Swanson, Larry - Vertical Systems from Spine to Cortex

The Structure of Concern Project compares many theoretical models from many disciplines to the Adizes PAEI model, arguing that they must all be reflecting the same underlying phenomenon. Several concern structure patterns are described below.

For several years now, research efforts headed by Larry Swanson at the University of Southern California have traced vertical anatomical connections in the rodent brain. Systems at the cortical and striatopallidal level have been linked to diencephalic and brainstem/spinal cord systems. The amygdala has played an important role in the mediation of these connections, and perhaps the most provocative claim arising from this research has been that the amygdala is not a structure unto itself, but rather several structures that can variously be assigned cortical, striatal, and pallidal functions.

Fourfold distinctions play important roles in these vertical anatomical and functional systems. Four systems feed into the hypothalamic area from the cortex, participating in the regulation of four hypothalamic functions. Those functions are involved in the regulation of three broad categories of behavior: ingestive, defensive/agonic and reproductive (and autonomic regulatory activity, considered independently from these three/four functions). These three categories might be expanded to four if defensive behavior and agonic behavior were differentiated. Framing all of these relationships is an overall zoning of the nervous system into four functional systems: motor, sensory, cognitive and behavioral state control. (Swanson, 2003, p. 95)

The various patterns of segmentation along the entire vertical nervous system demonstrate an oblique but consistent relevance for PAEI distinctions, well within the penumbra of the fuzzy concept being explored. Furthermore, Swanson’s model involves the divergence and convergence of information along this column between the various four-part layers. This networking among various four-part layers provides a possible mechanism for one quadrant of a system to modulate the three others, biasing responses towards a dominant or preferred style/subsystem.


Starting at the base of the column, Risold, Thompson and Swanson (1997) describe a visceral counterpart to the central pattern generators for somatomotor behavior in the hindbrain and spinal cord. Cell regions in the ventromedial diencephalon are organized and positioned such that they could generate similarly patterned activity over neuroendocrine and autonomic responses rather than somatomotor ones. They call these visceromotor responses, which are the output of a hypothalamic visceromotor pattern generator network (HVPG). The HVPG operates alongside a behavior control column (BCC) for controlling motivated behaviors, particularly ingestive, agonic, defensive and reproductive. The BCC involves both hypothalamic and midbrain/hindbrain nuclei. The hypothalamus is thus involved in generating both internally-directed visceromotor and externally-directed somatomotor patterns related to core survival and reproductive activities, and the strong motives that accompany those.

The hypothalamus does receive direct sensory information that would be relevant for releasing visceromotor responses. It gets information about environmental light from the retina, through its connection to the suprachiasmic nucleus (SCN) most notably, and also to the subparaventricular nucleus that is the heaviest target of SCN afferents, among other nuclei involved in circadian rhythms and autonomic responses.

A second pathway reaches these same two nuclei from the ventral lateral geniculate nucleus. The hypothalamus is also a major target for both main and accessory olfactory information. Caudolateral areas of the lateral hypothalamus receive olfactory information from the medial forebrain bundle (MFB), from sources such as the olfactory bulb, piriform cortex and amygdala (Risold et al., 1997). However, this direct sensory input is likely to be more modulatory or regulatory in nature. The staging of the visceromotor event itself involves higher brain systems that intersect with or converge upon the amygdala in important ways.

Swanson’s group views the amygdala as a name given to a group of nuclei pushed together in the brain by happenstance, forming neither a functional nor a structural unit. Rather, nuclei in the amygdalar region belong to three distinct groups:

  • The caudal olfactory cortex (cortical, piriform and postpiriform amygdala, nucleus of the lateral olfactory tract);
  • The ventromedial claustral complex (lateral, basal and posterior nuclei), and;
  • The caudoventral striatum (central, medial and anterior amygdala and intercalated nuclei). (Petrovich et al., 2001; Dong et al., 2001; Swanson & Petrovich, 1998).

These three amygdalar areas participate in four major telencephalic systems – frontotemporal, visceral, accessory olfactory, main olfactory – described in PAEI order below:

P: A frontotemporal system can be outlined that might mediate quick survival-related action. It includes many gustatory and visceroregulatory areas of the ventral forebrain, including the medial orbital area, insula, ventral temporal cortex and hippocampus. This system involves the anterior basolateral amygdala and lateral amygdala, with outputs mainly to the somatomotor system via the ventral and dorsal striatum. There is a projection from the lateral amygdala to the central amygdalar nucleus, which is involved in autonomic pattern generation, and which influences the striatum via the substantia nigra. The setup suggests a very quick shift from bio-relevant information processing into action, accompanied by supportive autonomic effects.

A: The visceral system is a narrower version of the frontotemporal system in some ways, and seems more geared towards the avoidance of unpalatable events. It includes agranular insular areas (primary gustatory/degustatory cortex), medial prefrontal areas directly involved in visceroregulation, and the subiculum, though by some to play a key role in anxiety and fault-detection (McNaughton & Gray, 2000). The main amygdalar target is the central nucleus, and below that the lateral hypothalamus, in regions that innervate many autonomic cell groups in the brainstem and spinal cord implicated in conditioned fear responses.

E: Accessory olfactory functions in the rat and other animals processes non-volatile chemical compounds with biological significance for the animal, most notably the pheromones released by conspecifics indicating reproductive status, territory markings and social status/dominance/eminence. (By contrast, the identification of specific group members such as mother, child, littermate etc. involves the more targeted chemical analytics of the main olfactory system.) The accessory or vomeronasal system is for broad-brush social status judgments relevant to one’s own motivational state. In humans the vomeronasal organ itself is vestigial. Social computations of all kinds have largely been captured by higher, non-olfactory, multimodal limbic and cortical systems of great complexity (so great that on some accounts it underpins all human cognitive expansion. Dunbar, 2003). Nevertheless, the old accessory olfactory pathways may maintain their social significance, under “new management”. The amygdalar components of the accessory olfactory system in rats – the anterior cortical amygdalar nucleus and the posterior nucleus – project to all functional zones of the hypothalamus, particularly reproductive and defensive areas, as well as lateral/autonomic areas and gonadotropic neuroendocrine areas targeted by the SCN (potentially involved in seasonal mating patterns). Thus this system seems to mediate social eminence, the staking out of territories and mating within social reference groups. Compared to the cycle-times for feeding or defense, reproduction is a much longer-cycle seasonal activity for most animals, only to be undertaken when the conditions and opportunities are right.

I: Main olfactory systems are finely-tuned evaluators of the biological relevance and significance of substances, and also of individualized social information. (In the early ontogeny of mammals, food value and social value information are phenomenologically identical, since our first food source is our mother.) The main olfactory systems involves five amygdalar cortical areas (the anterior and posterolateral cortical nuclei, nucleus of the lateral olfactory tract, a postpiriform transition area, and piriform-amygdalar area), along with parts of the claustral complex and the striatal anterior amygdalar area. Projections across the hypothalamus are comparatively light, with the notable exceptions of projections from the piriform-amygdalar, posterior basomedial and posterolateral areas, which heavily target the reproductive and defensive behavior control areas. This setup suggests the intensive, multivalent processing of socially-relevant information.

These fairly direct pathways represent one route for taking information from four forebrain systems down to four hypothalamic behavioral systems (ingestion, defense/agonism, reproduction, autonomic/neuroendocrine control). There are four more routes as well:

  • Via the bed nucleus of the stria terminalis (BNST);
  • Via hippocampal formation (HCF);
  • Via hippocampus and septum, and;
  • Via the medial prefrontal cortex (MPFC).

In the hippocampal formation, the entorhinal cortex is innervated throughout by amygdalar input, and it in turn innervates essentially the whole cortex, basal ganglia and hippocampus proper via the perforant path. It does not project to the hypothalamus or thalamus. The parasubiculum projects to the lateral mammillary bodies through the fornix. Projections into Ammon’s horn and the subiculum that traverse the septal area, however, target four hypothalamic areas. CA3 preferentially targets the caudal septum, and thus the lateral hypothalamus and supramammillary area. CA1 and subicular projections to the rostral and ventral septum project to the medial behavior control column and paraventricular/neuroendocrine zones, respectively.

Reciprocally connected amygdalar and hippocampal areas do not project in parallel to identical targets in the hypothalamus, offering a potential mechanism for considerations of one kind (amygdalar-focal) to influence or dominate another kind (hippocampal-situational) in directing motivated behaviors and self-regulatory responses. In the rat, the central amygdalar nucleus projects to the lateral hypothalamus, but not to the hippocampus. It is involved in conditioned fear. The hippocampal projection to the lateral hypothalamus arises in CA3, involved in novelty detection. The lateral hypothalamus influences autonomic reactivity. Fear versus fascination in response to novelty is a key differentiator between A and E. This interaction in the lateral hypothalamus could be one of the switches for setting up A or E style behavioral syndromes. Similarly, focal versus situation reactivity differentiates P and E. P and A share a negative bias towards novelty.

In terms of ascending projections, the hypothalamus uses at least four routes for sending information to the telencephalon: 1) a massive direct projection to the entire cortical mantle (a candidate for biasing the brain towards a dominant style?); 2) indirect relays through the thalamus (attention, learning, searching/foraging, activity switching); 3) the basal ganglia (action and motivation), and; 4) brainstem structures like the periaqueductal grey, superior colliculus, cuneiform nucleus and ventral tegmental nuclei of Gudden, through the medial ZI, ventral anteromedial thalamic nucleus and rostrodorsal nucleus reunions. A rough PAEI labeling of these rising projections could be made as follows:

P: Basal Ganglia – action and motivation
A: Brainstem – vigilance, quick corrective responses
E: Thalamus – information processing, scene-building
I: Global – all aspects of cognition involving interoception and social/visceral concerns.

All PAEI mappings in this section are highly provisional, but the resonances between Swanson’s framework and other concern structure models in psychology are worth emphasizing. A more careful analysis of these issues might contribute to a biological basis for temperamental differences, among other things. It would also be important to connect this biological organization to the ecological conditions of its emergence.


Swanson’s differentiation between four systems and four functions through the vertical brain is maintained in anatomical studies of the bed nucleus of the stria terminalis (BNST). Noting that the anterloateral BNST is composed of four cell groups with dense local interconnections, Dong and Swanson (2004) identify four subsystems that receive projections from this structure, given below in PAEI order:

P: Somatomotor system (nucleus accumbens, substantia innominata, ventral tegmental area, and retrorubral area and adjacent midbrain reticular nucleus)

A: Central ANS (central amygdalar nucleus, dorsal lateral hypothalamic area, ventrolateral PAG, parabrachial nucleus, and nucleus of the solitary tract)

E: Thalamocortical feedback loops (midline, medial, and intralaminar nuclei).

I: Neuroendocrine system (paraventricular and supraoptic nuclei, hypothalamic visceromotor pattern generator network)

The posterior BNST has three divisions, and seems more integrative, handling both topographically separate and converging projections to various cerebral structures.

PAEI themes are traceable in many studies of the vertical brain. For example, distinct, longitudinal neuronal columns have been identified within the midbrain periaqueductal gray (PAG). There are dorsolateral or lateral columns which are associated with active coping strategies (e.g. confrontation, fight, escape), and a ventrolateral column associated with passive coping strategies (e.g. quiescence, immobility, decreased responsiveness). Active strategies are usually recruited when the stressor is perceived as controllable or escapable, and passive strategies come into play when the stressor is perceived as inescapable. This maps very neatly onto the P-A distinction, in terms of both the behavior, and the ecological conditions that make the behavior adaptive. (Keay & Bandler, 2001)

The rostral lateral periaqueductal gray (PAG) has also been shown to play a role in the inhibition of hunting or predatory behavior and the release of maternal behavior. Lesions to this region strongly inhibit hunting and restore maternal behavior, indicating that some kind of P-I switch may be found in this region (Sukikara et al., 2006). A full mapping of these vertical relationships and PAEI “switches” would be a research project unto its own.

1. Swanson, L. W. (2003). Brain Architecture: Understanding the basic plan. Oxford: Oxford University Press.
2. Swanson, L. W., & Petrovich, G. D. (1998). “What is the amygdala?” TINS: Trends in Neuroscience, 21, 323–331.
3. Risold, P. Y., Thompson, R. H., & Swanson, L. W. (1997). “The structural organization of connections between hypothalamus and cerebral cortex.” Brain Research Reviews, 24, 197-254.
4. Dong, H.-W., Petrovich, G. D., & Swanson, L. W. (2001). “Topography of projections from amygdala to bed nuclei of the stria terminalis.” Brain Research Reviews, 38, 192-246.
5. Dong, H.-W., & Swanson, L. W. (2004). “Organization Of Axonal Projections From The Anterolateral Area Of The Bed Nuclei Of The Stria Terminalis.” The Journal of Comparative Neurology, 468, 277-298.
6. Petrovich, G. D., Canteras, N. S., & Swanson, L. W. (2001). “Combinatorial amygdalar inputs to hippocampal domains and hypothalamic behavior systems.” Brain Research Reviews, 38, 247-289.
7. McNaughton, N., & Gray, J. A. (2000). “Anxiolytic action on the behavioural inhibition system implies multiple types of arousal contribute to anxiety.” Journal of Affective Disorders, 61, 161–176.
8. Dunbar, R. (2003). “The Social Brain: Mind, Language, and Society in Evolutionary Perspective.” Annual Review of Anthropology, 32, 163-181.
9. Keay, K. A., & Bandler, R. (2001). “Parallel circuits mediating distinct emotional coping reactions to different types of stress.” Neuroscience and Biobehavioral Reviews, 25(7-8), 669-678.
10. Sukikara, M. H., Mota-Ortiz, S. R., Baldo, M. V., Felicio L. F., & Canteras, N. S. (2006). “A role for the periaqueductal gray in switching adaptive behavioral responses.” Journal of Neuroscience, 26(9), 2583-2589.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License