Central and autonomic nervous system integration in emotion

Abstract

Emotions involve physiological responses that are regulated by the brain. The present paper reviews the empirical literature on central nervous system (CNS) and autonomic nervous system (ANS) concomitants of emotional states, with a focus on studies that simultaneously assessed CNS and ANS activity. The reviewed data support two primary conclusions: (1) numerous cortical and subcortical regions show co-occurring activity with ANS responses in emotion, and (2) there may be reversed asymmetries on cortical and subcortical levels with respect to CNS/ANS interrelations. These observations are interpreted in terms of a model of neurovisceral integration in emotion, and directions for future research are presented.

Introduction

Emotions involve a complex mix of cognitive, affective, behavioral, and physiological responses (e.g., Birbaumer & Öhman, 1993; Oatley & Jenkins, 1996). At a physiological level of analysis, much current research in affective neuroscience seeks to elucidate the neural networks that underlie emotion. In this regard, results of numerous investigations of the central nervous system (CNS) concomitants of emotion suggest the involvement of multiple cortical (e.g., frontal, temporal, and parietal) and subcortical (e.g., basal ganglia, thalamus, amygdala, and hippocampus) regions across a variety of positive and negative emotions such as happiness, anxiety, anger, sadness, disgust (for reviews, see Borod, 2000; Lane & Nadel, 2000).

A separate body of psychophysiological research has sought to identify patterns of autonomic nervous system (ANS) correlates of emotion. Although several investigations have noted emotion-specific autonomic response patterns (e.g., Ekman, Levenson, & Friesen, 1983; Sinha, Lovallo, & Parsons, 1992), the majority of studies suggest greater similarities than differences among physiological activation patterns during various emotional states (for reviews see Cacioppo, Berntson, Larsen, Poehlmann, & Ito, 2000; Neumann & Waldstein, 2001; Stemmler, 1996).

Study of the linkage between central and autonomic correlates of emotion is relevant to numerous fields of investigation including affective neuroscience, neurocardiology, psychophysiology, and behavioral medicine. Yet, relatively few studies have simultaneously examined CNS and ANS responses during emotion. In this paper, we provide an overview of such investigations. Previously, it has been suggested that the right hemisphere may be dominant in eliciting autonomic responses during emotion (e.g., Borod & Madigan, 2000; Gainotti, 1989; Wittling & Roschmann, 1993). However, others have proposed more complex models of association between emotion-related central and autonomic response patterns (Lane & Jennings, 1995; Lane & Schwartz, 1987; Thayer & Lane, 2000); many of these models focus on both cortical and subcortical interconnections (such as frontal–subcortical systems). We present results from the literature on concomitant CNS and ANS response to emotional stimuli to suggest the viability of the latter position. We discuss findings from lesion studies, visual half field studies, and investigations that have directly measured CNS activity with the spontaneous electroencephalogram (EEG), evoked potentials (EPs), or neuroimaging techniques, along with measures of ANS activation. We also highlight a recent model of neurovisceral integration (Thayer & Lane, 2000) to enhance the findings of the present review. This model relies on principles of dynamical systems to suggest that different neural networks are flexibly recruited according to situational demands to integrate the central and autonomic response to emotional stimuli. We use findings from our overview of the CNS/ANS literature and the theoretical position of the model of neurovisceral integration to provide suggestions for future research directions.

Prior to the review of relevant studies, some remarks on the definition of “emotion” and additional technical issues may be helpful. First, defining “emotion” has been a controversial issue, both historically (e.g., Epstein, 1984), and currently (e.g., Scherer, 2000). For the purpose of the present review, we use a working definition that focuses on the functional aspects of emotions (e.g., Frijda, 1986, Frijda, 1988; Levenson, 1988; see Thayer & Lane, 2000): Emotions may be characterized as an organismic response to an environmental event that facilitates the rapid mobilization for action. This response involves multiple systems of the organism, such as cognitive, behavioral, and autonomic sub-systems. When these response systems are efficiently coordinated, they allow for goal-directed behavior in the service of flexible adaptation of the organism to changing environmental demands.

Second, several researchers have classified emotional responses in terms of a number of discrete “basic emotions” such as surprise, interest, happiness, rage, fear, sadness, and disgust (e.g., Ekman, 1984; Izard, 1977; Tomkins, 1962). In contrast, other investigators have argued that emotions—at least on the level of subjective experience—may be more parsimoniously described by only two dimensions, which are “valence” (pleasant–unpleasant) and “arousal” (low–high intensity) (e.g., Larsen & Diener, 1992; Russell, 1980). Several researchers have suggested that these two dimensions are hierarchically related to discrete emotions (e.g., Diener, Smith, & Fujita, 1995; Russell & Barrett, 1999). Nonetheless, these different theoretical orientations have given rise to different emotion induction procedures for experimental research (for general overviews, see Gerrards-Hesse, Spies, & Hesse, 1994; Oatley & Jenkins, 1996), and concomitant physiology has been contrasted in various experiments between basic emotions, or between pleasant and unpleasant states. For example, a variety of discrete emotions such as happiness, sadness, anger, fear, and disgust have been elicited with procedures such as presentation of film clips (Lane, Reiman, Ahern, & Thayer, 2000), presentation of pictures of respective facial expressions (Schneider et al., 1995), recall and re-experience of personal life episodes (Waldstein et al., 2000), anticipation of electric shock (Slomine, Bowers, & Heilman, 1999), or hypnotic induction (de Pascalis, Ray, Tranquillo, & D’Amico, 1998). In contrast, pleasant and unpleasant emotions have typically been induced by the presentation of affective slides (e.g., Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Lane et al., 1997; Palomba, Angrilli, & Mini, 1997).

Third, the studies described in this review have used a variety of variables to measure ANS activity. The ANS has two main branches—the sympathetic and parasympathetic (vagal) nervous systems—that innervate visceral organs, blood vessels, and glands, and which exert opposite effects on the innervated organs (for an overview, see Lovallo & Sollers, 2000). The most frequently used measure of ANS activity is heart rate (HR), which is antagonistically affected by sympathetic and parasympathetic activity. This dual influence renders considerable ambiguity to the interpretation of HR responses, but inclusion of other cardiovascular measures can help to avoid this problem: HR variability and particularly its high frequency component primarily reflect cardiac parasympathetic activity whereas measures of myocardial performance indicate primarily β-adrenergic sympathetic activity. Another widely used ANS measure is blood pressure, which is also influenced by sympathetic and parasympathetic activity, and skin conductance response (SCR), which primarily reflects sympathetic activity.