric Specialization in Humans. New York: Guilford, 1985. BISHOP, D.V.M. How sinister is sinistrality? Journal oj the Royal College oj Physicians oj London, 17: 161-172, 1983. BOROD, J.C., CARPER, M., NAESER, M., and GOODGLASS, H. Left-handed and right-handed aphasics with left hemisphere lesions compared on nonverbal measures. Cortex, 21: 81-90, 1985. CARLSON, D.F., and HARRIS, L.J. Perception of happiness and sadness in free viewing of chimeric faces. Paper presented at the 14th Annual Meeting of the International Neuropsychological Society, 6-9 Feb., 1986, Denver, Colorado (Abstract in Journal oj Clinical and Experimental Neuropsychology, 1985, 7, 636). DAVIDSON, R.J. Affect, cognition, and hemispheric specialization. In C.E. Izard, J. Kagan and R. Zajonc (Eds.), Emotion, Cognition, and Behavior. New York: Cambridge University Press, 1984. DRONKERS, N.F., and KNIGHT, R.T. Right-sided neglect in a left-hander: Evidence for reversed hemispheric specialization of attention capacity. Neuropsychologia, 27: 729-735, 1988. D'ELIA, G., and PERRIS, C. Cerebral functional dominance and depression. Acta Psychiatrica Scandinavica, 49: 191-197, 1973. EKMAN, P., and FRIESEN, W. V. Unmasking the Face: A Guide to Recognizing EmotionsJrom Facial Cues. Englewood Cliffs, New Jersey: Prentice-Hall, 1975. FLOR-HENRY, P. On certain aspects of the localization of the cerebral systems regulating and determining emotion. Biological Psychiatry, 14: 677-698, 1979. GAINOTTI, G. Emotional behavior and hemispheric side of lesion. Cortex, 8: 41-55, 1972.
Mood state and chimeric faces
481
GRUZELlER, J ., and VENABLES, P. Bimodality and lateral asymmetry of skin conductance orienting activity in schizophrenics: Replication and evidence of lateral asymmetry in patients with depression and disorders of personality. Biological Psychiatry, 8: 55-73, 1974. HARRIS, L.l. Cultural influences on handedness: Historical and contemporary theory and evidence. In S. Coren (Ed.), Left-handedness: Behavioral Implications and Anomalies. Advances in Psychology Series, Vol. 67. Amsterdam: Elsevier Science Publishers B.V.lNorth Holland Book Series, 1990. HEILMAN, K.M., and SATZ, P . Neuropsychology of Human Emotion . New York: Guilford Press, 1983. HELLER, W. The neuropsychology of emotion: Developmental patterns and implications for psychopathology. In N.L. Stein, B.L. Leventhal and T. Trabasso (Eds.), Psychological and Biological Approaches to Emotion. Hillsdale, N.l .: Lawrence Erlbaum Associates, 1990. l AEGER, 1., BOROD, 1.C., and PESELOW, E . Depressed patients have atypical hemispace biases in the perception of emotional chimeric faces. Journal of Abnormal Psychology, 96: 321-324,1987. LEVY, 1., HELLER , W .,BANICH, M .T ., and BURTON, L.A . Are variations among right-handed individuals in perceptual asymmetries caused by characteristic arousal differences between hemispheres? Journal of Experimental Psychology: Human Perception and Performance, 9: 329-359, 1983. LEY, R.G ., and BRYDEN, M.P . Hemispheric differences in processing emotions and faces. Brain and Language, 7: 127-138, 1979. Mc NAIR, D.M., LORR, M., and DROPPLEMAN, L.F. EITS Manual for the Profile of Mood States. San Diego, Calif.: Educational and Industrial Testing Service, 1971. PETERSON, R.A., and HEADEN, S.W. Profile of mood states. in D .l . Keyser and R.C. Sweetland (Eds .), Test Critiques Compendium. Kansas City, Mo.: Test Corporation of America, 1987. PRITCHARD, P.B. Personality and emotional complications of epilepsy. In K.M. Heilman and P. Satz (Eds.), Neuropsychology of Human Emotion. New York: Guilford Press, 1983. SACKEIM, H.A., GREENBERG, M .S., WEIMAN, A.L., GUR, R.C., HUNGERBUHLER, J.P., and GESCHWIND, N. Hemispheric asymmetry in the emotion of positive and negative emotions: Neurological evidence. Archives of Neurology, 39: 210-218, 1982. Lauren Julius Harris, Department of Psychology. Michigan Stale Uni versity. East L ansing. Michigan 48824, U .S.A.
Peter J . S nyder, De partment of Psychiatry, Hillside Hospital-Resea rch, Long Island Jewish Medical Center, P.O. Box 38,Glen Oaks , N. Y. 11004, U.S.A.
