accueil appel  thèmes  contributions lectures  commentaires recherche

 

AFFECTIVE NEUROSCIENCE :

POSSIBLE CONSILIENCE BETWEEN PSYCHOANALYSIS AND BRAIN RESEARCH. ^*

Jaak PANKSEPP

             Affective and cognitive neuroscientists are now in a position to link concrete neural entities to various abstract psychological and psychoanalytic concepts. Of course, we can anticipate that psychoanalytic theory and terminologies may need to be modified and fine-tuned when placed in the crucible of neuroscience. Likewise, functional neuroscience may become more refined by being ground with the pestle of psychoanalytic thought. Unfortunately, like the proverbial patient with a full array of defenses, practitioners of both may be reluctant to embrace change. The best that can be hoped for at present is that ongoing debates will promote positive interchange that leads to informative empirical inquiries. Clearly, psychoanalysis needs to anchor its thinking in a manner that will promote the empirical evaluation of ideas, and neuroscience needs to take emotional dynamics and defenses more seriously than it has in the past.
             The polarity between traditional psychoanalytic and neuroscientific views can be documented by a simple Medline search requesting papers from 1966 to the present (June 5, 1998) that concurrently discuss psychoanalytic (a total of 2,478 citations) and brain issues (a total of 240,567 citations). Such a search yielded only 40 papers, and it is noteworthy that none appeared in journals devoted primarily to neuroscience. Thus, psychoanalysis is beginning to seek a substantive connection to neuroscience (Schore, 1997; Solms, 1996) but there has been little reciprocity.
             Some of the causes for the lack of substantive interchange are straightforward. Neuroscientists are generally proud of possessing ears that are deaf to higher integrative issues, especially ones like «central states» that can only be indirectly measured. On the other hand, Freud intentionally chose to distance his thought from brain matters when he could not bring his alchemical «Project for a Scientific Psychology» to a satisfactory conclusion. When Freud was young and driven to relate mental and neural processes, available knowledge about the brain could sustain no credible fusion.
             Now that a century has passed since Freud abandoned his premature aspirations for a coherent psychobiology, the time is ripe, perhaps for the first time, to blend psychoanalytic and neural thought into a coherent synthesis (Bilder, 1998). To reemphasize the obvious--in order to make scientific progress, psychoanalysis needs to anchor many of its concepts in brain matters. It may also behoove neuroscience to begin seriously entertaining the existence of those subtle emotional processes, heretofore prodigal, that are absolutely essential for the construction of human and animal minds. Mainstream neuroscience should certainly become conversant with the fact that the mammalian brain can generate a variety of affective feelings which probably reflect long-term causes of behavior rather than epiphenomenal flotsam, as all too many behavioral neuroscientists are still prone to believe.
             Throughout the 20th century, psychoanalysis remained a bastion among the human sciences that acknowledged the deeper emotional currents of the human mind. In order to understand the brain, neuroscience must also become conversant with such realities. Affective processes are engraved deeply within the evolutionarily provided structures and dynamics of the mammalian brain. Psychoanalytic theory may now help guide neuroscientific thinking concerning various emotional dynamics that transpire within neural tissues. Conversely, neuroscience can provide the facts which may allow psychoanalytic theory to link up firmly with objective approaches, and thereby submit itself to the type of disciplined empirical inquiry that is the hallmark of all true sciences. In order for our concepts to have scientific substance, they must be capable of being quantified (whether it be by direct or indirect measures), and to specify systematic relationships to other variables than can also be quantified. How well such hybridization can work remains an open issue, but all who are interested in clarifying the affective mysteries of human existence will need to pay attention to how neuroscientific, psychoanalytic, psychobiological and sociobiological modes of thought can intersect and evolve (Leak & Christopher, 1982; Schore, 1997; Solms, 1996). As Freud sought to do through his depth psychology, we still need to deal credibly with the relatively invisible neurodynamic underbelly of the mind. Obviously, there are many subtle brain processes that help us maximize our fitness, and no single discipline has a comprehensive view of the relevant matters. The places where all relevant approaches can «conciliate» and work together (the domain I would call Affective Neuroscience) should yield some of the most fascinating scientific breakthroughs concerning brain organization in the next century.
             The preceding article provides fertile ground for planting hybrid seeds. In their contribution, Solms and Nersessian (henceforth, S & N) offer a synopsis of Freud's views on affective processes, providing a rich array of ideas and questions for generating testable associations between psychoanalytic views of emotions and emerging neuroscientific ones (e.g., Damasio, 1994; LeDoux, 1996; MacLean, 1990; Panksepp, 1998a). Their article may be considered an attempt to fulfill the hope of Freud's youthful 'Project.' Indeed, many of the spectacular new findings from modern neuroscience and psychobiology can be interpreted in ways that are generally coherent with many of Freud's ideas. However all proposed relationships must be deemed provisional until they are subjected to rigorous empirical evaluation. The critical question will remain: To what extent can objective breakthroughs and fruitful theories will emerge from the blending of approaches? At present, we can only hope that their intersections will be intellectually lively and scientifically productive.
             The journey ahead is unlikely to be smooth, not only because of the complexity of the underlying issues but also because so many biasses surrounding Freudian theory. Just consider the recent claim by by E.O. Wilson (1998) that «Freud's conception of the unconscious, by focusing attention on hidden irrational processes of the brain, was a fundamental contribution to culture. It became a wellspring of ideas flowing from psychology into the humanities. But it is mostly wrong. Freud's fatal error was his abiding reluctance to test his own theories--to stand them up against competing explanation--then revise them to accommodate controverting facts.»
             The failure to subject ideas to the acid-test of empirical evaluation remains the time-honored criticism of psychoanalytic views (Macmillan, 1997). However, in making such criticisms, we should also recognize that most of Freud's ideas, especially in the area of emotions, remain to be evaluated. From that perspective, it is nonsense for anyone to declare the ideas wrong--they simply have not been adequately tested. In fact, it is only with the advent of modern neuroscience, with its many brain scanning and psychopharmaceutical tools, that most of the ideas can be effectively evaluated. We can now estimate how unconscious processes operate within brain emotional systems (Morris, Ohman & Dolan, 1998). Without the ability to directly manipulate and measure neural activities, one could never provide causal evidence for the deep affective issues that Freud was most eager to clarify. For instance, the basic nature of «id-energies» (to coin a new catch-phrase) are incapable of being evaluated without concrete neuroscientific hypotheses. By linking Freud's ideas to neural systems (e.g., Schore, 1994), there are credible empirical ways to proceed.
             We should also remember that most of the criticism of Freudian perspectives is within the realm of therapeutic practice. Thus, we must clearly distinguish between Freud's theoretical ideas concerning the foundations of psychology and his advancement of a specific type of psychotherapy. Each set of ideas needs to taken on its own merits, and the failure of psychoanalysis to alleviate certain psychiatric symptoms (e.g., those associated with schizophrenic and autistic disorders) says little about the credibility of psychoanalytic theory in our attempt to understand human emotionality.
             S & N have now gathered together Freud's views on emotions from widely scattered sources, and I find much merit in those ideas. Still, it is a daunting task to juxtapose Freud's wide-ranging hypotheses, often metaphorically expressed, against the accumulating peppercorns of evidence from the basic and clinical neurosciences. Even though, neuroscience has little taste for incorporating such complex functional domains into its conceptual schemes, the time is ripe for doing so (Bilder, 1998; Schore, 1997). For example, a striking demonstration of how effectively psychological and neuropathological data can be joined to highlight the nature of dreams has recently been provided by Solms (1997). Comparable levels of neuropsychological insight are being achieved for affective processes (Damasio, 1994). In my commentary I will largely focus on how the basic neuroscience issues may promote such analyses, and I will seek potential conceptual connections where empirical ones do not yet exist.
             I will organize my commentary in parallel with the organizational structure of the target article--four major themes followed by a conclusion. To facilitate reading, I use sub-titles similar to those in the target article, modified to highlight the points of view I wish to share. Since some of the most relevant issues appear at several points in the target article, there will also be some recursions in my commentary, which may not be inappropriate when we discuss recursive, hierarchically self-organizing systems like the brain (Kelso, 1995), where no accepted scientific vocabulary yet exists for many of its «modular» functions. I will not provide extensive referencing for the basic neuroscience issues which are well covered in many other sources, including my recent effort in clarifying the neuroscience of emotions (Panksepp, 1998a).

I. A GENERAL ORIENTATION TO MODERN EMOTION THEORY: PREVAILING PREMISES AND BIASES.
             Before focussing on Freud's views, let me make a few general remarks about traditional theorizing in the psychological sciences. Modern emotion theories generally attempt to be safely catholic--covering all essential attributes of emotionality (with lots of inclusive information-flow block-diagrams), without attempting to highlight essential ingredients above others. Often the one factor that is missing, or deemed of secondary importance, is the internal feeling state of the emotions. While this attribute is commonly believed to be the essential ingredient by non-specialists (Figure 1), a position with which I agree, it is also the most elusive attribute when one tries to approach emotional matters objectively. However, I believe it can be probed credibly with neuroscientific approaches by 1) seeking essential brain processes that synchronize the visceral and somatic motor expressions of emotions, which 2) are also key ingredients in the cognitive and memorial aspects of emotions, and 3) combining these findings experimentally with a study of the subjective reports of humans.

(Figure 1)

             My personal assumptions, harmonious with Freud's, are as follows: I believe that biological values and the affective neural processes via which they are instantiated penetrate all of the cognitive structures of the mammalian brain/mind. Although certain sensory/perceptual processes may be free of affect, all of the higher association areas of the brain, at least as they operate in the natural world, are permeated by the sustaining and guiding effects of biological values. This is not to say that the conditioning of emotional responses cannot proceed without any mediation by consciously perceived affect; they certainly can (LeDoux, 1996; Morris, et al., 1998). However, most of the emotional behavior we see in the world is probably modulated by the background effects of low-level emotions (i.e., moods), and it is within these longer-term influences that affective experience may be critical. Thus, I believe that in early childhood there is no sustained line of thought without a sustained line of affect, and if we really want to understand how people and animals behave in the long-run, we must understand their emotional feelings. It is only in adulthood, when long-term behavior patterns and habits of thinking and defenses have been established, that the obvious linkages between affect and behavior diminish. In fact the affective regulators may have simply descended to preconscious levels of neural processing, still exerting fundamental controls over mind and behavior while leaving the propositional attitude functions (i.e., appraisals) that arise from our higher cognitive abilities quite confused about the types of psychological functions that actually control our behaviors (Griffith, 1997).
             I personally interpret the available evidence, culled from divergent empirical studies (for a thorough review, see Panksepp, 1998a) as being consistent with the conclusion that a variety of executive systems for distinct emotional processes do exist within the mammalian brain. Of course, the systems are not simple and modular, but widely ramifying, interacting with many other specific and non-specific processes of the brain. However, contrary to much of current theorizing, I do not think all of these interactive processes are of equal importance in our quest to understand emotions. To put it in Freudian terms, I believe the key issue in emotion research is the characterization of the instinctual «id-energies» or in my preferred terminology, affect generating «emotional command systems» that instigate and orchestrate the diverse aspects of emotionality within the brain.
             Freud conceptualized such id-processes not only as the foundation of emotions and personality, but also as essential substrates for the maturation of ego functions. Unfortunately, Freud had no credible way to scientifically taxonomize the various id-functions of the brain, and they remained the most ambiguous of the many psychological processes that he discussed. Affective neuroscience has now provided abundant evidence for the types of basic id-functions that are contained in the mammalian brain--ranging from basic SEEKING systems to those that elaborate PLAY (Panksepp, 1998a). If it were willing, cognitive neuroscience could now highlight the neural nature of some of the ego defenses that emerge when unbridled id-functions come under higher neural controls, partially via associative learning and partially via the evolution of more subtle emotion-regulating neuropsychological processes such as the various defenses conceptualized by Freud. Evolutionary psychology is just barely beginning to conceptualize how powerfully such sociobiological strategies reflect the inclusive-fitness pressures that permeate our higher cerebral rationalizations, but it has yet to deal effectively with the basic biological nature of affective feelings.
             By trying to cover all aspects, while disregarding the most troublesome and essential attributes of emotions (i.e., the nature of subjective states), modern psychological theories of emotions often do not provide the necessary analytic simplifications that can yield robust empirical evaluations of key issues. Although Freud's view was also quite broad, encompassing most of the now well-accepted attributes of emotions, he did take a harder stance on the primacy of emotional feelings. I respect the fact that he placed affect--the pleasure/unpleasure principle--at the center of his scheme. This can still be deemed courageously controversial, but in my estimation, it is the correct point of view. Without such functions, we can not even make computers truly intelligent (Clark, 1997) and humans seem to lose a lot of their common sense when emotional systems are damaged (Damasio, 1994; Picard, 1997). Without emotions we would probably be little more than the zombies of philosophical renown (Harnad, 1994).
             Of course, id functions need to be sub-divided and taxonomized more extensively than Freud was able to do. Such deep functional issues must be clarified through conjoint psychological and brain research, and it should be more widely recognized that animal work is absolutely essential to resolve the fine details of the underlying systems with adequate clarity. Because of the lack of any credible methodology or data base, Freud had remarkably little to say about the deep nature of the id-based foundation for his system of thought. He devoted his efforts to discussing the higher psychological processes--how ego and superego functions are molded through experience. My own work for the past quarter of a century has been based on the supposition that it will be through the neuroanatomical and neurochemical clarification of the various «id-energies» of the mammalian brain that our understanding of the fundamental nature of the basic forms of human emotional experience and behavioral control must be based. I work from the premise that the id-energies evolved long ago, and remain homologous in all of our close evolutionary relatives. This is not a view shared by many of my colleagues, especially when it comes to the importance of internal feeling, which cannot be directly observed, in the governance of animal behavior.
             Even though emotional systems, as all brain systems, have changed during speciation as a result of selective pressures, it is reasonable to assume that divergence has been most extreme among the more recent layers of brain evolution (i.e., the higher regulatory elements), while remaining most similar at the more ancient levels. Indeed, some of our higher processes, such as language, may often tend to homogenize a variety of distinct subcortical emotional entities into broader valence categories such as positive and negative affective arousal which can confuse basic brain issues as readily as clarify them. Those concepts, like «weather» and «motivation,» are little more than class-identifiers for a host of distinct sub-processes that need to be empirically specified. For some reason, these broad categories of approach and avoidance, positive and negative affect (reminiscent of Freud's global pleasure-unpleasure dimension) are accepted more readily than narrower categories, perhaps because they appear to simplify matters, when, in fact, they may be obfuscating our pursuit of the basic systems that actually exist in the brain. Most human have little difficulty distinguishing more than a dozen basic emotional feelings, and I believe most of them could be clarified by conjoint human and animal brain research (Panksepp, 1998a).
             Of course, the study of feeling states remain among the most troublesome issues in animal research. Animals can provide no verbal reports of their internal feelings, and only general measures of approach and avoidance are presently reasonably well accepted as indicators of valence. The use of more indirect indices, namely the quantification of various correlated spontaneous behavioral and autonomic changes is not generally accepted by the neuroscientific community as compelling evidence for the estimation of internal states. However, to differentiate the distinct processes that fall under the broad categories of approach and avoidance, we presently have few alternatives but to assume that various animal behaviors (perhaps a premier measure being their affective vocalizations and other straightforward instinctual responses) may be used to index the presence of various affective processes in their brains (Panksepp, 1998a). Obviously, in using such measures to base our theoretical inferences, every interpretative challenge needs to be experimentally contrasted and evaluated. Fortunately, triangulation between mammalian brain research, the study of animal behavior and the systematic analysis of human subjective experience is a sufficiently robust strategy to sustain a credible level of scientific progress (Panksepp, 1998a). Once this level of analysis has been satisfactorily consummated, we may be in a position to monitor affective processes more directly from brain activities (Panksepp, 1999).
             Accordingly, the issue of affective experience in other animals is gradually becoming a scientifically workable problem. Now that we are recognizing the deep homologies that exist in the 1) anatomical structure, 2) neurophysiological dynamics, and 3) neurochemical coding of brain processes across all mammalian species, we have the opportunity to validate many of the findings and conclusions derived from animal behavioral brain research with subjective reports derived from human research. Although, informative correspondences can be achieved at all those levels of analysis, the most useful ones are bound to emerge from a study of the neurochemical coding of behavior. Young Dr. Freud realized this when he discussed the pleasure and unpleasure of «sexual release» in his «Project,» and speculated that «a suspicion forces itself on us that in both instances the endogenous stimuli consist of chemical products, of which there may be a considerable number.» (Freud, 1895/1981, p. 321). Also, as S & N note in their 13th footnote, the elderly Freud still held out the hope that: «The future may teach us to exercise a direct influence, by means of particular chemical substances, on the amounts of energy and their distribution in the mental apparatus.» And he followed that with a statement which remains as true now as it was then--»It may be that there are still undreamt-of possibilities of therapy»-- especially since the day is approaching when we will be able to pharmacologically modulate specific neuropeptide-based emotional systems of the brain. Even with the currently splendid, but still quite limited, stockpile of accepted pharmacological tools for human research--from amphetamines to Viagra, so to speak--we can initiate many experimental inquiries that may highlight the causal underpinnings of id-energies in humans. Credible predictions can be derived from the many studies that have already been done in other species.
             Although considerably less useful for deriving causal connections, neurophysiological and neuroanatomical correspondences also help highlight affective homologies within human and animal brains. Localized brain stimulation has yielded remarkable correspondences (Panksepp, 1985), and more recently, non-invasive stimulation of the cortical surfaces with rTMS (George, Kettner, Kimbrell, Steedman, & Post, 1996) is providing a powerful ways to analyze the role of higher cerebral domains. Neuroanatomical correlates of emotional processing can now be achieved with various modern modes of brain imaging (George, et al., 1996; Lane, Reiman, Bradley, Lang, Ahern, Davidson, & Schwartz, 1997).
Unfortunately, many of these correlative techniques, which detect changes of activity in cerebral canopies more readily than subcortical trunk lines, tend to promote and reinforce corticocentric views of emotional systems. This has led to a neglect, during the current era, of the many subcortical systems that are absolutely essential for the basic emotions to be generated within the brain. In fact, it remains possible that cortical inhibition of lower processes actually suppresses the conscious experience of affect in humans, leading to the internalization or «binding» of emotional energies as Freud discussed extensively. In any event, most of Freud's theorizing was also devoted to the higher cognitive levels where the basic instinctual energies are dealt with various emotion-regulatory mechanisms including repression, projection, reaction formation, and so on (issues that are almost impossible to address in animal research except to the extent that different memory processes can be shown to govern emotional output). My personal view is that the shared subcortical heritage, from which the various id-energies emerge across mammalian species, provides an essential and solid foundation for understanding the nature of affective processes as well as higher emotion regulating functions of the brain. Unfortuantely, these centra-state controls can only be monitored, at present, with indirect behavioral measure.
             Although Freud did not take any clear positions on the brain organization of emotionality, I find it intriguing how closely so many of his ideas, as outlined by S & N, mesh with my own. Most importantly, in putting affect at the heart of his analysis, he recognized that the assignment of value to behavioral and higher psychological processes was the key function of emotions. Freud did not equivocate on the issue that still plagues most neuroscientists interested in the problem of emotions. Freud recognized that affect registers the importance of salient world events, and thereby, permeates the higher conscious functions of the brain/mind. He also viewed affect as arising from fundamental biological mechanisms (presumably brain circuits), which guide instinctual action tendencies. The affect programs of the brain that have now been revealed are probably the immediate infrastructures of such processes (Panksepp, 1998a). At the heart of these systems there are a variety of chemical codes (largely neuropeptidergic) that may eventually permit powerful new modes of psychiatric intervention, and new ways to evaluate how feelings are constructed in the human brain. These findings can now enrich psychoanalytic thought.

II. AFFECT IS AN INTERNALLY GENERATED NEURODYNAMIC PROCESS, PROBABLY CLOSELY RELATED TO SUBCORTICAL EMOTIONAL CIRCUITS.
             Freud's insights on the nature of affect were prescient. The notion that various types of affect are intrinsic value-coding functions of the nervous systems is, in my estimation, a view that needs to be incorporated into modern neuroscience, both computational and biologic, as well as into the emerging field of consciousness studies. So far, with a few exceptions, this essential dimension is lacking in both fields. Freud distinguished between the qualia emerging from exteroceptive modalities, and the affects which emerge much more from intrinsic internal functions of the nervous system. I prefer to call these affective brain functions, emotional or evolutionary qualia (equalia, for short). It seems likely that the neural process «that affect is a perception of» has to be fundamentally unconscious (during its early evolutionary stages), and that it became preconscious and then conscious as certain types of additional neural systems evolved. Unless one subscribes to the minority view that some intrinsic aspect of matter generates consciousness (e.g., emergent processes arising from the «quantum sea»), we must accept that a great deal of the neural infrastructure for emotionality operates automatically, with few conscious causes, but always, in mammals at least, with the potential for abundant effects on consciousness. Affective states, arising from a variety of emotional and motivational processes, may constitute the «ground» in the figure-ground relationships that constitute ordinary conscious experiences. Even though our conscious mind is not prepared to focus on the ground processes as readily as on the figurative contents of mind (indeed it has great difficulty remembering emotional intensity, even though it easily remembers the events associated with emotional episodes), those affective states may be absolutely essential for any type of consciousness to have emerged in brain evolution.
             In the mammalian brain, all higher forms of consciousness may still be grounded on the most primitive forms of consciousness, which I assume was affective in nature. As already mentioned, without the ground of affective experience, I suspect that individuals would present themselves as the proverbial «zombies» of philosophical discourse. This may also be the main reason it has been so difficult to build computers that behave as intelligent organisms (Clark, 1997; Picard, 1997). If this is so, rational thought may only be a relatively fragile tip on the iceberg of affective experience. In line with this metaphor, we should also assume, as did Freud, that most of the primitive sources of mind lie submerged below conscious appearances. However, we should also keep open the possibility that there exist several distinct arena's of consciousness in the brain that normally communicate poorly with each other (and not just of the right and left hemispheric variety), and that when one is on-line, the others are not.
             The neural substrate that permitted the emergence of affective feelings in brain evolution may, in agreement with basic psychoanalytic theory, provide the primitive foundation for ego development. Whether these ego functions are quite high or low in the neuroaxis, or as is most likely widely distributed, may eventually be capable of being empirically evaluated. So far there have been regrettably few explicit proposals as to how basic ego functions within the brain should be conceptualized, probably because they remain to be generally accepted as neurodynamic realities worthy of analysis. However, we can now envision the ego to sprout from rather primitive areas of the brain where basic emotional systems interact with basic neural representations of the body. This is not to deny that in neural evolution, as in the epigenetic growth of each human brain/mind, ego maturation is not based on additional vast and luxurious neural and psychological growths within the brain (i.e., leading the ego to be epigenetically, widely distributed, blending eventually with moral processes) (for recent reviews, see Mascolo & Griffin, 1998).
             My own take on this subtle matter is that at a low level of the neuroaxis brain evolution established a neurosymbolic «virtual body» which represents the organism as a coherent entity. I assume that this neural process was initially the key actor, as opposed to the key observer, within the Cartesian Theater (Panksepp,1998b). Affective feelings may be instantiated in the brain by emotional operating systems interacting with the neurosymbolic representation of such a «virtual body.» To facilitate communication, I have chosen to designate such an entity as the SELF (a Simple Ego-Type Life Form), and this process may correspond to the most primitive aspect of Freud's ego-structure. I have suggested that this hypothetical entity is grounded within centromedial areas of the brain stem--areas such as the periaqueductal gray (PAG) and surrounding collicular and tegmental zones--but its influence broadcasts widely in the brain through many direct and indirect influences such the strong two-way connections with frontal executive areas of the brain and widespread influences on sensory cortices through the Extended Reticular and Thalamic Activating Systems (ERTAS) conceptualized by Baars & Newman (1994). I believe that various emotional systems create affective experiences by interacting with such an extended ego structure. The resting level of reverberation within this system is an epicenter (homeostatic settling point) upon which the various basic pleasures and unpleasures of life may be predicated.
             Because of the centrality of such a value coding system for everything an animal does, I disagree with the widespread assumption among behavioral neuroscientists that feelings may be causally inefficacious epiphenomena within the human brain. Even though unconscious processes are obviously much vaster in the brain than conscious ones, the denial of affective consciousness in animals is certainly a dubious and potentially nefarious point of view. It will not serve our neuroscience well, in the long-run, to make wrongheaded choices about such matters. Such points of view compromise the ability of neuroscience to establish better relations with the social-sciences and humanities, as well as the everyday society in which our work is embedded. In any event, I believe that the neural grounding of the virtual body of the SELF is such that it can inter-relate external stimuli (simple perceptions), internal values (emotional states) with a coherent and stable motor representation of the body (a basic action system). In this scheme, feelings may ultimately reflect the various types of action readiness that permeate the extended neurodynamics of the SELF. In other words, the distinct, primal feeling states may simply reflect the natural types of neurodynamics (resonances) that can exist within the animate core of our being.

(Figure 2)

             The anatomy of the underlying neural structures are congruent with the view that the centromedial brainstem contains essential ingredients for organismic coherence and action readiness (Figure 2). 1) The PAG contains a massive convergence of many distinct emotional systems which are strongly connected to spinal sensory and motor systems. 2) The overlying colliculi contain a massive convergence of most of the major external sensory systems. 3) Nestled between them there exist motor maps that can yield coherent bodily movements, especially of the orienting and simple rhythmic locomotor varieties. I assume that this primal SELF anchors organisms as a coherent, feeling creatures with a basic form of self-identity, but also, the widespread neural connections of the primitive SELF, especially directed toward frontal cortical areas, may provide various forms of affective arousal and coherence throughout the neuroaxis. Also, as indicated, it has strong connections to the ERTAS which has been postulated to be the staging ground for a generalized work-station of sensory consciousness (Baars, 1996; Newman, 1997), whereby perceptions can become imbued with affect. With such an extended neural entity, it is easy to imagine how emotional and motiational values could percolate throughout the neuroaxis, and how it may be the foundation for higher forms of consciousness, and its various satisfactions and discontents. All that is required is strong modulation of ascending ERTAS components (including cholinergic and catecholaminergic) from the PAG/SELF system, and such neural connections have been demonstrated (xx). The wide neural extents of these systems can permits abundant types of emotional expression--from passionate addictions and fixations to repressions and rationalizations--to be constructed from raw feelings. In this view, the essence of affective feelings arises from various basic emotional operating systems interacting with the extended SELF networks, while higher associative/memory processes extent and modify these special forms of action readiness with cognitive contents. Such a view is compatible, at least in general principle, with Freudian thought.
             Although affective feelings may reflect the neurodynamics of such a system in action, a great deal of detailed neuroscientific work needs to be done before we shall have definitive evidence for such assertions. In any case, we presently have abundant evidence for more distinct emotional systems in the brain than those that simply subserve positive and negative valence, or pleasure and unpleasure. As already mentioned, those broad distinctions (which probably reflect the way our higher brain systems categorize affective events) may hide a vast number of affective types that need to be specified and studied. At present, only animal brain research in conjunction with reports of human affective experience has the resolution to distinguish basic from derivative processes.
             Thus, while for Freud a paradigmatic example of affect was anxiety, which came in three basic variants--neurotic, moral and reality anxieties--it is highly unlikely that they all arise from distinct emotional command systems within the lower reaches of the brain. As Freud recognized, they emerge from various higher intrapsychic/intrabrain interactions. To a large extent they may emerge from the conflicts and frustrations of thwarted id-energies that led to the internalization of emotions and ego-development within the Freudian scheme of things .
             However, existing evidence does suggest that there are several distinct primary-process (or reality) forms of anxiety within the brain. One massive system runs from central amygdala to the PAG (Panksepp, 1990a; Panksepp & Bekkedal, 1997; Rosen & Schulkin, 1998). Another is a distinct separation-distress systems, coursing from cingulate and preoptic/ventral septal regions through the dorsomedial thalamus to the PAG, which governs bonding processes throughout youth, and is probably a key ingredient in establishing a secure base that permits the growth of psychological functions such as adaptive and maladaptive social attachments (Panksepp, Siviy & Normansell, 1985). The probability that early arousal of such an emotional system can promote future depressions seems likely (for review see, Harkness & Tucker, 1999; Panksepp, Yates, Ikemoto & Nelson, 1991).
             As we come to the other essential questions raised by S & N, neuroscience can only provide some best guesses. In the murky world of the preconscious, there is abundant room for alternative interpretations, and it is not possible to make any firm claims how neuroscience knowledge relates to Freudian inferences. We can produce various forms of approach and escape/avoidance using brain stimulation applied to very specific trans-diencephalic circuits that course bidirectionally between the PAG and a variety of higher limbic areas (especially the amygdala, cingulate, frontal and insular areas). However, once again, approach does not simply mean «pleasure» and avoidance does not simply mean «unpleasure.» There are many distinct bads and goods within the brain, as indicated by the distinct type of coordinated emotional behavior patterns that can be evoked. For instance, the most vigorous forms of self-stimulation obtained from the MFB corridor that arises from the VTA to the ventral striatum, septum and several frontal cortical areas, seems to generate a behavioral state of anticipatory eagerness rather than a simple and unitary sensation of positive affect (Depue, 1998; Panksepp,1981). I have called this a SEEKING system (Panksepp,1998b) since it is a general purpose neural mechanism that can serve a large variety of distinct motivational urges (drives?). Others have called such neural substrates the basic «wanting» system of the brain (Berridge, 1999), which is also a good label. Affectively, this system probably mediates feelings of an obsessively energized sense of desire and power rather than any simple pleasurable sensation that we normally experience when we fulfill our needs.
             On the other hand, stimulation of other areas such as the medial septal nuclei, can provoke pleasurable feelings (much more akin to those seen during consummatory activities), but these brain sites are off the beaten track from the major MFB highway where the most vigorous forms of self-stimulation are obtained. Likewise, there are many distinct forms of aversion in the brain, and the «unpleasure» principle needs to be discriminate various forms of fear, rage, separation-distress and pain. More work needs to be done, especially with electrical and chemical brain stimulation in conjunction with place preference and avoidance paradigms, before we can be confident exactly which brain areas mediate which affective qualities. In any event, I believe the basic emotional affects may be more closely linked to basic instinctual action readiness systems (Panksepp, 1998a) while the motivational affects are more closely linked to sensory systems (Berridge, 1999; Solms, 1996). But ultimately, both types of affect are mediated by intrinsic, unlearned sensory-motor integrative abilities of ancient, subcortical regions of the brain that can establish various types of neurodynamic feeling states within the brain.
             We can be certain that there is much more affective information encapsulated in the brain (via extended, «modular» circuits) than could have been anticipated when Freud was pursuing his synthesis. Simply consider the fact that there are ludic circuits in the mammalian brain which generate joyous social engagement (i.e., play and laughter). Freud barely anticipated such id/ego-functions in his Wit and Its Relations to the Unconscious. At present, the infrastructure of the PLAY system includes the reticular nuclei of the posterior thalamus and centromedial brain stem. Indeed, ludic urges may reflect one of the earliest coherent manifestations of ego functions that emerged from the neural evolution of the primitive SELF. This is suggested by the dramatically assertive way in which organisms spontaneously throw themselves into playful activities when such brain systems are active (Panksepp, 1993). How these systems control psychological and neural maturation should be an interesting chapter in brain research (Panksepp, 1998b)
             From my vantage, Freud did not adequately recognize the existence of emotional systems devoted to distinct social processes. Although he gave abundant attention to sexuality, perhaps rather too creatively, he failed to acknowledge the probable existence of basic instinctual systems for maternal devotion (tenderness), social-attachment (lovingeness), separation distress (sadness) and playfulness (joyfulness), all of which are heavily represented in mid-line thalamic and limbic cortices (frontal, anterior cingulate and insular areas) (Panksepp, 1998a). Of course, later generations of psychoanalytic thought developed some of these themes under the rubric of «object relations,» but much of this was discussed in the context of derivative rather than basic emotional processes. We can now hypothesize that it is through the auspices of the various basic social-emotional systems that such psychodynamics emerge. Indeed, the types of self-serving cognitive patterns and deceptive tendencies promoted by exigencies of reproductive success are vaster than anyone could have imagined in Freud's day (Zahavi & Zahavi, 1997). How higher social emotions (presumably superego functions), such as envy, guilt, jealousy and shame emerge from these systems will be a fascinating chapter of neuropsychology, that has received some attention from psychoanalytically oriented investigators (Lewis, 1988). The refinement of such higher social emotions may proceed through «prepared» forms of learning that emerge from the neurodynamics of the more basic emotional tendencies reverberating through evolutionarily prepared channels of the extended SELF.
             Likewise, social dominance and submission arise from learning processes attached to these basic systems, and a deeper understanding of the social dimensions may provide some important insights as to why psychoanalysis works well in the treatment of some «neurotic» disorders. One may view the classic form of psychoanalytic therapy to reflect, and perhaps operate though, the establishment of dominance-submission relationships. A consistent observation from animal behavior is that submissive animals commonly exhibit more social solicitation (resembling behavioral free-association) while dominant animals exhibit less friendly behaviors and typically only respond if actively solicited. This type of dominant/passive role is all too commonly assumed by psychoanalysts. Sometimes, therapy ends rapidly when this asymmetrical relationship, typically characterized by a lack of reciprocity, becomes psychologically intolerable to the client. On the other hand, therapy comes to a more satisfactory conclusion when the initially submissive client, working through negative emotional energies, emerges with a more well-individuated dominant and/or accepting attitude toward life.
             We can be certain that the brain areas that generate all of these affective responses are anatomically distinct from the thalamo-cortical systems that mediate the basic qualia arising from exteroceptive sensations. This is not to say these systems do not functionally interact at many places in the brain, allowing values to permeate perceptions and the external stimuli access to internal value systems to help establish more sophisticated learned behavior patterns. These interactions open the door to various phobias, defenses and consequent neurotic behaviors. The acquired linkages have been most thoroughly analyzed in the amygdala (LeDoux, 1996), but we can anticipate that there will be many brain areas where the learning of anticipatory defensive strategies can transpire for this as well as other basic emotional systems. For instance, a great deal of social learning and emotional regulation transpires within frontal cortical and anterior cingulate areas of the brain, especially for frustrations and social losses (Devinsky, Morrell, & Vogt, 1995; Drevets, Price, Simpson, Todd, Rich, Vannier, & Raichle, 1997), and it has been gratifying to see documentation that psychotherapies can ameliorate the over-arousal of these brain areas (Schwartz, Stoessel, Baxter, Martin & Phelps, 1996).
             It will be most interesting to conceptualize how the more subtle defenses may be constructed in the brain. They probably are not simply the straightforward associative processes that have been the stock in trade for behaviorists, but may include various forms of internally mediated implicit learning, including condensations, displacements, projections, and transferences. For instance, as Freud also explicitly emphasized, excessive early experiences with certain negative emotional states may promote different routes of self-organization within the hierarchical processes that control behavior (Mascolo and Griffin, 1998). There is little relevant research, except for the fact that certain experiences can sensitize emotional responses (Rosen & Schulkin, 1998) perhaps by promoting the ability of higher functions to trigger subcortical emotional systems, a process that has been conceptualized as an increase in «limbic permeability» (Adamec, 1991, 1993). From this perspective, the supposition that many psychiatric ailments may reflect imbalances within and among the activities of the basic emotional systems deserves more attention (Panksepp,1988). In any event, the aim of therapy should be to establish more harmonious emotional resonances within the primitive neural infrastructures, and it presently seems that psychopharmaceuticals are most efficacious in achieving such results. However, systematic evaluation of the long-term benefits of non-traditional approaches ranging from rTMS to music and other somatic therapies need to be considered by those who wish to entrain benefical brain rhythms and thereby indirectly change behavioral patterns by inducing positive mood states. It would be most interesting to determine how affective id-energies might be recruited for therapeutic change when free-association therapies are combined with such modalities. In general, psychoanalytic approaches need to develop stronger relations to all of the somatic approaches that are known to modify moods (Parkinson, Tottersdell, Briner & Reynolds, 1996; Thayer, 1989)
             A key empirical issue at the present time is the clarification of the manner in which the various affective states are represented within the brain. Most likely the answer is «widely,» but one of the least likely places is the neocortex. This is not to say that cortical processing is not massively effected by emotions. To document this, one need only analyze how heavily the cortex is aroused during emotional episodes using techniques ranging from EEG to cFos expression. If you simply allow animals to play or force them to swim for their lives, the whole cortical mantle is aroused as indicated by the widespread expression of cFos throughout the cerebral cortex (Campeau, Falls, Cullinan, Helmreich, Davis, & Watson, 1997; Kollack-Walker, Watson, & Akil, 1997; Burgdorf & Panksepp, 1998, Unpublished data). Unfortunately the meaning of such effects remains ambiguous, and they are not prognostic, in any simple way, of where the basic affective feelings are elaborated in the brain.
             Since we are presently in a corticocentric era of emotion research, most investigators probably still tend to favor frontal lobe or higher limbic (anterior cingulate, frontal, insular and amygdalar) localizations for the neural processing of affect, and there should be little doubt that these brain areas do help regulate affective processes. Indeed, we can also obtain various correlations between affect and higher cortical activities (Lane, et al., 1997; George, et al., 1996; Panksepp & Bekkedal, 1997; Paradiso, et al., 1997), but it should require a great deal more evidence to convince us that those correlates reflect the essential neurodynamics of the affective states themselves. They may be more influential in the complex regulations of emotions and the abundant emotion-cognition interactions. A few, like myself, favor the view that the extensively ramifying subcortical command-circuits that generate and synchronize various emotional behaviors and accompanying bodily changes are essential for generating affective feelings, and there is abundant evidence for such conclusions (Panksepp, 1998a). Let me just mention one of the most recent nuggets: Opiates produce more rewarding effects in areas such as the PAG and ventral tegmental area than in higher areas such as the amygdala and frontal cortex (David & Cazala, 1996; Olmstead & Franklin, 1997).
             Unfortunately, for most basic feelings, it is not yet clear whether the higher reaches (e.g., the amygdala) or the lower reaches (e.g., the PAG) are more important. Adolph, Tranel and Damasio (1998) close their most recent contribution to this exciting area of research (in this case, the recognition of emotions in the face) with the provocative statement: «An intriguing question that remains to be addressed is the amygdala's relative participation in triggering information that is innate, versus information that is acquired through individual experience in a cultural setting.» I assume that the innate information referred to would be some type of perceptually induced internal feeling state induced by certain intrinsic aspects of perception, and I personally believe the evidence will eventually show that the amygdala cannot elaborate the feeling of fear or trepidation without its descending connections to the hypothalamus and PAG. On the other hand, those lower areas can generate unconditional fearful feelings, albeit not certain learned fears, without the amygdala.
             According to the subcortical view, all of the basic emotional systems are modulated by a large number of inputs, ranging from simple sign stimuli entering emotional systems (e.g., the smell of feline predators to rats) to higher-level associative inputs (appraisals and conditioned stimuli) that have received much attention from neuroscientists. There will also be intrinsic physiological and hormonal parameters that modulate circuit sensitivities, and most importantly, there appears to be a substantial form of neuropeptidergic coding within many emotional systems. For instance, endogenous opioids and oxytocin are essential for regulating social emotions, CRF mediates a very basic from of anxiety (Figure 3), Substance P is important in instigating aggression, and there are many other neuropeptidergic relations that have been revealed by pre-clinical work (Panksepp, 1998a). As mentioned before, the neurodynamics of these systems may be modulated in many ways, providing various opportunities for establishing emotion-regulatory effects that remain to be deployed for promoting positive emotional health (Ryff and Singer, 1998).

(Figure 3)

             In sum, to my way of thinking, the interaction of emotional systems with lower substrates such as the PAG may be essential for creating feeling states which are then broadcast widely in the brain. If so, a key to understanding the nature of feelings is the identification of the various neurodynamic and neurochemical correlates that accompany the various types of feeling, and the evaluation of their causal roles through parallel lines of manipulative research. Such lines of inquiry have barely been initiated (for summary, see Panksepp, 1999).
             This should make the study of affective processes a primary goal for both psychology and neuroscience, but it is understandable why progress has been so slow on the most basic issues--we cannot readily visualize the spontaneous and long-term internal neurodynamics of the brain as readily as we can image how the brain responds to events in the world. Being human, we prefer to look where the light is brightest, so we often focus on rapid onset emotional responses to conditioned stimuli, where affect clearly can not be the primary mediator of the observed behavior (LeDoux, 1996).
             Historically, the hope was that emotions could be explained by focussing on the autonomically induced cognitive commotion that accompanies emotional arousal (e.g., leading to the the James-Lange, Schacter-Singer, cognitive-attributional traditions). Some also hoped that we could ignore the likelihood that there were very specific brain mechanisms that were evolutionarily designed to instigate and coordinate affective responses. The same bias is still evident in modern emotion research, both psychological and neuroscientific, with more effort devoted to peripheral autonomic and cognitive issues than the lowest common denominators in the brain. I suspect that a clarification of the core brain issues (i.e., the pathways and neurochemistries for the instinctual id-energies) will be essential for establishing a substantive foundation for psychoanalytic thought, even though the other levels of analysis may be more important for developing ideas of how emotions become «bound,» potentially yielding various defense mechanisms.
             Although a definitive answer to the question posed by S & N-- «what are affects a perception of»--can not yet be provided, relevant empirical work can now be conducted, even in the social realm (Panksepp, Nelson & Bekkedal, 1997). There are bound to be some surprises--for instance neuropeptides such as GLP-1 (Glucagon Like Peptide-1) and urocortin that are thought to control feeding behavior may do so only indirectly by modulating emotional processes (Panksepp & Bekkedal, 1997). While many levels of the neuroaxis surely provide important ingredients for fully resolved affective responses, the integrity of the higher functions is bound to be more dependent on the integrity of the lower functions than vise versa. Accordingly, I strongly urge investigators to devote more research to brain zones like centromedial areas of the mesencephalon and the reticular nuclei of the thalamus where emotional values and external events are first coordinated with a coherent map of the body and exteroceptively triggered perceptions. These may be critical brain zones where id and ego processes begin their massively entangled battle for primacy that reverberates through all subsequent levels of neural development of each individual and species. This is where the various «energies» of the anima emerge. With each layer of development, there are new opportunities for the emergence of defenses, displacement activities, and neural sensitizations that will be most difficult to disentangle.

III. BASIC AFFECTS MAY REFLECT DISTINCT NEURODYNAMIC RESONANCES OF THE PRIMAL SELF.
             While Freud had to generally consider that affect was controlled in some way by «the quantity of excitation that is present in the mind,» investigators have now identified a large array of specific and nonspecific neurochemical participants in each of the «affect programs» of the brain. There are some systems that are very generalized, participating in practically all emotional and cognitive responses (e.g., norepinephrine (NE), serotonin (5-HT), acetylcholine (ACh) and glutamate), while others, especially the neuropeptide neuromodulators, have more discrete effects limited to particular emotional/affective states (Figure 3). Thus when we analyze these systems, it is evident that ß-endorphin and oxytocin promote positive affective states, while CRF and CCK promote negative ones. I assume that the arousal of these systems generate distinctive affective responses, but relevant empirical work is scarce. It is hard to imagine how the role of such systems can be fully evaluated without careful human research in which qualitative mental contents are evaluated with closed-ended questionaires as well as open-ended techniques such as psychoanalysis provides. Obviously, each emotional and motivational feeling will have to be taken on its own terms, and how such findings may relate to earlier generalizations such as Freud's pleasure-unpleasure principle will remain open issues for some time to come.
             While it is now quite easy to imagine how the degree of affect could be related to the degree of synaptic activation in certain neurochemical systems, the objective measurements remain difficult. For instance, take the so-called PANIC system: We can generate different amounts and durations of separation calls by varying the degree to which we experimentally activate specific neurochemical systems. In experimental animals we can promote calls with CRF (see Figure 4) and glutamate, and we can systematically decrease calling, for a set amount of time, by promoting opioid, oxytocin or prolactin activities or by decreasing brain glutamate arousal (for summaries, see Panksepp,1998b). There are many examples for other affective systems--for instance, a female rat can be persuaded into a high state of sexual receptivity for approximately a third of a day by infusing minuscule amounts of LH-RH into the brain (Moss & Dudley, 1984), freezing can be induced for many hours following small alpha-MSH infusions (Panksepp & Abbott, 1991), and the list goes on. Unfortunately, none of these chemistries have been measured in vivo in organisms that are actually undergoing separation or sexual experiences.

(Figure 4)

             How we might conceptualize 'bound' and 'free' excitation is, at present, strictly a matter for speculation. One could talk about these matters in various ways--from focussing on pools of transmitters that are readily available for synaptic release to neural systems that are under tonic inhibition through explicit as well as implicit learned influences, but none of these possible connections can be asserted with assurance, except to note that one of the major functions of the neocortex (especially the frontal cortex), in its role of processing exteroceptive information, is to inhibit more impulsive subcortical processes.
             As suggested by Solms & Nersessian, a provocative way to make the distinction between the 'quantitative' and 'qualitative' aspects of affective life may be to focus on the generalized systems (e.g., NE , 5-HT and ACh) shared by all of the discrete functional systems of the brain which contribute substantially to a quantitative dimension of affect while the more specific neuromodulators, like many of the neuropeptides systems, are more influential in establishing the qualitative differences among affects. This seems highly promising. However, it is unlikely that neuroscientific investigators would be eager to relate the nonspecific components to a vague, hydraulic concept like «drive,» for they already have more specific descriptions of how these systems operate in the brain. For instance, NE controls how efficiently the cortex processes information by increasing the effects of incoming signals as compared to background noise, 5-HT tends to diminish the impact of information on the cortex, and ACh focuses attentional resources. All of these generalized functions interact with a host of specific brain systems for discrete types of information processing (both exteroceptive/cognitive and interoceptive/emotional), which necessitates that we study their various localized effects only under the most exacting experimental conditions.
             Periodically there has been a desire to relate these systems to specific affects, especially fear, but the data remain most consistent with the existence of broad and non-specific affective as well as cognitive information-processing effects. Although these systems act globally, if one restricts focus to limited areas of the brain, they may appear to have distinct qualitative effects on emotional memories within very specific brain circuits such as the FEAR circuits of the amygdala (McGaugh, Cahill, & Roozendaal, 1996). However, it may well be that normally such effects are not seen within the brain except when FEAR systems have already been aroused by other, more specific stimuli. For the time being we must remain cautious in reaching any definitive conclusions, since it is certainly possible that a great deal more qualitative specificity may emerge through the overall circuit neurodynamics, as well as the effects of the remarkably diverse types of receptor sub-populations that aminergic transmitters act upon.
             The possibility that a single chemical system can have both qualitative and quantitative consequences is also a reasonable conceptual alternative. This is especially evident for glutamatergic transmission in the brain. Glutamate appears to operate directly in practically every cognitive and affective processes that has been studied in animals. Along with some other excitatory amino acids, it appears to construct the skeletal form of every thought and emotion of which the brain is capable. This is evidenced by the large number of emotional responses--from anger to fear to separation-distress--that can be activated by glutamatergic stimulation of different brain areas. However, it would seem that most of these functional potentialities are not manifested until instigating stimuli arouse more specific emotion control systems such as the various neuropeptide circuits. The way I would envision this process is that once a neuropeptidergic command influence is aroused in the brain (both at synaptic sites of interaction as well as in more nonspecific broadcasting via paracrine transmissions), the selected glutamatergic response within a subset of available response elements may carry the functional message forward both in qualitative and quantitative terms.
             Meanwhile, its metabolically related cousin GABA, the most prolific inhibitory transmitter of the brain, exerts local inhibitory controls over these same neuropsychological and behavioral potentials. Thus, one can envision glutamate as controlling both quantitative and qualitative excitatory component of each emotional response, while GABA can do the same, by controlling the inhibitory components of such responses, both in functionally restricted circuits as well as broader brain networks that mediate supportive psychological processes. No doubt, tonic activity in all of these systems might also contribute to a general concept like «drive» for it is known that brains with low GABA are very excitable, tending toward epileptic activity, while facilitation of glutamatergic activity also promotes epileptic arousal. Conversely, if we markedly reduce glutamate activity and increase GABAergic activity (as can be done with sedatives ranging from alcohol to barbituarates), organisms become unconscious. In sum, a variety of generalized brain influences as well as the arousal of specific emotional systems contribute to overall affective integration in the brain.
             Freud did anticipate that the neurochemical control of affective states could eventually become a reality. The rise of biological psychiatry has amply confirmed this prediction, and thereby diminished the influence of psychoanalysis. But, we are now on the verge of a second great revolution of biological psychiatry--one that may allow us to harness specific neuropeptidergic emotional controls within the brain. Such a revolution may eventually help re-open the doors to certain psychoanalytic concepts. However, none of those finding may be able to breath life into Freud's generalized concept of «drive.» It is simply too broad and imprecise for any compelling neuroscience treatment.The hydraulic concept of drives building up that was so popular in Freud's days now appears to be misleading in so many ways, even though it is not difficult to imagine how such metaphoric entities may correspond to the intensity of excitatory inputs to central integrative systems such as the SELF. Unfortunately, there are too many distinct influences to subsume them under a single concept, except perhaps as a general class-identifier. Ultimately Freud's drive probably reflected the way higher concept-mediating areas of the brain seek to create order from the complexities of subcortical neurodynamics. It is reminiscent of the arousal axis in current dimensional theories of emotions (Lang, 1995), which may also simply be another broad, higher-order abstraction that has no unidimensional representation within the subcortical dynamics of emotional systems.
             Also, it is worth noting that the term has been used in too many ways in the history of psychology to be resurrected as a major explanatory concept in any system. The traditional utilization of the «drive» concept in psychology fell into disfavor when it was realized that it was intrinsically ambiguous and might be devoid of explanatory power (Bolles, 1975). In my own recent overview of emotions and motivations (Panksepp,1998b), the «drive» concept was relegated to those specific regulatory motivational functions such as hunger, thirst and thermoregulation, where specific interoreceptive detector elements have been identified in medial strata of the diencephalon. One could also employ the drive concept for transmitters such as dopamine (DA) that help regulate generalized appetitive arousal functions (which I have conceptualized as a SEEKING function), but it should be noted that such circuits do not directly up-regulate other energetic affective behaviors such as playfulness and fear. Thus it seems conceptually fuzzy to label such systems as general behavioral arousal or facilitation system as some have done. We should avoid talking about these systems as if they «energize» every behavior or provide an infrastructure for an omnibus «positive affect» system of the brain. Of course, seeking an optimal conceptual structure that does not do injustice to the underlying complexities, remains a most difficult issue in psychology and functional neuroscience.
             Obviously, it is scientifically counterproductive to resurrect or sustain vague general terms when more specific concepts will suffice. However, at the present time we also have be able to talk about coherent processes that are incompletely understood. A similar debate is raging in the use of certain neuroanatomical terminologies. Some are deriding the continued use of the «Limbic System» concept when we have more precise anatomical terminologies, while other's point out that in the absence of an international convention, comparable ambiguities exist for the more distinct structures like the amygdala. My own take on the matter is that fairly broad class identifiers are still needed for heuristic and didactic purposes even after their explanatory value has declined. In this same way, we will never have precise neural boundaries for psychodynamic concepts such as the id, ego and superego, and various subjectively experienced «psychic energies,» but we must provisionally continue to use some of them as class-identifiers for the types of global categories or processes that the brain most certainly contains. Some will have to be discarded when their communicative utility has become degraded. At present, the Freudian «drive» concept retains little value, and in my estimation, should be put to rest. Still, it seems evident that certain types of emotional and motivational urges to build-up in the nervous system, so let me at least dwell a bit further on some other possible neuroscience linkages, especially for those inclined to believe that the burial of the «drive» concept is premature.
             In doing so, we should recognize that Freud was using the drive concept slightly differently than I would prefer to use the concept (i.e., as various brain states that arise directly from bodily need detectors). For Freud, «drive» seemed to be the generalized tension or arousal that accompanied all of the various homeostatic imbalances. Thus, if one is seeking to shore up this Freudian concept with modern neuroscience data, one would certainly focus heavily on the types of general arousal and dysarousal systems already discussed (glutamate and GABA) but also the well studied ascending DA, NE and 5-HT circuits, as well as the less appreciated ones such as histamine pathways. Most of these systems are responsive to certain bodily state variables, such as levels of circulating steroids and vigilance states. Also, it should be noted that while the NE system is highly responsive to all types of stimuli that alert animals, the 5-HT system is comparatively unresponsive to world events, increasing its activity only modestly to highly stressful stimuli. One could propose that NE and DA arousal and 5-HT quiescence promote drive, while catecholaminergic quiescence and 5-HT arousal tend to reduce the overall perceived feeling of drive tension.
             However, for all this to be a useful exercise, it should provide some explanatory power for existing observations concerning the normal human psyche. There simply is not sufficient data at that level to proceed very far. Indeed, carefully conducted studies that specifically modify these systems are just becoming available, and so far the results generally indicate that global affective/personality tendencies can be changed in predicted directions. For instance, increasing 5-HT in the normal human brain reduces negative affect and increases social cooperation without influencing positive affect (Knutson, et al., 1998). Thus, we can be confident that affect intensity can certainly be linked to the synaptic levels of such neurochemical activities, but I expect that the qualitative aspects of affective feelings do emerge largely from other, concurrently aroused, neuropeptide systems. In any event, the data base remains too meager to proceed confidently much farther on such fragile limbs of speculation.

IV. NEURO-AFFECTIVE BRAIN SYSTEMS ORCHESTRATE VARIOUS COHERENT INTERNAL AND EXTERNAL EXPRESSIONS CALLED THE BASIC EMOTIONS.
             I find Freud's distinction between the inner and outer perceptual and motor modalities an especially important one for modern neuroscience to consider. Obviously, within the behaviorist tradition in which most of neuroscience has been conducted, the outer expressions have received considerably more attention than the inner ones. Now that we are finally in a new era of consciousness research, the inner modalities that generate our subjective experiences and project them into the world are again becoming more prominent in discussions while remaining more difficult to analyze objectively. However, if we do not account for those inner dimensions in our theorizing (as neurodynamic entities rather than simply leaving them as mysterious or neglected forces), we shall come away with very faulty views of how the brain operates. I suspect that at least half of what is important in understanding mind/brain functions is to be found among the inner-directed processes that Freud courageously outlined.
             All serious neuroscientists should be willing to entertain how subjectivity--affective and otherwise--arises from brain matter, and how such processes are intermeshed with the long-term regulation of behavior. There is no question that a great deal of emotional behavior, usually the rapid-reflexive variety, requires no precipitous arousal of affective consciousness in order for emotional behaviors to be initiated (LeDoux, 1996), but we may be seriously amiss in our judgment if we believe longer-term emotional processes do not control future behaviors. And in that trailing role, affective states are not just epiphenomena; they surely have longer-term regulatory effects on future behavioral outputs. In other words, internally experienced emotional feelings may be seen as long-term modulatory influences which figure most heavily in the planning of behaviors and the more deliberate selection of future actions, rather than in the mere emission of eruptive emotional acts (which is only one component, and a transient one, of the overall emotional response). In the same way, in young people thoughts may more commonly follow impulsive actions than precede them. In other words, most conscious brain activity, affective as well as cognitive, is not simply devoted to generating behavior, but for dwelling on future behavioral strategies. In their long-term regulatory role, affective states can create inner turmoil and conflict in people's lives leading them to seek assistance (therapy) or to indulge in psychological and behavioral diversions (fixations, sublimations and displacement behaviors of various kinds).
             It is probably fair to say, that for humans, more theorizing has been devoted to the sensory/appraisal side of the affective integration that to the motor side. I believe that more effort should now be devoted to the motor side, recognizing that mind is typically embodied in actions. This may sound like a perplexing assertion, since behavioral neuroscience has in fact expended most of its resources to analyzing explicit motor processes. In clarification, I would suggest that in situations where explicit actions are inhibited by learned social-display rules, many motor responses of the nervous system continue to be reflected in various bodily tensions, small twitches, postures, gestures as well as many other local bodily effects, from autonomic to hormonal. These small motor responses should figure as heavily in our analysis of affective states of consciousness as the grosser forms of action readiness that are typically manifested in the varieties of instinctual motor actions that characterize the eruptive emotional states of animals. Indeed, psychoanalysis is better positioned than most other disciplines to try to analyze these small but powerful outputs that seem to reflect a battle between the various id-energies and ego-controls, but verbal approaches may need to be supplemented with ethological analysis of both gross and subtle neuromuscular tendencies. The nervous twitches of facial muscles and gaze aversions may eventually allow us to read a persons affective state more accurately than what they say, especially if it turns out, as some lines of research suggest, that the speech functions of the left hemisphere evolved as much for lying, cheating and deception as for straightforward communication.
             Freud recognized such issues in his concept of «bound» affective energy, and in this context S & N encourage us to conceptualize how voluntary/instrumental actions emerge developmentally from the more primitive affective actions of the nervous system. A simple and straightforward answer to this may be that changes in internal affective states, as they relate to both implicit and explicit motor actions, probably reinforce the preceding behavior patterns. Emotional states may constitute the major reinforcers for the development of long-term behavior patterns. An underutilized but complementary approach to the problem may be to search for the neural instantiation of certain emotional affects more within the motor organization of the brain rather than among neural systems that are closely related to the afferent side (which may contribute more to the classical motivational rewards, as in the pleasures of gustatory and other bodily sensations).
             It is widely recognized by embryologists that motor competence emerges in the nervous system earlier than sensory guidance, suggesting its primacy in brain evolution. For instance, in the chick embryo, we observe coherent whole-body motor movements before they are capable of being modulated by sensory inputs (Oppenheim, 1991; Provine, 1980). This primacy of the motor system should be of considerable interest for those who wish to clarify affective consciousness, and Damasio (1994) has emphasized the role of the whole body in affective experience. However, rather than focussing on the material body (which is obviously important for harvesting all manner of sensory and reafferent impressions), I suspect that it is the neurosymbolic representations of the «virtual body» of the SELF within fairly low levels of the brain that will give us the greatest leverage in understanding the fundamental nature of affect within the brain.
             The distinct resonances of emotional systems on such a «virtual body» may yield emotion specific neurodynamics, which are broadcast widely in the brain, thereby constituting the very essence of emotional feeling. As feelings are transmitted through the brain, they presumably interact with the information buffering abilities of higher brain areas. It may be through these associative interactions that «free» (instinctual) emotional energies become «bound» into higher networks of inhibition and regulation. In other words, in line with traditional Freudian thinking, the primitive id functions as well as basic SELF-identity functions are closely related to the «free» affective energies emerging from the emotional command systems already discussed, while the more deliberate, volitional and morally directed actions (the higher ego functions) may only emerge through the auspices of various form of internal inhibition, the information buffering abilities of working memory and higher brain modules such as those that can elaborate theories of mind (Baron-Cohen, 1997). In this view, much of affective feelings as well as the ensuing thought patterns follow rather than precede eruptive behaviors. In other words, most of affective feeling and thinking is devoted to the generation of future behavioral strategies rather than the generation of instinctual actions.
             I suspect that the binding of «free» affective energies is largely related to how well local GABA and 5-HT systems throughout the higher reaches of the brain are able to keep a tight lid on the primitive emotional energies that can well up from below. Obviously, the more effectively an organism can buffer and plan its behavioral responses, the more adaptive it will be in competing for resources. But, as Freud realized, in humans this is a double-edged sword. By actively suppressing primitive emotional energies in the service of long-term ego goals, one is at risk of having those emotional energies redirected in unexpected ways. They may then percolate up neurotically in spheres of life where they may disrupt and modify other aspirations and goals. This heart of Freudian theory remains a fertile field for conceptual and empirical development in brain research. Indeed, displacement behaviors are common occurrences in animals that are emotionally aroused but not allowed to dissipate their «energies» in goal-appropriate ways.
             As S&N emphasize, one brain region that certainly exerts regulatory control over many basic affective energies is the ventromedial region of the frontal lobes. Because of their decisive role in regulating emotions by promoting certain types of obsessive ruminations, these brain regions became the ultimate targets of destruction during the bygone era of psychosurgery. The clinical benefits seen from frontal lobe excisions were optimized by restricting damage to the ventromedial quadrants of the frontal lobes (Scoville & Bettis, 1977), and consisted mainly of freeing patients from dwelling on certain negative life events. These individuals could still exhibit practically all forms of simple emotional arousal, but they would not become entrapped in such episodes. Across the years, it has become clear that other higher limbic areas serve similar roles for other affective processes (Figure 2). For instance, social-affective issues, which characterize depression and obsessive-compulsive disorders, find a focus of cogitation within anterior cingulate areas (Drevets, et al., 1997; Mayberg, et al. 1997). Anxious and angry thoughts probably find executive domains in lateral and medial temporal lobe/amygdala zones, respectively, even though frontal areas are not without influence (Gloor, 1990).
             There are reasons to suspect that the primary excitant for all such cogitations is glutamate, and the main inhibitor is GABA, with a more general upward and downward tuning of information processing in these areas exerted by NE and 5-HT, respectively, and ACh helping construct an attentional searchlight which is directly under the control of the emotional-SELF. Individual neuropeptides apparently bring restricted parts of the widely distributed affect control mechanisms to bear on specific types of life problems. A detailed understanding of how so many systems work together, and how they construct the abstract neurodynamics of higher psychological processes, remains a great challenge for all of the mind sciences.

V. CONCLUSIONS
             It is a daunting task to build solid linkages between psychoanalytic ideas derived from clinical observations, especially ones that have not been adequately subjected to standard modes of scientific validation, and the incredibly rich empirical findings of modern neuroscience, which have typically been harvested in the most austere spirit of logical positivism. Regrettably, modern neuroscience has not been adept at conceptualizing how the internal neurodynamics of the brain weave psychological realities by blending evolutionarily provided abilities with neurodynamic symbolizations of ongoing world events. This is understandable; to accept the potential existence of hidden functions, neuroscience would have to drop it's ultrapositivistic veil. It will not do so, until the data massively and unambiguously mandates acceptance of a deeper integrative reality/complexity behind the surface appearances that we can directly measure. In my estimation, compelling evidence for this has long been available for both humans and related animals, but the neuroscientific community has chosen not to pay attention.
             In any event, during early phases of vertebrate brain evolution, a variety of special-purpose neural mechanisms were created that allowed animals to automatically cope with a limited set of life-challenging situations. Many of these processes reside in the spinal cord and lower brain stem. On top of such relatively reflexive and predictable abilities, evolution added more general-purpose orchestral functions that could coordinate various bodily functions, to allow organisms to cope more flexibly with key survival issues. Some of these coordinating systems are the basic and archetypal emotional circuits that all mammals share, heavily concentrated in midline mesencephalic and diencephalic structures and in higher zones traditionally conceptualized as the limbic system. As competition for resources became more intense, general-purpose learning mechanisms added layers of cognitive flexibility, permitting animals to conceptualize their circumstances so they could behave with varying degrees of foresight and hindsight. These biologically subtle, abstract abilities arise from most recent neural developments of the mammalian brain such as the neocortex.
             In the human brain, with its truly remarkable degree of encephalization that transpired only within the passage of a mere few million years, higher order emotional regulatory mechanisms (superego functions) have been added in such profusion that it is a major scientific challenge to disentangle the resulting strata of influence and counterinfluence. However, we can be certain that the emotional values that were established during earlier phases of brain evolution have remained intimately, coordinated with more recently evolved functions. Among these new affective skills we find a large number of socially constructed emotional tendencies which owe their existence definitively to cognitive functions. Thus, it is impossible to imagine feelings like jealousy and shame without there existing certain type of thoughts, and hence working memories, in the brain. On the other hand, the basic emotions can probably all be experienced without any preceding cognitive activity (at least of the variety that requires neocortical participation), even though there are abundant cognitive changes when basic emotionals are aroused, leading to a variety of secondary emotional constructions. I believe that the primal nature of emotions, conscious and unconscious, must be premised on our understanding of the subcortical neural processes which coordinate certain types of action readiness, seen in their most dramatic but transient forms in various eruptive emotional outbursts. However, the volcano of active mood states can continue to smolder even when such short-lasting emotional eruptions are not evident. In humans, higher ego and super-ego functions can keep these eruptive forces from being expressed outwardly, but much less so inwardly.
             Freud was among the first to probe into these darker areas of the human psyche, and he sought to untangle, at a conceptual level, how such functions might generally be organized and internalized within the psyche. Our concern here is to discuss whether his insights can be linked credibly to modern neuroscience. I believe many can. But will such exercises provide new avenues of thought that can promote the maturation of functional neuroscience and psychoanalysis? We must wait and see. There is much to gain and little to lose in pursuing such endeavors, especially if the work remains empirically focussed.
             I believe the «great intermediate net» of the brain can only be disentangled if global psychodynamic, and molecular and systems neuroscience approaches can be brought to bear on problems of mutual interest. To bring this to pass, psychoanalysts will have to invest more intensely in the study of experimental manipulations, especially psychopharmacological ones, where human verbal reports of internal experiences and dynamics are studied after systematically induced transient changes in the arousability in specific brain systems. We finally have a great many experimental tools to pursue such studies well. Psychoanalysis has carefully nurtured a listening approach that can minimize the use of loaded questions and other demand characteristics. Unfortunately, the transcription and content analysis of free associations is a tedious and a tricky business, but computerized transcription techniques can now be implemented. Such approaches should be supplemented with spectral analyses and electronic processing of the acoustic data, as well as the development of new qualitative tools to systematically probe the psyche.
             But for all this to be useful, we must encourage individuals to speak rather directly and systematically about their internal experiences, perhaps in controlled settings where the experimenter/therapist is not in the physical presence of the person being studied. We must cultivate a greater willingness to ask appropriate questions concerning the affective changes that various individuals experience in various situations. There may also need to be some careful personality descriptions and perhaps selection of individuals for such studies. At times it may be essential to have well-primed, cooperative individuals who are willing to try to communicate with their less deceptive right hemispheres. In order for us to access their minds effectively, our subjects should feel like collaborators in the research as opposed to simply guinea-pigs to be observed. The establishment of such working alliances may run against the grain of certain neuropsychological and psychoanalytic traditions, but I think both the ethics of research and methodological exigencies mandate we fully consider such courses of action.
             What might protypical experiments be like? First, the questions asked should be based on solid, and theoretically interesting lines of work from animal brain research which suggest especially clear and safe ways to modify emotional reactivity. Not only would it be interesting to catalog the effects of the many drugs that modify cholinergic, glutamatergic, and biogenic amines systems, but soon we will be able to study neuropeptide systems as well. Already there are non-peptide antagonists for Substance P, and corticotrophin releasing factor (CRF), and several other systems. The preclinical data suggest that these drugs should reduce anger, fear and separation anxiety in humans (Baram, Koutsoukos, Schultz, & Rivier, 1996; Panksepp, 1998a; Siegel, Schubert, & Shaikh, 1997).
             Of course the peptide system that has had a clinically available antagonist for the longest time is the opiate receptor system, and I would suggest that careful psychoanalytic studies are warranted with drugs such as naloxone and naltrexone. For instance, from animal research, it is evident that brain opioids constitute a major part of a generalized «pleasure principle» in the brain. Animal work has clearly indicated that opioids mediate the positive incentive properties of tasty foods, sex, play and other positive social feelings (Panksepp, 1998a). Surprisingly, past research has suggested that such drugs have remarkably modest affective consequences in humans (Panksepp, 1981), but in fact, there is a paucity of carefully collected data to analyze how blocking brain opioid receptors with long-acting antagonists, such as naltrexone, modifies human affective experiences, especially social ones. A careful psychoanalytic probing of the feelings of individuals challenged with opiate antagonists should reveal many subtle changes in the underlying fabric of affect, especially when they encounter various emotion provoking stimuli, ranging from those that arouse our basic appetites to more cultural ones such as art and music. One should be prepared to observe a variety of subtle changes, including an increase in social sensitivity and a tendency to behave differently to the feelings of others.
             Indeed, opiate antagonists have provided such effects in the treatment of autistic symptoms (Kolmen, Feldman, Handen & Janosky, 1997). In our experience with this medication in the treatment of autism (Panksepp, Lensing, Leboyer & Bouvard, 1991; Bouvard, et al., 1995), modest benefits have been evident in about half the children, especially the high-functioning ones. I am especially fond of insight provided by a highly self-centered and emotionally-aloof 17 year old teenager who had good language skills. On naltrexone, she became more sensitive and concerned with the feelings of her parents, but when asked what had changed inside her, she seemed perplexed and answered that nothing had changed within her. Only the world had changed from her point of view--other people were simply behaving differently. And perhaps they were, through subtle interactions, that arose from her increased intimacy with their lives.
             This, of course, is the great dilemma of the subjective phenomenological view. Our affects are value constructions of the brain, and our emotional systems are designed in such a way as to project our feelings into the world. We are not just angry; we are typically angry with someone who seems hateful. We are not just in love, but we love someone who seems lovable. We are not just emotionally moved by the music we enjoy, but the emotions actually appear to flow directly from the music. Even as we recognize that the information triggering the feelings is encapsulated within the well-interpreted musical score, the resulting mood changes arise from the dynamic responses of our brains. Recently, Fried and colleagues (1998) stimulated the human prefrontal cortex and provoked feelings of hilarity in humans with nothing special having happened in the environment, and these feeling were also projected outwards. During the brain stimulation, everything the individual focussed on seemed funny.
             In short, our brains are designed to project affect (as well as perceptions, of course) back into the world, and even our animal subjects seem to imbue neutral environmental events with affective salience for they exhibit contextual conditioning remarkably easily. This is the way the brain generates its highly adaptive illusions of emotional realities, and that may also be why behavioral scientists are so timid in trying to deal forthrightly with such brain processes. After all, by accepting the probable existence of emotional feelings in other creatures, we may only be studying our own anthropomorphistic tendencies, rather than the functional competencies of other animals. However, the available data, taken together, do suggest that we are not deluding ourselves, for a neurochemical understanding of the brain mechanisms that generate the affective behavioral tendencies in other animals can predict many of our own feelings (Panksepp, 1998a). Similar types of correspondence maps cannot be generated for thoughts. Thus, in the restricted arena of basic emotions and motivations, careful, scientifically-advised anthropomorphism is a viable preliminary strategies for identifying which types of value-encoding processes exist in other brains. Through a detailed study of the underlying neural mechanisms in animals, we can now achieve a general understanding of how homologous neural processes operate in our own minds. Now we need to figure out how the animal brain actually generates affect, but that work is still proceeding remarkably slowly, partially because of the conceptual blinders that behavioral neuroscience has imposed on itself and the resulting research funding policies.
             In pursuing parallel phenomenological studies of subjective emotional experiences in humans, we are confronted by one enormous obstacle: The speaking hemisphere appears to be a master of confabulation and deception in its appointed role of verbally communicating with others in the world (LeDoux, 1985). Social-desirability factors appear to be much more important to the left hemisphere than to the emotionally deeper and more sincere, right hemisphere. This hemispheric specialization of emotional values was strikingly evident in a recent study by Ross, Homan, & Buck (1994), where individuals changed their heartfelt emotional confessions toward more superficial/cavalier directions when their right hemispheres were selectively anesthetized.
             What problems does this pose for future psychoanalytic work? Would subjects coming to participate in a study be more left-hemisphere oriented, focussing their psychological resources toward sustaining a high level of social desirability? If so, would they be less likely to reveal their deeper emotional feelings? If the left-hemisphere is, in fact, an expert at emotional repression (i.e., having a preference for an alexithymic communication style, especially in males), research would have to try to work past that barrier. This is where the psychoanalytic tradition may be a special blessing to empirical pursuits in the area. Testing situations where individuals are willing to confide their inner lives may help us see, more clearly, the deeper feelings that reside beyond the surface veneer of social desirability. Indeed, one could imagine that such studies could eventually be used to effectively probe, with full client feedback, which types of psychotropic medications might have optimal effects for which patients.
             In closing, I would again emphasize that it is unlikely that neuroscientific findings will match up precisely with psychoanalytic concepts. Also, the levels of complexity already revealed at the neuroscience level, mostly derived from animal brain research, are so vast that no one can have confidence in relating them to human psychodynamics that arise from the immeasurably complex interactions of many neural systems. Still, there is an incredibly rich neuroscientific data base from which we can derive testable ideas, and after pre-clinical testing, the best concepts should eventually be taken to psychoanalytic laboratories.
             Obviously, most of the basic knowledge concerning human emotions remains to be collected. We do not yet understand the natural time-courses of emotional episodes (even though some progress is being made: Potegal, Kosorok, & Davidson, 1996); we know little about how various emotions interact; few have tried to characterize the real oscillatory neurodynamics of the affective states (using not only electronic technologies but various depth psychological approaches). A great deal of careful behavioral, psychological and psychoanalytic work is needed to answer such questions, but ultimately, our explanations for many human feelings must be based on our understanding of the neural processes that control homologous processes in related animals.
             In pursuing such lines of inquiry, we should be under no illusion that neuroscience has progressed farther toward a substantive understanding of emotional processes than it actually has. Within the grand edifice of modern neuroscience, the study of emotions remains little more than a cottage industry, even though leaders in the field are recognizing that «the study of emotion is enjoying a renaissance» (Hyman, 1998, p. 417). Still, to get support for such work, it is best not to mention that you are interested in understanding emotions, especially in animals. Feel free to highlight learning or any variety of other accepted brain mechanisms, but please, please do not mention that you are interested in understanding the fundamental nature of affective experiences.
             The majority of neuroscientists (and hence peer-review panels) still believe such questions are outside the purview of standard science. This is most perplexing, for we can be certain that a great deal of brain activity is devoted to creating the affective infrastructure upon which our cognitive abilities are built. Psychoanalysis has recognized that fact, and the systematic study of human affective experience is one of the major pillars upon which a lasting understanding of emotions must be built (the others being careful studies of the brain and the accompanying bodily/behavioral changes). It seems evident that in brain evolution, as in the psychological development of each individual, feelings preceded thinking, and that practically all thoughts continue to be channeled by emotional and motivational processes, both conscious and preconscious.
             To understand how mind emerges from brain functions, we must be willing to conceptualize new coherent entities such as teh SELF. We should be willing to consider that the brains of all mammals can generate various basic feelings, including joy. By accepting such possibilities, we have recently discovered what seems to be a primitive form of laughter in laboratory rats (Panksepp and Burgdorf, 1998). If such research approaches grow and thrive, they can serve as a basis for useful cross-fertilization with psychoanalysis as well as the social sciences and humanities. It is time for neuroscientists to fully consider the evidence that points to the existence of internally experienced affective states which are causally efficacious in the governance of behavior. We should now seek a new consilience (Wilson, 1998), among the many disciplines that are honestly seeking to reveal the deeper layers of human nature. There should be a major scientific role for psychoa