Functional MR imaging of regional brain responses to pleasant and unpleasant odors.

BACKGROUND AND PURPOSE: Odors can elicit a range of behaviors and emotions. Our purpose was to identify regional activation of the human cerebral cortex in response to pleasant (positive hedonic value) and unpleasant (negative hedonic value) odors. METHODS: Thirteen neurologically normal adults underwent functional MR imaging of frontal and anterior temporal brain regions with a gradient-echo echo-planar technique. Eleven candidate regions of interest (ROIs) were identified on the first half of the data set based on t-map comparisons of signal intensities during administration of clementine (pleasant odor), isovaleric acid (unpleasant odor), and clear air (control odor). These ROIs were applied to the second half of the data set, and the number of voxels activated with the odorants was compared with the number of voxels activated during clear air trials, using independent t-tests. RESULTS: Clementine activated five cortical areas: Brodmann's area (BA) 8, BA 32 (lateralized to left), BA 46/9, BA 6 (lateralized to right), and the insula. Isovaleric acid activated four of the five regions without lateralization; no BA 8 activity was seen. Clementine produced more activity than isovaleric acid in the left insula, and isovaleric acid produced more activity than clementine in the left BA 6. No activation was detected in the orbitofrontal cortex or in the medial temporal lobes. Subjects rated clementine, isovaleric acid, and clear air as being pleasant, unpleasant, and neutral, respectively. CONCLUSION: Activation in frontal regions may represent brain processes linked to olfactory networks. There may be regional specialization based on odorant hedonic values.

Startle reflex modulation by pleasant and unpleasant odors in a between-subjects design.

Participants were presented with an acoustic startle probe while they smelled either a pleasant odor or an unpleasant odor (n = 40 per condition). Within each condition, participants were presented with four blocks in which odor was present and four in which odor was not present (no-odor control). Significant differences were found in both conditions between the odor and no-odor blocks. In the unpleasant odor condition, blink magnitude was greater during the odor blocks; in the pleasant odor condition, blink magnitude was smaller during the odor blocks.

The contingent negative variation in an odor labeling paradigm.

Fifteen subjects participated in an experiment designed to assess the contingent negative variation (CNV) during the labeling of odors and shapes. Odors or shapes were presented (S1) and followed 3 s later by a lexical label (A, B, or C) (S2). In 75% of the trials, the S2 was the correct label for the odor or shape. In the remaining trials, the S2 was an incorrect label. Subjects' olfactory performance was correlated with both the CNV during in the S1/S2 interval and also the P300 following the S2 stimulus. The CNV over the left frontal area was significantly larger in the olfactory phase of the experiment. CNV activity also correlated with olfactory performance such that subjects with the largest odor-related CNVs had the best olfactory performance. Although P300 differed as a function of label matches versus mismatches, no odor-specific effects or correlations were found.

Startle reflex modulation during exposure to pleasant and unpleasant odors.

The eyeblink component of the startle response to acoustic probes was measured while subjects smelled pleasant odors, unpleasant odors, or no odor. Peak electromyogram (EMG) 20-90 ms after probe onset was greater during unpleasant than during no odor conditions; EMGs for pleasant odors did not differ from those for no odor. Base tension in orbicularis oculi muscles was also higher during unpleasant odors. The results for unpleasant odors parallel those found in previous studies that have used a variety of methods for producing negatively valenced experiences. The results for pleasant odors diverge from those of most previous studies, which have found startle attenuation during positively valenced experiences. However, the results may be compatible with the suggestion (Lang, Bradley, & Cuthbert, 1992) that the motivational state that modulates startle involves both valence and arousal. Pleasant odors may require appropriate contexts to produce the positive motivational state required for startle attenuation.

Decreased perceptual sensitivity to emotion-evoking stimuli in depression.

We measured two aspects of emotional response in depressed patients, as a preliminary study of the potential usefulness of such measures for elucidating pathophysiological mechanisms. First we used electromyography to measure the automatic mimicry on an individual's own face of facial displays of emotion observed on the faces of others. Next we used the fused dichotic listening paradigm to measure selective perception of both positive and negative emotion-related words as opposed to neutral words. Patients failed to show the normal facial mimicry of both positive and negative facial displays, despite normal cognitive processing of the stimuli. They also heard significantly fewer positive and negative words on the dichotic tests than did healthy controls. This suggests that depressed patients are hyposensitive to emotion-related stimuli in general.

EEG and EMG responses to emotion-evoking stimuli processed without conscious awareness.

Dichotic stimulus pairs were constructed with one word that was emotionally neutral and another that evoked either negative or positive feelings. Temporal and spectral overlap between the members of each pair was so great that the two words fused into a single auditory percept. Subjects were consciously aware of hearing only one word from most pairs; sometimes the emotion-evoking word was heard consciously, other times the neutral word was heard consciously. Subjects were instructed to let their thoughts wander in response to the word they heard, during which time EEG alpha activity over left and right frontal regions, and muscle activity (EMG) in the corrugator ("frowning") and zygomatic ("smiling") regions were recorded. Both EEG and EMG provided evidence of emotion-specific responses to stimuli that were processed without conscious awareness. Moreover both suggested relatively greater right hemisphere activity with unconscious rather than conscious processing.

Defensive coping and blood pressure reactivity in medical patients.

Two defensive coping styles, denial of illness and repressive coping, were studied in two groups of medical patients whose blood pressure (BP) was measured during a stress interview. Denial of illness was measured using the Levine Denial of Illness Scale (LDIS), and repressive coping was measured using a combination of the Marlowe-Crowne (MC) Social Desirability Scale and the SCL-90R anxiety subscale (ANX). Consistent with our prior research indicating that LDIS was associated with adaptive outcomes in the short run, high deniers manifested reduced systolic BP reactivity compared to low deniers. Although not related to repressive coping, systolic BP reactivity was correlated positively with MC and ANX separately. The results demonstrate that LDIS and MC measure different types of defensive coping. Current theories of the MC scale suggest two possible interpretations of the MC findings, one that focuses on avoidant coping and the second on attentional coping in high MC scorers.

The role of denial in recovery from coronary heart disease.

This longitudinal study investigated the relationship between denial of illness and the course of recovery in patients with coronary heart disease. Using a newly developed interview instrument, the Levine Denial of Illness Scale (LDIS), the level and modes of denial were assessed in 45 male patients who were hospitalized for myocardial infarction or for coronary bypass surgery, of whom 30 were followed for 1 year after discharge. The reliability, internal consistency, and validity of the LDIS were found to be satisfactory. Furthermore, the LDIS showed discriminant validity from trait measures of denial. LDIS scores were not associated with severity of illness or risk factors. High deniers spent fewer days in intensive care and had fewer signs of cardiac dysfunction during their hospitalization relative to low deniers. However, in the year following discharge, high deniers adapted more poorly than low deniers: high deniers were more noncompliant with medical recommendations and required more days of rehospitalization. The findings suggest that denial of illness is adaptive during acute hospital recovery, but is maladaptive in the long-run after hospital discharge.

Effects of emotion on perceptual asymmetry: interactions with personality.

Perceptual asymmetry on a series of four specially constructed dichotic word tests was found to change as a function of the emotional quality of the words in the tests (P = 0.05). This was most pronounced in the case of positively valued words which produced an increase in asymmetry consistent with facilitated left-hemisphere function (P less than 0.004). Changes in asymmetry with emotion differed as a function of personality characteristics of the subjects, with repressors and high anxious subjects showing an increase with emotion while true low anxious subjects showed a decrease (P less than 0.02). Personality groups also differed in asymmetry on an emotionally neutral test (P less than 0.04) and in changes in asymmetry over time independent of emotion (P less than 0.001). These data suggest that emotion mediated activation of the left hemisphere may facilitate information processing within that hemisphere. Moreover, they indicate that dichotic listening tests may provide a non-invasive and inexpensive method for assessing emotion mediated changes in brain state that are clinically relevant.

Hemisphere functioning and motor imitation in autistic persons.

Previous research has found that a high proportion of autistic individuals exhibit an atypical pattern of hemispheric specialization suggestive of impaired left hemisphere functioning: namely, right hemisphere dominance for both verbal and visual-spatial processing. Studies of brain-damaged persons have suggested that the left hemisphere is specialized for the use of nonverbal gesture. Since a major characteristic of autism is an impairment in the use of gesture, it was predicted that autistic persons would also show atypical hemispheric specialization for motor imitation. To test this hypothesis, hemispheric activation was measured using EEG recordings of alpha rhythm in autistic and matched normal control subjects during four motor imitation tasks. Autistic subjects showed significantly greater right hemisphere activation during the imitation tasks, than normal subjects. This pattern was particularly evident in younger autistic subjects and during oral, rather than manual, imitation tasks.

A comparison of somatic relaxation and EEG activity in classical progressive relaxation and transcendental meditation.

Oxygen consumption, electroencephalogram (EEG), and four other measures of somatic relaxation were monitored in groups of long-term practitioners of classical Jacobson's progressive relaxation (PR) and Transcendental Meditation (TM) and also in a group of novice PR trainees. All subjects (1) practiced relaxation or meditation (treatment), (2) sat with eyes closed (EC control), and (3) read from a travel book during two identical sessions on different days. EEG findings indicated that all three groups remained primarily awake during treatment and EC control and that several subjects in each group displayed rare theta (5-7 Hz) waveforms. All three groups demonstrated similar decrements in somatic activity during treatment and EC control which were generally of small magnitude (e. g., 2-5% in oxygen consumption). These results supported the "relaxation response" model for state changes in somatic relaxation for techniques practiced under low levels of stress but not the claim that the relaxation response produced a hypometabolic state. Despite similar state effects, the long-term PR group manifested lower levels of somatic activity across all conditions compared to both novice PR and long-term TM groups. We concluded that PR causes a generalized trait of somatic relaxation which is manifested in a variety of settings and situations. Two likely explanations for this trait were discussed: (1) PR practitioners are taught to generalize relaxation to daily activities, and/or (2) according to a "multiprocess model," PR is a "somatic technique," which should produce greater somatic relaxation than does TM, a "cognitive technique." Further research is required to elucidate these possibilities.

Sleep during transcendental meditation.

Five experienced practitioners of transcendental meditation spent appreciable parts of meditation sesions in sleep stages 2, 3, and 4. Time spent in each sleep stage varied both between sessions for a given subject and between subjects. In addition, we compare electroencephalogram records made during meditation with those made during naps taken at the same time of day. The range of states observed during meditation does not support the view that meditation produces a single, unique state of consciousness.