We think Lindquist et al.'s model would benefit from a more evolutionary and developmental perspective (cf. Tinbergen Reference Tinbergen1951). Attention to emotion in infants and other species would highlight the continuity in emotional mechanisms from ancestral species and infants to adult humans. Then, more weight might be given to limbic system mechanisms and less to cortical processes. Certainly, people use the cortex to identify emotionally relevant stimuli on the basis of conceptual learning and cognitive interpretation, but the limbic system reacts emotionally to unconditioned stimuli such as tastes and facial expressions, to conditioned stimuli, and to stimulus familiarity (Campos & Barrett Reference Campos, Barrett, Izard, Kagan and Zajonc1984; Zajonc Reference Zajonc, Scherer and Ekman1994).
The authors assert that “emotion words are essential to our model” (sect. 3, para. 7) and claim that infants possess enough language capacity to experience emotion, but many emotions emerge by 9 months of age. Infant emotions, such as fear, unfold through epigenetic programs according to a precise, universal timetable (LaFreniere Reference LaFreniere2000; Reference LaFreniere2010; Sroufe Reference Sroufe1996) well before the onset of language. Young infants rely mainly on subcortical behavioral mechanisms, not language (infans, L., without speech), for registering and communicating their specific emotional needs.
Likewise, vertebrates depend heavily on unlearned, limbic system responses to execute specific emotional behaviors such as fighting a rival (Butler & Hodos Reference Butler and Hodos2005; Tinbergen Reference Tinbergen1951). Identifying particular adaptive needs is not completely surrendered to the vagaries of cognitive interpretation and experience. Reptiles, lacking a neocortex and abstract cognitive abilities, nevertheless possess emotional systems for flight, attack, and feeding. Specific emotional behaviors and brain structures show considerable continuity across mammals (Panksepp Reference Panksepp1998). Mammals possess reward and punishment limbic areas (Olds & Milner Reference Olds and Milner1954) homologous to structures that, when stimulated in conscious patients, elicit reports of specific affects (anger: King Reference King and Sheer1961; fear: Gloor Reference Gloor1997; humor appreciation: Black Reference Black1982; Martin Reference Martin1950). Specific affects also arise during psychomotor seizures (Gloor Reference Gloor1997).
The fact that the brain often instantly registers specific affects weakens the claim that interpretation is necessary for affective experience. The authors refer to interpretation of sensory feedback from the body and endorse the James–Lange position that affects occur after the motivated behavior and accompanying visceral changes. Presumably, discrete affects evolved to direct adaptive motivated behavior, so there would be little benefit from telling the brain what it should do after it has acted. In James' example, why would fleeing a bear be necessary for fear – if one froze in terror, would one not be afraid? People may reflect on their fear after escaping, but typically report having been terrorized immediately.
Cannon (Reference Cannon1927) and Bard (Reference Bard1928) challenged James–Lange with laboratory research showing that affect precedes visceral adjustments. Visceral feedback takes several seconds – but affects are instantaneous, so visceral feedback cannot be the primary source of affect. In spinally transected experimental animals and patients, behavioral responses and affective self-reports are appropriate for the situation. Engaging in exercise with its visceral changes does not elicit particular affects. Of course, bodily input initiates some affects, such as low blood glucose or stomach contractions eliciting hunger. But hunger is not triggered following interpretation of feeding or visceral adjustments to food intake.
Recent evidence supporting Cannon and Bard comes from LeDoux (Reference LeDoux1996). A mammal receives emotionally salient information through sensory systems that activate the thalamus. There, a quick appraisal of the information takes place. If the stimulus constitutes a conditioned or unconditioned stimulus of, say, fear, a rat exhibits the panoply of fearful behaviors. The central nucleus of the amygdala orchestrates the various facets of the emotional response. The central gray activates the overt behavior of freezing, and the hypothalamus initiates visceral adjustments such as increased heart rate. But fear comes first: the affectively sensitive amygdala presumably registers fear before these behavioral and visceral events occur. LeDoux's model does not include feedback from the voluntary muscles or viscera to the limbic system. With the neocortex removed, a rat will still respond to appropriate fear releasers or conditioned stimuli. With the neocortex intact, it makes finer discriminations of these stimuli; it may not exhibit fear of a tone different from that which it was conditioned to fear. The neocortex also allows finer motoric responses. So the neocortex only refines emotional behavior, but is not the essential mechanism, which is limbic.
But why is so much of the brain, including the neocortex, activated by an emotional stimulus? Observable behavior consists of addressing specific emotional needs seriatim. Like other animals, we feed, sleep, mate, defend ourselves, compete, and so forth through the day. We use our neocortex to perform these behaviors more efficiently, but our behavior is motivated by the limbic system. Because it is for the fulfillment of biological needs that, arguably, our brains evolved, large areas of the brain, including language areas, are mobilized by emotional stimuli, especially verbal ones. We selectively attend to, recall, and react to emotionally salient stimuli. These reactions include not only affect but also visceral, hormonal, expressional, cognitive, mnemonic, and overt behavioral responses. This widespread activation of the brain makes it difficult to pinpoint affective experience, especially given the technical difficulties of neuroimaging subcortical structures. Affects may be better localized by brain stimulation, experimental lesions, and clinical research.
Another problem is that the brain structures studied are highly differentiated. The amygdala has 15 nuclei (Gloor Reference Gloor1997) and is implicated in aggression and sexual behavior, not just fear. The orbitofrontal cortex is involved in pride and shame (Fuster Reference Fuster1997), as well as anger. Finer anatomical analysis might reveal more emotional specificity.
Also problematical are the emotion terms in Lindquist et al.'s meta-analysis. Whereas the amygdala is activated in fear, fear of giving a speech deactivates the amygdala. This is understandable if one thinks of the amygdala as mediating fear of bodily harm, not fear of any unpleasant outcome, such as embarrassment in this case (which might activate the orbitofrontal cortex). Many researchers relied on Ekman and Friesen's (Reference Ekman and Friesen1971) list of six emotions with universal facial expressions. Other emotions have no distinct facial expression, as Ekman (Reference Ekman, Ekman and Davidson1994a) acknowledged, and some facial expressions – happiness and sadness – can be observed following any pleasant or unpleasant experience. More specific localization might result from using more precise emotion terms.
We think Lindquist et al.'s model would benefit from a more evolutionary and developmental perspective (cf. Tinbergen Reference Tinbergen1951). Attention to emotion in infants and other species would highlight the continuity in emotional mechanisms from ancestral species and infants to adult humans. Then, more weight might be given to limbic system mechanisms and less to cortical processes. Certainly, people use the cortex to identify emotionally relevant stimuli on the basis of conceptual learning and cognitive interpretation, but the limbic system reacts emotionally to unconditioned stimuli such as tastes and facial expressions, to conditioned stimuli, and to stimulus familiarity (Campos & Barrett Reference Campos, Barrett, Izard, Kagan and Zajonc1984; Zajonc Reference Zajonc, Scherer and Ekman1994).
The authors assert that “emotion words are essential to our model” (sect. 3, para. 7) and claim that infants possess enough language capacity to experience emotion, but many emotions emerge by 9 months of age. Infant emotions, such as fear, unfold through epigenetic programs according to a precise, universal timetable (LaFreniere Reference LaFreniere2000; Reference LaFreniere2010; Sroufe Reference Sroufe1996) well before the onset of language. Young infants rely mainly on subcortical behavioral mechanisms, not language (infans, L., without speech), for registering and communicating their specific emotional needs.
Likewise, vertebrates depend heavily on unlearned, limbic system responses to execute specific emotional behaviors such as fighting a rival (Butler & Hodos Reference Butler and Hodos2005; Tinbergen Reference Tinbergen1951). Identifying particular adaptive needs is not completely surrendered to the vagaries of cognitive interpretation and experience. Reptiles, lacking a neocortex and abstract cognitive abilities, nevertheless possess emotional systems for flight, attack, and feeding. Specific emotional behaviors and brain structures show considerable continuity across mammals (Panksepp Reference Panksepp1998). Mammals possess reward and punishment limbic areas (Olds & Milner Reference Olds and Milner1954) homologous to structures that, when stimulated in conscious patients, elicit reports of specific affects (anger: King Reference King and Sheer1961; fear: Gloor Reference Gloor1997; humor appreciation: Black Reference Black1982; Martin Reference Martin1950). Specific affects also arise during psychomotor seizures (Gloor Reference Gloor1997).
The fact that the brain often instantly registers specific affects weakens the claim that interpretation is necessary for affective experience. The authors refer to interpretation of sensory feedback from the body and endorse the James–Lange position that affects occur after the motivated behavior and accompanying visceral changes. Presumably, discrete affects evolved to direct adaptive motivated behavior, so there would be little benefit from telling the brain what it should do after it has acted. In James' example, why would fleeing a bear be necessary for fear – if one froze in terror, would one not be afraid? People may reflect on their fear after escaping, but typically report having been terrorized immediately.
Cannon (Reference Cannon1927) and Bard (Reference Bard1928) challenged James–Lange with laboratory research showing that affect precedes visceral adjustments. Visceral feedback takes several seconds – but affects are instantaneous, so visceral feedback cannot be the primary source of affect. In spinally transected experimental animals and patients, behavioral responses and affective self-reports are appropriate for the situation. Engaging in exercise with its visceral changes does not elicit particular affects. Of course, bodily input initiates some affects, such as low blood glucose or stomach contractions eliciting hunger. But hunger is not triggered following interpretation of feeding or visceral adjustments to food intake.
Recent evidence supporting Cannon and Bard comes from LeDoux (Reference LeDoux1996). A mammal receives emotionally salient information through sensory systems that activate the thalamus. There, a quick appraisal of the information takes place. If the stimulus constitutes a conditioned or unconditioned stimulus of, say, fear, a rat exhibits the panoply of fearful behaviors. The central nucleus of the amygdala orchestrates the various facets of the emotional response. The central gray activates the overt behavior of freezing, and the hypothalamus initiates visceral adjustments such as increased heart rate. But fear comes first: the affectively sensitive amygdala presumably registers fear before these behavioral and visceral events occur. LeDoux's model does not include feedback from the voluntary muscles or viscera to the limbic system. With the neocortex removed, a rat will still respond to appropriate fear releasers or conditioned stimuli. With the neocortex intact, it makes finer discriminations of these stimuli; it may not exhibit fear of a tone different from that which it was conditioned to fear. The neocortex also allows finer motoric responses. So the neocortex only refines emotional behavior, but is not the essential mechanism, which is limbic.
But why is so much of the brain, including the neocortex, activated by an emotional stimulus? Observable behavior consists of addressing specific emotional needs seriatim. Like other animals, we feed, sleep, mate, defend ourselves, compete, and so forth through the day. We use our neocortex to perform these behaviors more efficiently, but our behavior is motivated by the limbic system. Because it is for the fulfillment of biological needs that, arguably, our brains evolved, large areas of the brain, including language areas, are mobilized by emotional stimuli, especially verbal ones. We selectively attend to, recall, and react to emotionally salient stimuli. These reactions include not only affect but also visceral, hormonal, expressional, cognitive, mnemonic, and overt behavioral responses. This widespread activation of the brain makes it difficult to pinpoint affective experience, especially given the technical difficulties of neuroimaging subcortical structures. Affects may be better localized by brain stimulation, experimental lesions, and clinical research.
Another problem is that the brain structures studied are highly differentiated. The amygdala has 15 nuclei (Gloor Reference Gloor1997) and is implicated in aggression and sexual behavior, not just fear. The orbitofrontal cortex is involved in pride and shame (Fuster Reference Fuster1997), as well as anger. Finer anatomical analysis might reveal more emotional specificity.
Also problematical are the emotion terms in Lindquist et al.'s meta-analysis. Whereas the amygdala is activated in fear, fear of giving a speech deactivates the amygdala. This is understandable if one thinks of the amygdala as mediating fear of bodily harm, not fear of any unpleasant outcome, such as embarrassment in this case (which might activate the orbitofrontal cortex). Many researchers relied on Ekman and Friesen's (Reference Ekman and Friesen1971) list of six emotions with universal facial expressions. Other emotions have no distinct facial expression, as Ekman (Reference Ekman, Ekman and Davidson1994a) acknowledged, and some facial expressions – happiness and sadness – can be observed following any pleasant or unpleasant experience. More specific localization might result from using more precise emotion terms.