Elsevier

NeuroImage

Volume 23, Issue 1, September 2004, Pages 64-74
NeuroImage

Dissociable effects of arousal and valence on prefrontal activity indexing emotional evaluation and subsequent memory: an event-related fMRI study

https://doi.org/10.1016/j.neuroimage.2004.05.015Get rights and content

Abstract

Prefrontal cortex (PFC) activity associated with emotional evaluation and subsequent memory was investigated with event-related functional MRI (fMRI). Participants were scanned while rating the pleasantness of emotionally positive, negative, and neutral pictures, and memory for the pictures was tested after scanning. Emotional evaluation was measured by comparing activity during the picture rating task relative to baseline, and successful encoding was measured by comparing activity for subsequently remembered versus forgotten pictures (Dm effect). The effect of arousal on these measures was indicated by greater activity for both positive and negative pictures than for neutral ones, and the effect of valence was indicated by differences in activity between positive and negative pictures. The study yielded three main results. First, consistent with the valence hypothesis, specific regions in left dorsolateral PFC were more activated for positive than for negative picture evaluation, whereas regions in right ventrolateral PFC showed the converse pattern. Second, dorsomedial PFC activity was sensitive to emotional arousal, whereas ventromedial PFC activity was sensitive to positive valence, consistent with evidence linking these regions, respectively, to emotional processing and self-awareness or appetitive behavior. Finally, successful encoding (Dm) activity in left ventrolateral and dorsolateral PFC was greater for arousing than for neutral pictures. This finding suggests that the enhancing effect of emotion on memory formation is partly due to an augmentation of PFC-mediated strategic, semantic, and working memory operations. These results underscore the critical role of PFC in emotional evaluation and memory, and disentangle the effects of arousal and valence across PFC regions associated with different cognitive functions.

Introduction

The domain of cognitive neuroscience of emotion has grown dramatically during the last decade. As a result of this development, various effects of emotion on brain activation associated with different perceptual and cognitive functions have been revealed (reviewed in Davidson and Irwin, 1999, Lane and Nadel, 2000, Phan et al., 2002). To understand these diverse effects, researchers have often divided emotion into its basic underlying constructs. One dimensional approach to emotion emphasizes the contribution of two orthogonal components, namely arousal and valence Lang et al., 1993, Russell, 1980. Arousal refers to a continuum that varies from calm to excitement, whereas valence refers to a continuum that varies from positive to negative with neutral in the middle (for methods to assess these two dimensions, see Bradley and Lang, 1994). The vast majority of studies has focused on the limbic system and particularly on the amygdala, whereas other components of the emotional processing network, such as the prefrontal cortex (PFC), have received relatively less attention. Recent studies of amygdala function have attempted to tease apart the relative contributions of the aforementioned affective dimensions to task performance (e.g., Anderson et al., 2003, Hamann et al., 2002, Phelps and Anderson, 1997, Phelps et al., 1998). However, the contribution of these factors to emotional processing in other frontolimbic regions is not well understood, and the available evidence is contradictory. To address this imbalance, the present functional MRI (fMRI) study focused on the role of PFC regions in emotional processing.

In particular, we investigated the effects of arousal and valence on emotional evaluation and emotional memory. Emotional evaluation refers to the perception and categorization of emotional stimuli, and emotional memory refers to the modulatory effect of emotion on different stages of memory processing, including encoding, consolidation, and retrieval. In the domain of emotional evaluation, the amygdala is assumed to be involved in the rapid detection of the basic emotional properties of incoming stimuli, whereas PFC is assumed to be involved in higher-order emotional evaluation processes, which operate in close interaction with other cognitive functions and with behavioral goals (Davidson and Irwin, 1999). In the domain of emotional memory, the existing studies have focused on the amygdala and identified arousal-mediated effects at encoding that predict subsequent memory Cahill et al., 1996, Canli et al., 2000, Canli et al., 2002, Dolcos et al., 2003, Dolcos et al., 2004, Hamann et al., 1999, but there is little understanding of the contribution of other brain regions, such as PFC regions. Although it is assumed that activation during emotional evaluation tasks plays a role in memory, few studies have explicitly examined the relationship between emotional evaluation and memory. Thus, the overarching goal of the present study was to carefully investigate the contribution of the PFC to arousal and valence effects on emotional evaluation and memory.

Different PFC subregions are likely to make distinct contributions to emotional evaluation, but information about this issue is scarce. A basic anatomical distinction in this domain is between lateral and medial PFC regions. According to one prevailing view, the role of lateral PFC regions in emotional evaluation is primarily related to valence. The valence hypothesis states that the left PFC is dominant in the processing of positive emotions, whereas the right PFC is dominant in the processing of negative emotions Davidson, 1995, Davidson and Irwin, 1999. This hypothesis is inspired by evidence from lesion literature and is mainly supported by electrophysiological evidence from EEG recordings. Neuropsychological evidence shows that patients with left hemisphere lesions tend to experience negative emotions, such as sadness Morris et al., 1996, Paradiso et al., 1999a, whereas patients with right hemisphere damage are biased toward experiencing positive emotions, such as euphoria (e.g., Starkstein et al., 1989). The results of some electrophysiological studies are consistent with the valence hypothesis and support the idea that this valence-related PFC lateralization may depend either on transiently induced affective states or on stable personality traits Aftanas et al., 2001, Davidson, 1995, Davidson and Irwin, 1999, Tomarken et al., 1992, Wheeler et al., 1993.

However, electrophysiological studies do not provide an accurate localization of the sources of these valence effects, and, overall, the evidence supporting the valence hypothesis has been mixed. First, neither lesion nor electrophysiological evidence has always been consistent with the valence hypothesis (e.g., Borod, 1992, Borod et al., 1998, Dolcos and Cabeza, 2002, Hagemann et al., 1998). Second, functional neuroimaging evidence is also inconclusive. Whereas some studies support the valence hypothesis (e.g., Canli et al., 1998), some studies do not report valence-related hemispheric asymmetry in PFC (e.g., Baker et al., 1997, George et al., 1995, Lane et al., 1997a, Lane et al., 1997b, Lane et al., 1997c, Pardo et al., 1993, Teasdale et al., 1999). One possible reason why the results have been mixed is that arousal and valence are often not distinguished carefully (but see Canli et al., 1998); hence, valence effects might have been confounded with arousal effects. Thus, the first goal of the present study was to investigate the valence hypothesis and identify the specific PFC regions involved, in conditions where positive and negative stimuli were matched in arousal and other potentially confounding factors were controlled.

As for the role of medial PFC regions in emotional evaluation, different hypotheses have been suggested. For instance, orbitofrontal areas of medial PFC have been linked to the rewarding nature of stimuli (e.g., O'Doherty et al., 2001, Rolls, 2000), and anteromedial areas have been related to more personal and subjective aspects of experiencing internal states (e.g., Frith and Frith, 1999). Although medial PFC regions have been strongly associated with emotional processing, it is unclear whether the role of these regions is related to arousal or to valence. Given that medial PFC regions are systematically activated by emotional stimuli, regardless of their valence (for a review, see Phan et al., 2002), PFC involvement could be attributed to its role in the processing of arousal. This notion is consistent with evidence supporting medial PFC involvement in processing emotionally arousing stimuli irrespective of valence Lane et al., 1997a, Lane et al., 1997b, Lane et al., 1997c, Reiman, 1997, Reiman et al., 1997, Schneider et al., 1995, Teasdale et al., 1999. On the other hand, there is also evidence suggesting valence-related specificity in medial PFC (e.g., George et al., 1995, Paradiso et al., 1999b). In particular, medial PFC has been associated with affiliative behaviors and appetitive or reward circuits (e.g., Rolls, 2000). A recent metaanalysis of functional neuroimaging studies of emotion (Wager et al., 2003) found that, overall, medial PFC activity was associated with approach or appetitive tasks. To address this issue, the second goal of the present study was to determine whether the role of medial PFC in emotional processing is primarily related to arousal or to valence, or whether there are subregions within medial PFC differently involved in arousal and valence.

Turning to emotional memory, the most basic phenomenon to explain in this domain is why arousing events (both positive and negative) are better remembered than neutral events Bradley et al., 1992, Christianson, 1992. This effect has been attributed to the modulatory effect of the amygdala on the medial temporal lobe (MTL) memory system (McGaugh et al., 2002), and this modulation hypothesis has been confirmed by functional neuroimaging studies Cahill et al., 1996, Canli et al., 2000, Canli et al., 2002, Dolcos et al., 2003, Dolcos et al., 2004, Hamann et al., 1999, Kilpatrick and Cahill, 2003. For example, we investigated this hypothesis using event-related fMRI and the subsequent memory paradigm (Paller and Wagner, 2002). In this paradigm, memory performance on a subsequent memory task is used to sort encoding items into two categories: remembered versus forgotten. Greater encoding activity for remembered than forgotten items, sometimes known as “the Dm (difference in memory) effect”, is assumed to reflect successful encoding operations. Consistent with the modulation hypothesis, we found that the Dm effects in the amygdala and the MTL memory regions were greater for emotionally arousing pictures than for neutral pictures, and that the two Dm effects were more strongly correlated in the case of arousing pictures than in the case of neutral pictures (Dolcos et al., 2004).

Although this evidence strongly links the memory-enhancing effect of emotion to an MTL mechanism, it does not exclude the possibility that other brain regions, such as PFC, also play a major role. In fact, in functional neuroimaging studies, PFC regions are as strongly associated with successful encoding operations as MTL regions (e.g., Brewer et al., 1998, Paller and Wagner, 2002, Wagner et al., 1998). Also, the effects of several factors affecting encoding success, such as organizational strategies and attention, have been found to be mediated by changes in PFC activity Anderson et al., 2000, Fletcher et al., 1998, Kensinger et al., 2003. Moreover, studies using transcranial magnetic stimulation (TMS) have shown that PFC activity is actually necessary for successful encoding Epstein et al., 2002, Grafman and Wassermann, 1999, Rossi et al., 2001. Thus, it is quite likely that the enhancing effect of emotion on encoding is mediated not only by MTL but also by PFC. Yet, very little is known about the role of PFC on emotional memory formation (see, however, Canli et al., 2002, Kilpatrick and Cahill, 2003). For example, it is uncertain if the Dm effect in PFC is enhanced by emotion, similar to what we found in MTL (Dolcos et al., 2004). It is also unclear whether this putative effect is due to arousal or valence, and which specific PFC regions are involved. Thus, the third goal of the study was to investigate the role of PFC in the formation of emotional memory.

The method we employed has two main features: (1) it distinguishes between activity associated with emotional evaluation and emotional memory, and (2) it distinguishes between the effects of arousal and valence. Participants were scanned while rating the pleasantness of arousing pictures (positive and negative) and non-arousing pictures (neutral), and after scanning, they recalled the contents of the pictures. Stimuli were selected from a standardized set of pictures that allows experimental control over arousal and valence characteristics (Lang et al., 1997), which has been largely used in neuroimaging studies of emotion (e.g., Dolan et al., 2000, Hamann et al., 1999, Hamann et al., 2002, Lane et al., 1997a, Lane et al., 1997b, Lane et al., 1997c, Lane et al., 1999, Liberzon et al., 2000, Paradiso et al., 1999b, Taylor et al., 1998, Taylor et al., 2000). Evaluation activity was measured by comparing activity during picture rating to the baseline activity, and successful encoding activity was measured by comparing activity for subsequently remembered versus subsequently forgotten pictures (Dm effect). Given that positive and negative pictures were both more arousing than neutral pictures, arousal effects should affect both positive and negative pictures. Given that positive and negative pictures were matched in arousal, differences between them should reflect valence effects rather than arousal effects. Thus, the effect of arousal was defined as greater activity for both positive and negative pictures than for neutral pictures, and the effect of valence, as differences between activity for positive and negative pictures.

To summarize, we investigated three main issues. First, we investigated the valence hypothesis, and in particular, what specific left and right PFC subregions would be sensitive to valence effects. Second, we investigated the role of medial PFC in emotional evaluation, and specifically, whether activity in this region is primarily sensitive to arousal or valence, or whether subregions can be distinguished. Finally, we investigated the role of PFC in emotional memory, and particularly, the relationship of Dm effects with stimulus arousal and valence, and their localization within PFC.

Section snippets

Subjects

Sixteen young (25 + 4.6 years), right-handed, healthy women participated in the study. Female participants were chosen because evidence suggests that they are more likely to display strong physiological responses to emotional stimuli (Lang et al., 1993) and report more intense emotional experiences (Shields, 1991) than men. All participants consented to a protocol approved by Duke University Institutional Review Board.

Materials

Stimuli consisted of 60 positive, 60 negative, and 60 neutral pictures

Valence ratings

The average valence scores (1 = negative, 2 = neutral, 3 = positive) as rated by the participants in the scanner were 1.14 (SD = 0.16) for negative pictures, 2.18 (SD = 0.40) for neutral pictures, and 2.64 (SD = 0.26) for positive pictures. All pairwise comparisons were significant (P < 0.0001). Thus, the subjects' rating scores were consistent with those provided in the IAPS norms (Lang et al., 1997). Further validating this consistency, the correlation between our subjects' average scores and

Discussion

The present study yielded three main findings relevant for understanding PFC contributions to emotional evaluation and memory. First, consistent with the valence hypothesis, during emotional evaluation, specific left dorsolateral PFC areas showed greater activity for positive than for negative pictures, whereas right ventrolateral PFC areas showed the converse pattern. Second, also during emotional evaluation, dorsomedial PFC activity was sensitive to arousal (greater activation for both

Conclusions

Using an fMRI paradigm that distinguished between activity related to emotional evaluation and emotional memory and between the effects of arousal and valence, the present study yielded three main results. First, during emotional evaluation, PFC activity showed a hemispheric asymmetry consistent with the valence hypothesis. A left dorsolateral PFC region was sensitive to positive valence, possibly reflecting the maintenance of positive information in working memory, whereas a right

Acknowledgements

We thank David Beckmann for assistance with data analysis. This study was supported by NIH grants R01 AG19731 (RC) and R01 DA14094 (KSL), and a NARSAD Young Investigator Award (KSL). FD was supported by a Chia PhD Scholarship and a Dissertation Fellowship from the University of Alberta (Canada), and a Research Assistantship from Duke University (USA).

References (90)

  • R.D. Lane et al.

    Common effects of emotional valence, arousal and attention on neural activation during visual processing of pictures

    Neuropsychologia

    (1999)
  • I. Liberzon et al.

    Limbic activation and psychophysiologic responses to aversive visual stimuli. Interaction with cognitive task

    Neuropsychopharmacology

    (2000)
  • J.L. McGaugh et al.

    Amygdala modulation of memory consolidation: interaction with other brain systems

    Neurobiol. Learn. Mem.

    (2002)
  • K.A. Paller et al.

    Observing the transformation of experience into memory

    Trends Cogn. Sci.

    (2002)
  • K.L. Phan et al.

    Functional neuroanatomy of emotion: a meta-analysis of emotion activation studies in PET and fMRI

    NeuroImage

    (2002)
  • E.A. Phelps et al.

    Emotional memory: what does the amygdala do?

    Curr. Biol.

    (1997)
  • R.A. Poldrack et al.

    Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex

    NeuroImage

    (1999)
  • F. Schneider et al.

    Mood effects on limbic blood flow correlate with emotional self-rating: a PET study with oxygen-15 labeled water

    Psychiatry Res.

    (1995)
  • S.F. Taylor et al.

    The effect of emotional content on visual recognition memory: a PET activation study

    NeuroImage

    (1998)
  • S.F. Taylor et al.

    The effect of graded aversive stimuli on limbic and visual activation

    Neuropsychologia

    (2000)
  • T.D. Wager et al.

    Valence, gender, and lateralization of functional brain anatomy in emotion: a meta-analysis of findings from neuroimaging

    NeuroImage

    (2003)
  • L. Aftanas et al.

    Event-related synchronization and desynchronization during affective processing: emergence of valence-related time-dependent hemispheric asymmetries in theta and upper alpha band

    Int. J. Neurosci.

    (2001)
  • N.D. Anderson et al.

    The effects of divided attention on encoding- and retrieval-related brain activity: a PET study of younger and older adults

    J. Cogn. Neurosci.

    (2000)
  • A.K. Anderson et al.

    Dissociated neural representations of intensity and valence in human olfaction

    Nat. Neurosci.

    (2003)
  • S.C. Baker et al.

    The interaction between mood and cognitive function studied with PET

    Psychol. Med.

    (1997)
  • J.C. Borod

    Interhemispheric and intrahemispheric control of emotion: a focus on unilateral brain damage

    J. Consult. Clin. Psychol.

    (1992)
  • J.C. Borod et al.

    Right hemisphere emotional perception: evidence across multiple channels

    Neuropsychology

    (1998)
  • M.M. Bradley et al.

    Remembering pictures: pleasure and arousal in memory

    J. Exp. Psychol., Learn, Mem. Cogn.

    (1992)
  • J.B. Brewer et al.

    Making memories: brain activity that predicts how well visual experience will be remembered

    Science

    (1998)
  • R. Cabeza et al.

    Imaging cognition: II. An empirical review of 275 PET and fMRI studies

    J. Cogn. Neurosci.

    (2000)
  • R. Cabeza et al.

    Brain activity during episodic retrieval of autobiographical and laboratory events: an fMRI study using a novel photo paradigm

    J. Cogn. Neurosci.

    (2004)
  • L. Cahill et al.

    Amygdala activity at encoding correlated with long-term, free recall of emotional information

    Proc. Natl. Acad. Sci. U. S. A.

    (1996)
  • T. Canli et al.

    Hemispheric asymmetry for emotional stimuli detected with fMRI

    NeuroReport

    (1998)
  • T. Canli et al.

    Event-related activation in the human amygdala associated with later memory for individual emotional experience

    J. Neurosci.

    (2000)
  • T. Canli et al.

    Sex differences in the neural basis of emotional memories

    Proc. Natl. Acad. Sci. U. S. A.

    (2002)
  • S.-A. Christianson

    The Handbook of Emotion and Memory: Research and Theory

    (1992)
  • R.J. Davidson

    Cerebral asymmetry, emotion and affective style

  • R.J. Davidson et al.

    The functional neuroanatomy of emotion and affective style

    Trends Cogn. Sci.

    (1999)
  • M. D'Esposito et al.

    Prefrontal cortical contributions to working memory: evidence from event-related fMRI studies

    Exp. Brain Res.

    (2000)
  • Dillon, D.G., Cooper, J.J., Grent-t-Jong, T., Woldorff, M.G., LaBar, K.S. Dissociation of event-related potentials...
  • F. Dolcos et al.

    Event-related potentials of emotional memory: encoding pleasant, unpleasant, and neutral Pictures

    Cogn. Affect Behav. Neurosci.

    (2002)
  • F. Dolcos et al.

    Coactivation of the amygdala and hippocampus predicts better recall for emotional than for neutral pictures

    Brain and Cognition

    (2003)
  • P.C. Fletcher et al.

    The functional roles of prefrontal cortex in episodic memory: I. Encoding

    Brain

    (1998)
  • N.H. Frijda

    The laws of emotion

    Am. Psychol.

    (1988)
  • C.D. Frith et al.

    Interacting minds—A biological basis

    Science

    (1999)
  • Cited by (344)

    View all citing articles on Scopus
    View full text