Basic functioning

Panksepp has proposed that the SEEKING System is the primary task-oriented pathway by which affective goals are met (Panksepp, 1998). This generalized system activated by primary biological needs and characterized by homeostatic detection mechanisms, can also function in a non-specific manner. Other primary emotional systems (in our proposal labeled E3–E9), each characterized by a genre of intention and intensity of desire, feed information to the SEEKING system, as do secondary emotions thereby affecting overall motivational state². Through inclusion of conscious volitional goals in E1, intentions and resulting purposive action gain emotive power: “I want that job,” “I need that house,” and so on, have affective components. This ultimately influences the way that ethical choices and values (the basis on which we choose acceptable actions) become effective in guiding action, as they too have underlying affective components.

Evidence now suggests (e.g., Berridge and Kringelbach, 2008; Berridge et al., 2009) that reward functions are parsed into two components – a motivational, appetitive system (corresponding to Panksepp’s SEEKING/expectancy system), and a distinct hedonic appraisal or consummatory system. This “wanting” vs “liking” distinction is supported by the fact that addiction involves craving but not necessarily satisfaction (Robinson and Berridge, 2003). The relevant neural pathways normally function together, but can be behaviorally dissociated (e.g., Knutson et al., 2001; Cannon and Bseikri, 2004), even in fish (Spector, 2000) suggesting the division is of ancient origins.

The earliest protovertebrate ancestors would of necessity had consummatory appraisal responses even before appetitive seeking behaviors evolved, hence “liking” might be considered the primordial reward system. In mammals the lowest level for hedonic appraisal must reside in the brainstem as decerebrate rats and anencephalic infants also show hedonic responses (Steiner et al., 2001). Although taste, which relies on medullary nuclei, was probably the earliest effective stimulus, the ventral pallidum, nucleus accumbens, and several other forebrain structures evolved to process and respond to pleasurable stimuli in many modalities. Thus in our model the multi-modal nature of hedonic appraisal parallels that of the general purpose SEEKING system described by Panksepp (see, e.g., Kelley and Berridge, 2002; Burgdorf and Panksepp, 2006).

The mesolimbic dopamine system, extending from midbrain ventral tegmental area (VTA), lateral hypothalamus, nucleus accumbens shell to orbitofrontal cortex, was traditionally implicated in the neural basis of reward; however it is now recognized that both “wanting” and “liking” mechanisms also utilize endogenous opioids (Levine and Billington, 2004; Peciña, 2008) and the ventral pallidum is a necessary component of the pathway (Smith et al., 2009). These systems play a fundamental role in learning, possibly by associating arousal with specific activities, thereby attaching a positive affective value to them and then acting as a stimulus for repetition of these activities (see, e.g., Wise, 2004; Rolls, 2005).

Following Panksepp’s (1998) suggestion that the SEEKING system evolved to provide a common currency of reward, we suggest that facilitated by associated hedonic appraisals there exists a primitive system shaping general coordination of affective responses. The dopamine system is activated not only by food, drugs, sex, electrical stimulation, and monetary reward, but also by aversive stimuli with response segregation according to positive or negative valence occurring in separate regions of the nucleus accumbens and pallidum (Kelley and Berridge, 2002; Smith et al., 2009). The ventral pallidum also responds to a wide variety of food, sexual, and affiliative cues (e.g., Rauch et al., 1999; Dillon et al., 2009; reviewed in Smith et al., 2009).

Psychological illnesses that Panksepp associates with malfunctioning of the SEEKING system include addictions and cravings, eating disorders, and possibly schizophrenia (see Panksepp, 1998; Panksepp and Harro, 2004). Panksepp (2002) considers obsessive compulsive behaviors to be malfunctions of this system, however, we suggest below that washing symptoms may arise instead from malfunctions of a primordial DISGUST system, while symptoms of checking and hoarding may arise from circuitry originating phylogenetically in a POWER/dominance system.

Basic survival

In previous papers (Toronchuk and Ellis, 2007a,b) we described a primitive emotional system which, analogous to SEEKING, can be activated by various sensory modalities and ideational components, but with opposite functions to SEEKING. We designate this as the DISGUST system rather than AVOIDANCE or AVERSION because of the previous inclusion of disgust in various lists of basic emotions from Darwin onward. Together SEEKING and DISGUST would have been the primal opposing operating systems for ancestral organisms. In our model DISGUST evolved from primitive chemosensory mechanisms adapted to both avoid pathogens and their toxins and eject them from the gut if necessary. We propose this occurred in conjunction with the development of interaction between the immune and nervous systems (c.f. discussion in Ellis and Toronchuk, 2005; Rubio-Godoy et al., 2007) because non-specific (innate) immune cells are found in almost all multicelled organisms. Evolutionary adaptation would then lead to learned avoidance of toxic or infected material before ingestion, as opposed to vomiting afterward, ultimately giving rise to the subjective human experience of disgust as an anticipatory mechanism for avoidance.

The DISGUST system meets the criteria we enumerate above for primary emotional operating systems (Toronchuk and Ellis, 2007a,b, but see Panksepp, 2007). This system is intended as a parallel mechanism to appetitive SEEKING. Much broader than the sensory affect distaste, disgust is elicited by olfactory, gustatory, auditory, tactile, or visual cues (see Curtis and Biran, 2001; Curtis et al., 2004). Therefore it is not restricted to avoiding bad taste, but hinges on avoidance of contamination (Rozin and Fallon, 1987; Haidt et al., 1997; Rozin et al., 2009). Our proposal is that nutritional-, sexual-, and socially-related stimuli plus ideational components are all able to activate either the SEEKING or DISGUST systems in analogous ways.

We suggest that the disgust response did not arise merely as a reaction to bad taste, but due to association with increased likelihood of illness as proposed above (Curtis et al., 2004). Serotonin release is essential for development of disgust reactions and its use by the gut, vagus nerve, brainstem disgust mechanisms, and immune signaling suggests a defensive continuum of immune and disgust systems linked by serotonin (Rubio-Godoy et al., 2007). Negative hedonic value is not necessary for disgust as shown by conditioned taste aversions (CTA) elicited to sweet tastes paired with illness (e.g., Garcia et al., 1974; Parker et al., 2008). Conditioned immune responses may form from single pairings of novel taste with antigens (Pacheco-Lopez et al., 2004) further suggesting to us an evolutionary role for the immune system in disgust. Formation of a CTA requires activation of the insula (Ferreira et al., 2006) a structure which compares incoming and stored tastes (Koh and Bernstein, 2005). Rats with insular lesions fail to learn anticipatory discrimination although they remain capable of hedonic responses (Kesner and Gilbert, 2007).

Human anterior insula (AI), containing the gustatory cortex, also plays a role in self-awareness (reviewed in Toronchuk and Ellis, 2007a). It is activated during experience, observation, and imagination of disgust (Jabbi et al., 2008). Insular responses to aversive tastes vary according to expectations (Nitschke et al., 2006) and disgust sensitivity (Calder et al., 2007) as predicted for a primary emotional system. Although many behaviors integrating higher cognitive processing with “gut level” feelings also activate it (see Saper, 2002), AI does not provide a simple module for disgust but functions together with other areas including the basal ganglia to instantiate a variety of interoceptive bodily and conscious feeling states (Calder et al., 2001). The adjacent orbitofrontal cortex also plays a major role in taste, smell, and visceral responses and in assigning reward values to stimuli (Rolls, 2005; Rolls and Grabenhorst, 2008). Recent brain imaging research consistent with our proposal for separate DISGUST and SEEKING systems, suggests two regions of dorsolateral prefrontal cortex with different patterns of connectivity are differentially involved in approach and avoidance as parts of a network including orbitofrontal cortex, cingulate, amygdala, and basal ganglia (Spielberg et al., 2012).

Increased cortical integration of body states occurs in primates because the insula receives direct thalamocortical taste and visceral input which in rodents arrives indirectly via the amygdala (Craig, 2005). Emotional contagion or “resonance” is a further adaptation promoting disgust activation when observing socially relevant disgust responses in conspecifics (Wicker et al., 2003; von dem Hagen et al., 2009). The insula may also play a role in social evaluation cognition as a viscero-motor center which simulates the activity of others in a manner similar to “mirror” neurons previously described in monkeys (Gallese et al., 2004; Keysers and Gazzola, 2007). Cognitive processing allows primitive DISGUST impulses to be melded with social learning to produce secondary emotions incorporating social status and morality. Activation patterns of facial muscles in response to unpleasant tastes, contaminated objects, and unfair treatment are consistent with the suggestion of shared neural evaluative mechanisms for distaste, olfactory disgust and moral disgust (Chapman et al., 2009). The role of the insula in awareness of disgust in self and others, possibly facilitated by von Economo neurons in AI and adjacent orbitofrontal area, may be a preadaptation setting in motion the development of theory of mind and moral reasoning. We are thus suggesting that the primitive emotive circuit which originally functioned to defend the self by regulating consummatory behaviors contributes to an emotional system which facilitates some of the most highly developed human capacities of social evaluation (Toronchuk and Ellis, 2007a,b).

Disgust sensitivity is inversely correlated with sensation seeking (Haidt et al., 1994) further suggesting that DISGUST and SEEKING represent opposing operating systems which might then be reflected in contrasting behavioral traits and psychiatric disorders. Genetic influence in disgust sensitivity (Olatunji and Broman-Fulks, 2007; Kang et al., 2010) is consistent with the finding that even presymptomatic genetic carriers of Huntington’s disease show selective deficits in recognition of disgust (Hennenlotter et al., 2004) and insular size in presymptomatic patients is correlated with disgust recognition (Kipps et al., 2007). In contrast increased activation of the insula in OCD patients is associated with contamination, specifically washing-related symptoms (Stein et al., 2001; Shapira et al., 2003). Phillips and Mataix-Cols (2004) find that patterns of brain activation in OCD patients in response to disgust or anxiety-producing objects vary according to the patient’s major symptom type. Brain activation to disgust stimuli occurs independently of anxiety again supporting the view that OCD typified by contamination/washing symptoms represents an underlying malfunction of the DISGUST system (Husted et al., 2006; Lawrence et al., 2006).

In contrast to internal threat, protection from external sources was phylogenetically provided by the RAGE E3 and FEAR Systems E4 described in detail by Panksepp (1998). Previous experience and assessment of the present circumstances determines the final behavioral outcome of these systems in higher organisms. As with DISGUST certain stimuli are predisposed to easily activate FEAR (Öhman and Mineka, 2001) and as with DISGUST the same structures activated during the production of fear and anger are also activated in humans during recognition of fear and anger in others (Calder et al., 2001; see discussion in Goldman and Sripada, 2005). Panksepp suggests that certain psychological disorders such as aggression, psychopathic tendencies, and related personality disorders might be correlated with the malfunctioning of the RAGE system; while anxiety disorders, phobias, and PTSD variants might be associated with malfunctioning in the FEAR system (Panksepp, 2002).

Reproduction

Sexual reproduction, an obvious necessity for evolutionary selection, is the outcome of the LUST system E5 which also infers “wanting” and “liking” components. As with the SEEKING system the appetitive and consummatory aspects can function independently with corresponding behavioral dissociation (Pfaus, 1996; Kippin et al., 2004). The transition from reptiles to mammals brought about modifications in this ancient system to produce the attachment necessary for lactation (Panksepp, 1998). Thus the circuitry for LUST is biochemically related to that of both PANIC/attachment (or separation) and CARE. Due to different mating strategies, the complex relationship between LUST, PANIC, and CARE likely differs in males and females giving rise to different attachment styles in adult humans (see Taylor, 2006; Del Giudice, 2009).

All the social emotional systems are mediated by hormones, synaptic signaling, and other biochemical signals as employed by other elements of the underlying primordial value systems. Social emotions can be therefore causally effective in terms of contributing to neural selection because these signals also affect brain plasticity. Oxytocin and vasopressin in particular function as neuromodulators in the value systems associated with both adult social bonds and parental behavior (reviewed in Carter, 1998, 2003). For example during mating vasopressin is released in the ventral pallidum and nucleus accumbens (areas associated with SEEKING) of male prairie voles; its blockade prevents pair bond formation (Lim and Young, 2004). In addition, vasopressin plays a role in sexual competition (see e.g., Sewards and Sewards, 2003a).

In parallel with the SEEKING system, dopamine is associated more with the appetitive phase of LUST and endogenous opioids more with the consummatory phase. Dopamine secretion increases with sexual arousal, but at consummation dopamine decreases while secretion of oxytocin and opioids increase (Pfaus, 1996; Van Ree et al., 2000). The neural circuitry of this system includes among other areas prefrontal, insular, and anterior cingulate cortex along with amygdalar, hypothalamic, and other subcortical areas (Stoléru et al., 2012). These cortical areas are highly interconnected with the amygdala and are also prominent in social bonding. The malfunctioning of this system might be reflected in fetishes and sexual addictions (Panksepp, 2002), and disorders of desire and orgasm resulting from either over-activation or under-activation of the subcomponents.

Primary emotional systems of social bonding

CARE, PANIC, PLAY are also important emotions of social behavior discussed in detail by Panksepp (1998, 2002) Here we discuss primarily their crucial roles in sculpting mammalian social and cognitive development. According to MacLean (1990, p. 247) the differentiation of mammals from reptiles involved (1) lactation and associated maternal care, (2) vocal communication to maintain mother-infant contact, and (3) playful behavior facilitating social learning. Because lactation and maternal care are essential for mammalian survival, significant selection pressure would act on the neural mechanisms controlling these behaviors. In mammals social bonding is initially effected primarily by the PANIC/attachment (E6) system in the young, which triggers emotional panic during separation, but signals contentment during closeness. This necessitated the tandem development of a complementary CARE (E7) system, through which parents respond to the young. Panksepp’s (1998) PANIC system basically concerns separation distress; however, we suggest that distress is only one mode because the system also operates in a positive fashion when needs are satisfied. We propose that subjective feelings in the human infant include both panic/distress during separation, and also contentment/comfort during contact (E6); and in the care-giver, tenderness/affection, carrying over to reciprocal distress when the infant is perceived as in distress (E7).

The biological origins of human sadness may be rooted in an extended system used earlier in phylogeny for pain perception but which now includes the cingulate gyrus that mediates separation distress in infant mammals (Panksepp, 1998, 2003b). This origin is supported by the role of endogenous opioids in both pain reduction and positive feelings brought on by contact. Similar to the case of distaste and DISGUST, an early sensory affect (pain) perhaps gave rise to a basic emotional system (PANIC/attachment). Relational loss is thus perceived in humans as similar to pain and panic. Both adult and infant bonding are mediated by oxytocin (Carter, 1998, 2003; Insell and Young, 2001; Lim and Young, 2004; Taylor, 2006; Bos et al., 2012) perhaps also due to phylogenetic origin. Estrogen further enhances the effects of oxytocin, which may provide a basis for gender differences in human attachment styles (see Taylor, 2006; Del Giudice, 2009).

The long period of helplessness in hominid infants combined with a need for training in foraging and social behaviors likely put considerable selection pressure on the neural circuitry for emotional attachment between mother and infant. Ape mothers’ use of gestures and facial expressions plays a key role in communication with infants, who correspondingly develop an intense interest in the mother’s face Falk, 2004). Chimp mothers teach which foods are edible, and perhaps even tool use (Goodall, 1986; Falk, 2004). Chimp infants communicate various types of distress through specific vocalizations. Hominid evolution therefore involved tandem evolution of emotional circuitry in adults, in order to provide emotional nurturance and instruction, and parallel circuitry in the young to seek and respond to caregivers. This entailed increasing use of gesture, facial expression, tactile, and vocal communication. According to this scenario, the lengthening period of dependency in early humans due to increased brain size and subsequent immaturity at birth selected for even more skilled caretaking and teaching ability on the part of adults. Although the effects of early emotion on individual adult emotional and social behavior has been well-studied, the evolutionary role of infant emotion as a selection factor in cognitive development has been less well researched. According to the theory we suggest, mother-infant communication provided emotional motivation for the development of language. Learning in infants is critically enabled by reciprocal interaction with the care-giver (e.g., Schore, 1994). The ability for shared attention becomes necessary for the development of a theory of other minds and language use. Thus social emotions provide the valenced state necessary for infant learning, initially developed to predict and respond to actions and emotions of the care-giver. This provides a major example of how cognition is molded in part by emotion, and emotion in turn comes to be regulated by cognition.

Tactile stimulation in infancy mediates upregulation of glucocorticoid receptors in the hippocampus through DNA methylation, an effect which in rodents persists throughout life (Kaufman and Meaney, 2007). Upregulation decreases reactivity of the hypothalamic-pituitary-adrenal (HPA) axis via negative feedback. Early tactile stimulation of female rodents regulates expression of estrogen receptors in the medial preoptic area, resulting in adulthood in increased oxytocin receptor binding with subsequent increased licking, grooming, and nursing. Nurturant maternal care also enhances learning by enhancing NMDA receptor activity in the hippocampus. These mechanisms may all be conserved from rodents to humans (Kaufman and Meaney, 2007). Consistent with this, human adults who experienced childhood adversity both rate reward cues less positively and show less response in the reward-evaluating pallidum (Dillon et al., 2009). Together these effects illustrate the long-term influence of the PANIC and CARE circuitry on emotional and cognitive behaviors.

It seems worth investigating whether gender differences in stress responses (e.g., Dalla et al., 2008) might be related to demethylation of estrogen receptors in early life as described by Kauffman and Meaney (or other nurturant behaviors), thus affect attachment style. Specifically human males show more avoidant attachment, and females more ambivalent attachment from middle childhood through adulthood (Del Giudice, 2009). Fight and flight behavior is more typical of males under stress, while oxytocin facilitates tend and befriend behavior in females (Taylor, 2006).

Separation of infant and caretaker can lead to physical and emotional stunting as described by Rene Spitz in the 1940s, a concept that was later expanded by Harlow while the long term impact of maternal care on children was confirmed by Bowlby (Van der Horst and van der Veer, 2008). Panksepp predicts malfunctioning in these social systems, combined with other environmental factors, are associated with panic attacks, pathological grief, some types of depression, agoraphobia, social phobias, dependency, attachment disorders, and perhaps also be a contributing factor to autism (Panksepp, 2002). Nemeroff et al. (2006) seem to suggest an association with post-traumatic stress disorder. Thus, the PANIC and CARE systems facilitated evolution of human social bonding, altruistic behavior, and possibly influenced language evolution, while disruption is predicted to lead to numerous psychological disorders (see Toronchuk and Ellis, 2009).

Learning and development

The interaction between emotion and cognition becomes especially apparent in the life-long processes involved in learning. Although learning is enabled by the primordial SEEKING system (see Berridge et al., 2009) it is ontologically dependent in mammals on the circuitry for PLAY E8. The importance of mammalian play as preparation for adult food procurement and social roles, suggests that play should be considered a basic emotional program necessary for normal human cognitive development. It was facilitated evolutionarily by cortical enlargement and the prolonged interaction necessitated by lactation (MacLean, 1990). Developmentally play is also influenced by the interaction between CARE and PANIC. Play involves learning social roles and behaviors (e.g., Bekoff and Byers, 1998; Brosnan, 2006; Keltner et al., 2006) and allows assessment of social commitments and dominance.

Extensive mother-infant play in chimps and especially bonobos is characterized by facial gestures, vocalizations, and laughter (reviewed in Falk, 2004). It is possible that optimal cortical plasticity depends on the rewarding effects provided by these interactions. Allowing juvenile rats 30min of “rough and tumble play” results in increased BDNF transcription in the amygdala and dorsolateral frontal cortex (Gordon et al., 2003). Although much anecdotal information suggests an important role in human learning, the underlying neurophysiological mechanisms are not yet defined.

Panksepp (1998) includes “rough and tumble play” in the list of primary emotions, but we speculate that in humans the neural activation has been extended to facilitate representational and intellectually imaginative play. Evidence exists that both captive and wild apes may engage in representational play in which one object comes to stand for another (Lyn et al., 2006). We conjecture that the phylogenetic transition from “rough and tumble play” to representational play facilitated the evolution of language through development of a mechanism allowing symbolic representation. A theory of other minds, another factor in imaginative play, is likely necessary for language evolution. Play thus remains particularly important in the ontogenetic processes of language development (see Bruner, 1983; Paley, 2004; Zigler et al., 2004).

Many aspects of human culture such as the performing arts, ceremonial, and celebratory behavior, and even creativity in science ultimately depend on extensions of these mechanisms (see Frost et al., 2001). Humor with embedded dominance paradigms in which we “make fun” of others could be another expression of rough and tumble activity. Social aspects such as turn taking and feinting may also contribute to development of altruistic behavior. A detailed assessment of the PLAY system, and the associated role of opioids, is given by Panksepp (1998, 2002). He suggests that attention-deficit hyperactivity disorder, mania, and perhaps impulse control disorders may be associated with its malfunctioning.

Implications

The nature of the phylogenetic influence on emotional behavior has consequences in many areas of human life, e.g., economics, politics, and education (Ellis, 2008). Given the assumptions of Affective Neural Selection, we have a more nuanced version of motivational theory than simple conditioning theory provides. Each major survival and reproductive need, we propose, is related to one or more of the specific primary emotional systems either directly, or indirectly through secondary emotions; hence we can analyze psychological and psychiatric issues in terms of their relation to these primary operating systems in order to develop new forms of both therapy and prevention.