The omnipresence of autonomic modulation in health and disease

Abstract

The Autonomic Nervous System (ANS) is a critical part of the homeostatic machinery with both central and peripheral components. However, little is known about the integration of these components and their joint role in the maintenance of health and in allostatic derailments leading to somatic and/or neuropsychiatric (co) morbidity. Based on a comprehensive literature search on the ANS neuroanatomy we dissect the complex integration of the ANS: (1) First we summarize Stress and Homeostatic Equilibrium – elucidating the responsivity of the ANS to stressors; (2) Second we describe the overall process of how the ANS is involved in Adaptation and Maladaptation to Stress; (3) In the third section the ANS is hierarchically partitioned into the peripheral/spinal, brainstem, subcortical and cortical components of the nervous system. We utilize this anatomical basis to define a model of autonomic integration. (4) Finally, we deploy the model to describe human ANS involvement in (a) Hypofunctional and (b) Hyperfunctional states providing examples in the healthy state and in clinical conditions.

1. Introduction

1.1. Autonomic function – a pervasive system in health and disease

The autonomic nervous system (ANS) is primarily entrusted with the maintenance of homeostasis. The peripheral and central ANS components are indispensable for the survival of the individual and species by controlling multiple vital functions including hemodynamics, thermoregulation, caloric/electrolyte balance or sexual/reproductive activities. The ANS is comprised of sympathetic and parasympathetic divisions usually with yin- and yang-type complementary roles responding to a wide range of endogenous and exogenous challenges. The sympathetic ANS division is actively recruited during threatening or stressful situations while at the same time the parasympathetic division subserving the resting state-related vegetative and anabolic processes is restrained. Whether the sympathetic-parasympathetic interactions are adaptive or maladaptive depends on their nature i.e., truly reciprocal and balanced or otherwise.

1.2. Structure and function

Under physiological conditions the parasympathetic system serves the trophotropic function aimed at recovery and resting purposes, whereas the sympathetic counterpart mediates the ergotropic effects, metaphorically termed “fight-or-flight”. This phrase coined by Cannon (1915) describes the rage and fear emotional correlates of the rapid homeostatic adjustment aimed at rescuing an organism at the time of an acute challenge to its integrity. Thus, the immediate response to stress is the sympathetic nervous system mediated release of catecholamines (Romero, 2010) and other neurotransmitters to maintain physiological and psychological homeostatic equilibrium - or simply homeostasis. If a stressful condition persists, the secondary response mechanism involves the activation of the hypothalamic-pituitary-adrenal (HPA) axis (Becker and Rohleder, 2019).

Despite the ubiquity of stress in daily life and extensive advances in our understanding of the molecular and cellular mechanisms underlying the consequences of stress, there is a relative paucity of data (Herman et al., 2003; Ulrich-Lai and Herman, 2009) of potential central ANS processes that may play a role in physiological processes as well as somatic- and neuropsycho-pathology. The majority of research has focused on the multidirectional interactions among the cardiovascular (CV)-, immune- and central nervous systems arising in the context of psychosocial stressors (Crestani, 2016; De Bosscher et al., 2008; Duivis et al., 2013; Fagundes et al., 2013; Hansson et al., 2015; Harrison et al., 2013; Heidt et al., 2014; Hodes et al., 2014; Lagraauw et al., 2015; McKlveen et al., 2015; Muscatell et al., 2015; Pace et al., 2020; Perova et al., 2015; Rosengren et al., 2004; Schaeuble et al., 2019; Song et al., 2019; Tawakol et al., 2017; von Kanel et al., 2007; Wang et al., 2010; Zhang et al., 2010). The psychosocial stress has shown predictive value for adverse CV outcomes that is equivalent to other major CV risk factors (Lagraauw et al., 2015; Nabi et al., 2013; Rosengren et al., 2004; Steptoe and Kivimaki, 2012; Steptoe and Kivimäki, 2013). It appears that acute (Baumhakel et al., 2007; Carroll et al., 2002; Gullette et al., 1997; Jones et al., 2016; Kloner, 2004; Kloner et al., 2009; Leeka et al., 2010; Leor et al., 1996; Meisel et al., 1991; Rawish et al., 2021; Reich et al., 1981; Schwartz et al., 2012; Trichopoulos et al., 1983; Wilbert-Lampen et al., 2008) and chronic (Bernberg et al., 2012; Gu et al., 2012; Ranjit et al., 2007; Rosengren et al., 2004; Schaeuble et al., 2019; Tawakol et al., 2017) stress exposure contributes to cardiovascular morbidity and mortality.

Models dissecting the central and peripheral autonomic nervous function within different topics such as inflammation, neurocardiology as well as significant contributions towards the understanding of the integrative role of the ANS are presented elsewhere (Jänig, 2006, 2014, 2016). In order to adapt to intrinsic or extrinsic influences targeting the ANS the central nervous system employs a cascade of operations to process sympathetic and parasympathetic information. In this review, we propose that this includes either an up- or down-regulation of the neuronal regions (ie. modulators/integrators), which are targeted by both autonomic counterparts, or the processing of either sympathetic or parasympathetic information by entirely different structures (ie. para-sympathetic/sympathetic effectors; see Fig. 1). However, caution needs to be exercised with a stringent designation of the brain regions as either sympathetic or parasympathetic as the majority of regions may express both sympatho- or parasympathomimetic features when down- or upregulated. Nonetheless, in following classification we provide a schematic overview of how some key brain ANS regions are functionally integrated. The sympathetic effectors encompass the infralimbic medial prefrontal cortex, the central nucleus of the amygdala, the caudate nucleus, the habenula, the rostro ventrolateral medulla, the locus coeruleus and the parabrachial nucleus. The major parasympathetic effectors include the prelimbic medial prefrontal cortex, the nucleus ambiguus and the motor nucleus of the vagus nerve. Other modulators/integrators orchestrate both sympathetic and parasympathetic information, examples of such regions include the insula, the paraventricular nucleus of the hypothalamus and the nucleus of the solitary tract.

Fig. 1. Components of the ANS.

The ANS entails an integrative as well as modulatory closed loop processing of efferent and afferent signals. Central integrators and modulators are pivotal for connecting central effector regions and peripheral afferent/efferent structures.

In this review, we dissect the complex integration of the ANS: (1) We summarize Stress and Homeostatic Equilibrium – elucidating the responsivity of the ANS to stressors; (2) We describe the overall process of ANS involvement in Adaptation and Maladaptation to Stress; and; (3) We discuss, – Four Modules: Contributions across the Nervous System – the ANS can be hierarchically partitioned into the peripheral/spinal, brainstem, subcortical and cortical components of the nervous system. We utilize this anatomical basis to define a model of autonomic integration. (4) Finally, we deploy the model to describe the human ANS involvement in (a) Hypofunctional and (b) Hyperfunctional states providing examples in the healthy state and in clinical conditions.

The preclinical and clinical English language peer reviewed literature search on ANS physiology and on ANS role in somatic- and neuropsycho-pathology along with the mechanisms of normative stress processing and their potential impairments in patients with representative ANS-associated disorders was undertaken using PubMed (http://www.ncbi.nlm.nih.gov/pubmed) from inception (June 2020) until February 2021. Medical Subject Headings’ terms used included ANS, catecholamine, norepinephrine (adrenergic), acetylcholine (cholinergic), sympathetic, parasympathetic plus cortex, subcortical, brainstem, afferent, efferent, insula, amygdala, striatum, habenula, locus coerules, nucleus of the solitary tract and the parabrachial nucleus. Information on the mechanisms and neurobiology of ANS and their derailment in various conditions were also drawn from recent seminal reviews of these topics (Averill et al., 2018; Charney, 2004; Goldstein, 2019; Hadaya and Ardell, 2020). The scopes of the review were adjusted based on consultations with scientists and clinicians, manual searches for relevant articles from the selected papers’ reference lists along with the utilization of PubMed’s “similar articles” function.

2. Stress and homeostatic equilibrium

Stress is an integral aspect of modern life and may be a contributing factor to the pathogenesis of multiple clinical disorders including CV disease, pain, respiratory disease, autoimmune disease, anxiety and depression (Chrousos, 2009; Cohen et al., 2007; de Kloet et al., 2005; Flaa et al., 2008; Golbidi et al., 2015; Gu et al., 2012; McEwen and Gianaros, 2011; Pitman et al., 2012; Rosenberg et al., 2014; Segerstrom and Miller, 2004; van Campen et al., 2014; Yehuda, 2002). Deepening insights into co-occurring stress and psychosomatic morbidity faces a major challenge inherent in the definition and operationalization of “stress.” This commonplace concept has been given nearly as many definitions as there are treatises that have dealt with it (Goldstein, 1995) “Stress” may be defined from the cognitive, emotional, endocrinological, psychophysiological, neurobiological, and molecular standpoints (to name a few). Accordingly, the term “stress” has been argued to be misused and abused and thus lead to confusion as reflected by the apposite chapter “Stress is a Noun! No, a Verb! No, an adjective!” of the book Stress and Coping by (Engel, 1985). Here we responded to this question by borrowing the psychological “intervening variable” construct denoting a causal link between a stimulus and response (Harmon-Jones and Gable, 2018; Mercer and Holder, 1997). Pain, for instance, is an intervening variable, in the form of a motivational state, which links an acute injury with the behaviors aimed at the avoidance of the tissue damage (Elman et al., 2018). In this theoretical framework, stress may be conceptualized as an intervening variable linking the discrepancy between genetically-programed, learned or situationally-deduced expectations and the actual or perceived internal- or external states with consistent compensatory patterns (Goldstein, 1995) c.f., the neuroscience prediction error concept reflecting the neuronal coding of the expectation vis-à-vis outcome discrepancy (Niv and Schoenbaum, 2008). In more general terms, stress predominantly refers to a situation that requires additional resources, which is why stress is often perceived negatively. The circumstance that adequately dosed stress may have a positive impact should not be neglected. In fact, the influence of “healthy” stress (ie. eustress) can be observed in various aspects of life such as in the speaker or actor that needs a certain level of stage “fright” for a strong presentation. Another example is the right amount of nervousness experienced by students before an exam that may improve the outcome especially before an oral exam. Athletes also report that a certain amount of nervousness enhances their performance. The concept of “healthy” eustress and pathological distress has been proposed by Selye (Selye, 1974). Additionally, he proposed the conditions of hypo- and hyperstress and concluded that balancing these conditions would lead to eustress (Selye, 1983), which in other words would foster the homeostatic equilibrium of stress response. A revised presentation of this concept pertaining to stress response within a healthy adaptive and a diseased environment is displayed in Fig. 2.

Fig. 2. The concept of eu- and distress.

Displayed is the gradual ANS response to a stressful event in an adaptive environment as well as the constrained ANS response in a maladaptive environment. The balancing of hypostress and hyperstress endorses eustress whereas a dysbalance in maladaptive processes induces gain of hypostress or hyperstress, this increases distress and thus reduces homeostatic equilibrium. The ANS response in health and disease will depend on the resilience, robustness and adaptation of the integrated functioning of the peripheral and central nervous systems controlling autonomic function.

Thus said, failure to establish a homeostatic equilibrium through the sympathetic and parasympathetic arms of the autonomic nervous system may lead to a buildup of allostatic load, which engenders maladaptation. A classic example may be a stress-related elevation of the blood pressure, which is an adaptive mechanism in terms of coping with short-term stress, but in the long-term, may lead to pathophysiological changes (allostatic load) e.g., atherosclerosis with an ensuing stroke or heart attack. As such ANS control in health and disease is a major contributor to allostasis and allostatic load that may be harnessed/trained to improve function in health and disease or left untethered and contribute to a deteriorating state of health. In other words, when referring to the above example; physical exercise, dietary adjustments as well as the administration of antihypertensive medication may have harnessed/trained the ANS and thus either prevented or decreased elevated blood pressure.

However, the ANS itself also may underlie central or peripheral alterations and thus impair the response to stress. In this review, the traditional concepts of autonomic stress response will be extended and an integrated approach will be provided to determine how changes in central and peripheral circuitry translate to changes in autonomic key physiological outcomes. The lack of a well-defined, integrative model that can be used to define the processes involved in the response to and recovery from stress as well as their magnitude and time course may be a major obstacle to overcome. Thus, this review addresses these challenges by developing a model which depicts the spectrum of central and peripheral autonomic changes within two conditions: (i) the adaptive state – entailing a return to resting state following acute perturbation via normatively functioning regulatory mechanism and (ii) the maladaptive state – altered or inadequate responses failing to restore or achieve homeostasis.

3. Adaptation and maladaptation to stress

The brain is a target of stress and is a key determinant of the response to and the recovery from stress (de Kloet et al., 2005; Henckens et al., 2015; Hermans et al., 2014; Joels and Baram, 2009; McEwen et al., 2015). We refer to the term “central stressors” as to a range of processes and states including biochemical, endocrinological, inflammatory or structural dysfunctions either interacting with or directly within the central stress circuitry. The activation of central stressors may exhibit a spectrum of stress responses ranging from autonomic nervous system (ANS) activation to HPA axis responses. Although some progress has been made showing that stress leads to change of gene expression as well as structural and functional remodeling of the brain (McEwen, 2007, 2013; McEwen and Gianaros, 2011). The effects of central stressors on the autonomic nervous system, the recovery from these stressors, and how to mitigate allostatic load (cumulative wear and tear of stress on the body) are not fully understood (McEwen, 2013). Interestingly, studies demonstrated that common brain disorders are characterized by disturbances in brain circuitry (Fox and Greicius, 2010; Vemuri et al., 2012). These findings emphasize the importance of a circuit-based approach to study autonomic brain function. Interestingly, there is no direct linkage of central structures such as the limbic system involved in stress processing to autonomic nerve fiber terminals in the periphery (Herman et al., 2003; Ulrich-Lai and Herman, 2009). Thus, the stress response presumably depends on a polysynaptic organization. In the present review, we further propose that the orchestration of the autonomic stress response involves a multimodal configuration (ie. through effectors and modulators/integrators) at different levels (Fig. 3). Components involved in ANS response include (1) cortical modules (e.g., insula) responsible for conscious and subconscious processing, (2) subcortical effectors (e.g, hypothalamus), (3) brainstem modules for in- and out-put integration (e.g, locus coeruleus) and (4) peripheral modules for in- and out-put servosystem feedback.

Fig. 3. Components of the ANS response function.

Displayed are the 4 different modules (cortical, subcortical, brainstem and spinal/peripheral) involved in stress processing. Each central module (2–4) includes parasympathetic and sympathetic effectors and modulators which orchestrate the ANS response. An operating principal can be attributed to each individual module.

From the above evidence, it is clear that ANS is not a unitary entity defined by a simple configuration of physiological processes aimed at homeostatic stability. On the contrary, it is both hierarchical and multidimensional, consisting of higher order homeostatic factors and lower order effectors and monitored variables, each of which plays a specific role within a successful compensatory adaptation for external/internal challenges or a creation of allostatic built-up. Modulation by the homeostatic vs. allostatic states as well as by such regulators as background medical ailments, neuropsychopathology and resilience to stress may produce a complex pattern of interactions leading to preserved health or to disease states discussed in the following sections.

4. Four modules: contributions across the nervous system

While the ANS is an integrated system, it is useful to consider it in the context of 4 modules. These can be evaluated in their contributions to ANS responses. The evaluation includes studies using functional approaches such as electrical or chemical stimulation of certain brain regions in combination with general autonomic measures (e.g. heart rate variability, heart rate, blood pressure change, etc.) as well as neuroanatomical measures of central autonomic circuits which mainly pertain to animal studies. So far, only few has been reported using functional brain imaging to assess autonomic circuits in the human (Chouchou et al., 2019; Dichtl et al., 2020; Hiser et al., 2021; Napadow et al., 2008). Two imaging approaches had a more functional approach by implementing cardiovagal modulation or applying electrical stimulation to certain brain areas (Chouchou et al., 2019; Napadow et al., 2008) whereas one study followed an anatomical and resting-state functional MRI approach (Dichtl et al., 2020).

4.1. Module 1: peripheral nervous system (PNS)

The spinal/peripheral module is the most exposed portion towards extrinsic factors such as disease of the peripheral nervous system or injuries and thus probably one of the most vulnerable parts of the autonomic nervous system. This module accomplishes three important tasks, which account for a servo-like feedback system (i) First, the perception of the external and internal environment via different receptors (e.g., C- and Aδ-fibers, chemo- and baroreceptors). (ii) Second, the conduction of afferent and efferent signals from the peripheral receptors/effectors towards the integration of module M2 within the brainstem and vice-versa; and (iii) the execution of autonomic responses via efferent autonomic nerve terminals (Fig. 4).

Fig. 4. Module 1: PNS.

The efferent pathways are fundamentally differently executed, the sympathetic arm origins at spinal levels C₈/T₁ – L₁/L₂ in the intermediolateral cell column and conveys signals via splanchnic nerves to the effector organ, whereas the parasympathetic arm directly functions via branches of the oculomotor (III), facial (VII), glossopharyngeal (IX) and vagus (X) nerve. The cutaneous innervation within the peripheral ANS consists of efferent autonomic C- fibers which are predominantly sympathetic (except for the face) as well as afferent polymodal C- and Aδ-fibers (high threshold mechanoceptors [HTMR], low threshold mechanoceptors [LTMR]) (Abraira and Ginty, 2013; Li et al., 2011; Roudaut et al., 2012). On a molecular level the efferent sympathetic C-fibers mainly function via adrenergic messengers to regulate vasoconstriction or pilomotor stimulation, but also regulate cholinergic sudomotor function and vasodilation. The sensory polymodal nerve fibers innervate hair follicles and the epidermis as well as convey the sympathetic axon reflex via collaterals to blood vessels. Importantly, a decreased density of these nerve fibers is an important marker for the presence of autonomic small fiber neuropathies, which has been associated with increased morbidity and mortality in diabetes and cardiovascular disease (Chan and Wilder-Smith, 2016; Freeman, 2020; Peters et al., 2013).

4.2. Module 2: the brainstem

The brainstem is one of the most integrative portions of the ANS. Its integrative function is a likely result of its strategic location in-between spinal/peripheral and subcortical/cortical autonomic structures. However, the role of the brainstem by far exceeds its exclusive role as an integrator of peripheral and central autonomic in- and out-put. The brainstem itself includes complex networks of autonomic circuitry which integrate central and peripheral stimuli as well as produce a significant amount of functional output.

Thus, the role of the brainstem may be considered an integrative generator. Additionally, it should be noted that in this location, subdivision of autonomic nervous system into sympathetic and parasympathetic components is an oversimplification because these divisions are anatomically and physiologically interconnected. Still in accordance with an analysis of connections towards distinct neuroanatomical structures (e.g. IML, vagus nerve, etc.) and the information of activation quality (ie. inhibitory or excitatory), a rough classification can be achieved (Fig. 5).

Fig. 5. Module 2: Brainstem.

The sympathetic regions include the locus coeruleus, which relays to the intermediolateral nucleus (IML) as well as shows inhibitory function towards the parasympathetic nucleus ambiguus (Samuels and Szabadi, 2008). Moreover, the rostral ventrolateral medulla, which relays directly to sympathetic preganglionic neurons within the IML (Oshima et al., 2008) and receives inhibitory input from the parasympathetic NA (McKitrick and Calaresu, 1996). Additionally, the lateral parabrachial complex mediates sympathetic outflow (Chamberlin and Saper, 1992) and modulates sympathetic activity via decrease of baroreflex response (Hayward and Felder, 1998). In contrast, the motor nucleus of the vagus nerve and NA both connect to the primary parasympathetic vagus nerve and interact parasympathomimetic with other regions in the brainstem (McKitrick and Calaresu, 1996; Samuels and Szabadi, 2008). The NTS has a special role as it relays to both sympathetic and parasympathetic integrative regions within the brainstem and thus can be considered as the primary modulator within the domain of integrators. References for afferent and efferent connections: LC- locus coeruleus (Gatter and Powell, 1977; Lopes et al., 2016; Samuels and Szabadi, 2008); RVLM- rostral ventrolateral medulla (Brown and Guyenet, 1984; Carrive et al., 1988; Dampney et al., 1987; Dempsey et al., 2017; McMullan et al., 2008; Oshima et al., 2008); PBN- parabrachial complex (Fulwiler and Saper, 1984; Krukoff et al., 1993; Tokita et al., 2009); NA- nucleus ambiguus (Nosaka et al., 1982; Núñez-Abades et al., 1990; Stuesse, 1982; Stuesse and Fish, 1984); DMX- motor nucleus of the vagus nerve (Ciriello, 1983; Holstege, 1987; Holstege et al., 1985; Hopkins and Holstege, 1978; Kalia and Mesulam, 1980; Takeuchi et al., 1983); NTS- nucleus of the solitary tract (Altschuler et al., 1989; Blessing et al., 1981; Chan et al., 2000; Ciriello and Caverson, 1986; Ciriello et al., 1981; Craig, 1995; Finley and Katz, 1992; Housley et al., 1987; Kalia and Mesulam, 1980; Kapp et al., 1979; Loewy et al., 1981; McDougall et al., 2017; Ricardo and Koh, 1978; Ross et al., 1981; Shin et al., 2008; Shipley, 1982; van der Kooy et al., 1984; Viltart et al., 2006; Wallach and Loewy, 1980).

4.3. Module 3: subcortical regions

The subcortex is of great importance for the preservation of the homeostatic equilibrium. Moreover, regions of this module are sharing circuits which are frequently altered in neuropsychiatric disorders (CNS-Type I) as well as neurodegenerative disease (CNS-Type II) and thus such overlap with autonomic processing circuits may alter the autonomic stress response in various conditions (Fig. 6).

Fig. 6. Module 3: Subcortex.

The central nucleus of the amygdala (AMYce), the caudate nucleus (CAN) and the habenula (HAB) are considered as autonomic effectors. An electric stimulation of the AMYce demonstrated sympathomimetic effects (heart rate ↑, blood pressure ↑). Additionally, a lesion of the AMYce prevented development of neurogenic hypertension after a 4-week daily stress intervention in rats (Baklavadzhyan et al., 2000). Interestingly, such a sympathomimetic effect was only shown in awake and not in anaesthetized rats (Iwata et al., 1987). The caudate nucleus (CAN) modulates blood pressure (Pazo and Medina, 1983) and salivary secretion, both presumably sympathetically, which has been corroborated by an attenuation of effects due to sympathectomy (Pazo et al., 1981). Interestingly, a stimulation of the rostral caudate nucleus increases blood pressure, whereas the stimulation of the caudal part induced hypotension (Pazo and Medina, 1983). This association of cardiovascular function with the striatum has even been argued to be the underlying mechanism for the cardiac sympathetic denervation in Parkinson’s disease (Goldstein et al., 2000; Pazo and Belforte, 2002). The role of the habenula within the autonomic response pertains to thermoregulation. It was shown that sympathetically driven brown adipose tissue hyperthermia and skin vasoconstriction was induced due to lateral habenula stimulation (Brizuela and Ootsuka, 2020; Ootsuka and Mohammed, 2015). Moreover, a similar thermogenesis pattern was observed due to emotional stress induced hyperthermia (Ootsuka et al., 2017). The paraventricular nucleus of the hypothalamus (HYPOpvn) relays to all of the previously mentioned regions and has emerged as a primary autonomic modulator involved in cardiovascular, renal and gastrointestinal, stress, metabolism, growth, reproduction and immune response (Buijs et al., 2003; Coote, 1995; Ferguson et al., 2008; Kalra et al., 1999; Swanson and Sawchenko, 1980; Williams et al., 2000). References for afferent and efferent connections: AMYce- central nucleus of the amygdala (Bienkowski and Rinaman, 2013; Fudge et al., 2017; Paredes et al., 2000; Rosen et al., 1991; Zséli et al., 2018); CAN- caudate nucleus (Çırak et al., 2020; Kotz et al., 2013; Rakić and Ivanus, 1985; Royce and Laine, 1984); HAB- habenula (Herkenham and Nauta, 1977; Namboodiri et al., 2016; Parent et al., 1981; Shelton et al., 2012); HYPOpvn- paraventricular hypothalamic nucleus (Geerling et al., 2010; Herman et al., 2003; Hernández et al., 2015; Li and Sawchenko, 1998; Ono et al., 1978; Pittman et al., 1981; Robert et al., 2013; Sawchenko et al., 1996; Silverman et al., 1981; Zhang et al., 2010; Ziegler et al., 2012).

4.4. Module 4: cortical regions

The cortical module is unique in that it may convert predominantly unconscious autonomic response to a conscious level, it was shown that the response to acute emotional stress was modulated due to lesioning either ventral or dorsal medial PFCs, suggesting that the ventral portion relates to autonomic modulation and the dorsal part to neuroendocrine stress response (Radley et al., 2006). Other than that, certain abilities such as the heartbeat control in high performance sharp shooters or archers (Açıkada et al., 2019; Caterini et al., 1995; Guillot et al., 2004; Tremayne and Barry, 2001), may relate to a very conscious autonomic control (Fig. 7).

Fig. 7. Module 4: Cortex.

The termination and initiation of autonomic responses is mediated by different areas of the medial prefrontal cortex (MPFC) (Powell et al., 1996; Tavares et al., 2009). A more detailed characterization of the vMPFC, targeting the infralimbic (ilMPFC) or prelimbic (plMPFC) showed that the ilPFC is involved in the initiation of the sympathetic ANS due to its diminished cardiovascular response (in terms of decreased tachycardia and pressor response) when inactivated (Frysztak and Neafsey, 1991, 1994), whereas an electrical stimulation endorsed blood pressure increase (Resstel et al., 2006). Conversely, the termination (i.e. activation of the parasympathetic ANS) of autonomic stress response may be ineffective due to an inhibition of the prelimbic PFC (Frysztak and Neafsey, 1994; Resstel et al., 2006). In line with these notions, the efferent cardiovagal response to a grip task and the central activation patterns assessed via fMRI were correlated, indicating a positive correlation with a prelimbic PFC activation, whereas a negative correlation was found towards an infralimbic PFC activation (Napadow et al., 2008). Interestingly, the insula was suggested to subserve both sympathetic and parasympathetic pathways. Thus, in epilepsy patients, electrical stimulation of the anterior insula increased parasympathetic tone, whereas the sympathetic control seemed to be located at the posterior insula region (Chouchou et al., 2019). References for afferent and efferent connections: ilMPFC- infralimbic medial prefrontal cortex (Hoover and Vertes, 2007; Vertes, 2004), plMPFC- prelimbic medial prefrontal cortex (Hoover and Vertes, 2007; Vertes, 2004), aINS (Evrard, 2019; Gogolla, 2017).

4.5. Integrating function – a model

An integrative model that includes function as well as morphological properties and characteristics is necessary to describe alterations in health and disease. The application of such a model in different conditions may extend the current view on this topic and provide a systematic overview. Thus, different conditions within health and disease as well as the respective origin of autonomic modulation (peripheral vs. central) can be distinguished.

In this model, the physiological properties are translated into a bioengineering setting (see Fig. 8). Thus, the term receptor in a biological context is extended by translating it into sensors/transducers within the bioengineering setting, which adheres to the complex pre-processing steps already at peripheral levels. Importantly, after peripheral and central signal processing the interpretation of the readout signal is based on the set point (i.e. allostasis), which however may be altered by itself due to disease. Comparable to a servosystem which entails a positive or negative feedback, parameters are then adapted to reach this set point. This model gives a holistic view on the topic and allows an application in various states with ANS involvement. Other than that, such an approach fosters the systematic characterization and categorization into normofunctional, hypofunctional and hyperfunctional ANS states.

Fig. 8. Engineering approach to ANS function.

The engineering model includes (i) sensors, (ii) transducers and (iii) a servosystem. The first two components relate to the afferent peripheral and central ANS, whereas the servosystem relates to the executive function of the efferent ANS pathways. The adequate response of the ANS depends on the allostatic load and thereby determines adaptive or maladaptive functionality. (1) Sensors/Transducers (= peripheral receptors of the epidermis, dermis and endothelium); (2) Autonomic Afferents; (3) peripherocentral relay (4) M1 (= spinal cord); (5) M2 (= brainstem); (6) M3 (= subcortex); (7) M4 (= cortex); (8) Readout; (9) Autonomic Efferents; (#9a) negative feedforward mechanism of sympathetic/parasympathetic activation towards CNS; (#9b) positive feedforward mechanism of sympathetic/parasympathetic activation towards CNS; (#9c) negative feedforward mechanism of sympathetic/parasympathetic activation towards PNS; (#9d) positive feedforward mechanism of sympathetic/parasympathetic activation towards PNS. This conceptual model can be applied within states of health and disease. For each example given in sections 4.1. - 4.3. the application to the integrative model is further explained.

Key: Σι = summation of afferent signals; Σo = setpoint of efferent signals; ∫ = integration of information; = resistor; = capacitor; =sender.

5. Normofunctional, hypofunctional and hyperfunctional ANS states

The ANS is an integrated system but within a condition there may be central or peripheral dominance of alteration. Moreover, the quality of the autonomic activation can be considered as normofunctional within certain boundaries as well as reach hypo- and hyperfunctional states with regard to a sympathetic or parasympathetic activation. The majority of conditions exhibit one distinct underlying pathomechanism, however other condition may also present two different underlying mechanisms which may augment each other and lead to a distinct clinical presentation such as the sympathoexcitatory effects in PTSD which may result from both sympathetic hyperfunction and parasympathetic hypofunction (Buckley and Kaloupek, 2001; Fonkoue et al., 2020; Park et al., 2017; Pole, 2007). Other than that, in some conditions such as in migraine the current state of literature provides a very heterogenic and inconclusive concept of underlying autonomic patho-mechanisms involved (Avnon et al., 2004; Cambron et al., 2014; Gass and Glaros, 2013; Gotoh et al., 1984; Rauschel et al., 2015; Tabata et al., 2000). A summary of selected conditions with their underlying patho-mechanism is provided in Table 1. Moreover, we discuss selected conditions in the context of stress, by including them into our integrative model we provide a systematic insight.

Table 1.

Examples of Autonomic Processing in Health and Disease.

Health		Disease
PNS	CNS	PNS	CNS-Type 1	CNS-Type 2
Food/feeding:	Nutritional status:
- Postprandial state in Young 1and Elderly 1↓ (Lipsitz et al., 1991, 1993; Oberman et al., 2000; Rokowski and Spodick, 1989) - Starvation 3(Landsberg and Young, 1983; Young and Landsberg, 1977) - Hypohydration 1 (Leib et al., 2016; Watso and Farquhar, 2019)	Periprandial state (taste evaluation) 1 (Harthoorn and Dransfield, 2008; Rousmans et al., 2000)	Baroreceptor Irradiation a–br↓ (Fagius et al., 1985; Heusser et al., 2005; Timmers et al., 2004)	Idiopathic Parkinson’s disease 3, a–br↓ (Goldstein, 2007; Orimo et al., 2002; Satoh et al., 1999)	Schizophrenia a–br↓ (Bär, 2015; Bär et al., 2007)
Exercise:	Exercise:
- during exercise 1 (Christensen and Galbo, 1983; Fisher, 2014), - after (aerobic) exercise 2,3 (Akselrod et al., 1985, 1981; Delius et al., 1972; Hagbarth and Vallbo, 1968)	-preparation phase 1,4 (Fisher et al., 2015; Williamson et al., 2006)	Guillain-Barré Syndrome 1 (Ahmad et al., 1985; Ventura et al., 1986; Zaeem et al., 2019)	Multiple System Atrophypre–g3 (Jordan et al., 2015; Parikh et al., 2002)	Panic Disorder 1 (Cameron et al., 1990; Villacres et al., 1987)
Gravitation:			Dementia:
- short-term microgravity 3 (Mano, 2005; Mano et al., 1998, 1985; Miwa et al., 1996) - long-term microgravity 1 (Fu et al., 2000; Kamiya et al., 2000a, b; Kamiya et al., 2000c; Mano et al., 1998) - mild hypergravity a–br↓ (Ueda et al., 2015; Yanagida et al., 2014)	Anger 1(Kreibig, 2010)	Post prandial hypotension a–br↓ (Kawaguchi et al., 2002; Lipsitz et al., 1986; Ryan et al., 1992)	-Alzheimers disease 1,4(Aharon-Peretz et al., 1992)	PTSD 1,4 (Buckley and Kaloupek, 2001; Fonkoue et al., 2020; Park et al., 2017; Pole, 2007)
			-Lewy body dementia 3 (Orimo et al., 2005)
			Acquired brain injury 1:
	High-Performance Sharpshooter 1↓,2↑ (Elkins et al., 2009; Saus et al., 2006; Thompson et al., 2014)	Electrical Injury 1 (Kim et al., 2014a; Lim et al., 2006)	-Concussion 1 (Boeve et al., 1998), -Hypoxic Brain Injury 1 (Adeoye et al., 2015), -Subarachnoid Hemorrhage 1 (Lambert et al., 2002; Liu et al., 2013; Naredi et al., 2000)	Anxiety 1 (Kaplan and Sadock, 1998; Pohjavaara et al., 2003)
	Emotional/Physical Thread:
	Stage Fright 1,4 (Berntson and Cacioppo, 2004; Brantigan et al., 1982; Yoshie et al., 2009) Insomnia 1,4, a–br↓ (Carter et al., 2018; Jerath et al., 2019; Tsai et al., 2015)		Dysreflexia postspinal cord injury 1(Eldahan and Rabchevsky, 2018; Hou and Rabchevsky, 2014) Diencephalic Syndromes1 (Andy and Jurko, 1983; Bullard, 1987; Carr-Locke and Millac, 1977; Fox et al., 1973; Giroud et al., 1988; Rovne and Williams, 1962; Solomon, 1973)	Malignant Serotonin Syndrome 1 (Ener et al., 2003; Francescangeli et al., 2019; Mills, 1997) or Neuroleptic Syndrome 1 (Feibel and Schiffer, 1981; Gurrera, 1999) Alcohol/Opioid drug withdrawal 1 (Hawley et al., 1994; Kienbaum et al., 2002, 1998; Linnoila et al., 1987)
Meditation:
	-Vajrayana and Hindu Tantric 1 (Amihai and Kozhevnikov, 2014; Kozhevnikov et al., 2013, 2009) -Theravada and Mahayana 2 (Ditto et al., 2006; Krygier et al., 2013; Tang et al., 2009)		Migraine1,2,3,4 (Avnon et al., 2004; Cambron et al., 2014; Gass and Glaros, 2013; Gotoh et al., 1984; Rauschel et al., 2015; Tabata et al., 2000)	Fatal Insomnia 1 (Benarroch and Stotz-Potter, 1998; Lugaresi et al., 1998)
	Sexual Function 1, 4 (Alwaal et al., 2015; Baldo and Eardley, 2006)			Anti-NMDA-R Encephalitides 1 (Byun et al., 2015; Hinson et al., 2013) Takotsubo 1 (Pelliccia et al., 2017; Raj, 2010; Wang et al., 2020)

Displayed are examples of normo—, hypo— and hyperfunctional sympathetic and parasympathetic states in disease and health. In some conditions, the underlying pathology includes more than one mechanism as indicated by multiple superscript numbers. Other than that, the literature points to a controversial current state of evidence in some conditions (in grey). Note: Although the listings of the various conditions may be noted under PNS or CNS they are usually integrated as a consequence of their activity, either as a primary or secondary effect. CNS type I includes all conditions that are primarily brain related neurological conditions and CNS type II are primarily psychiatric conditions as defined by DSM—5.

Key: ¹ sympathetic hyperfunction, ² parasympathetic hyperfunction, ³ sympathetic hypofunction, ⁴ parasympathetic hypofunction, ^a–br afferent –baroreflex function, ^pre–g pre-ganglionic, ^↑ indicates increase of function, ^↓ indicates decrease of function.

5.1. The normofunctional adaptive ANS state

The normofunctional state is defined as temporal change of sympathetic or/and parasympathetic activity in which autonomic homeostasis is maintained or restored after suspension of a stressful event. As noted in Table 1 there are a number of conditions that can be considered within this definition. While both central and peripheral processes are involved, examples noted include response to physical (e.g., exercise) or emotional or both (e.g., sex) some engage central control of modulating or controlling autonomic processes, implicating cognitive control in these cases (e.g., competitive sharpshooting). As such, individuals can slow heart rate and control breathing to enable a shot to be fired without physical effects of these processes (Açıkada et al., 2019; Tremayne and Barry, 2001). There is a direct correlation of performance and autonomic control (Caterini et al., 1995). The brain regulates the cardiovascular system by two general means: 1) feedforward regulation, often referred to as “central command,” and 2) feedback or reflex regulation (Kamiya et al., 2011; Matsukawa, 2012; Tan and Taylor, 2011; White and Raven, 2014).

An example of normofunctional state, is the stress of shooting someone in response to a threat of bodily harm. The ability to successfully point and fire a gun at a target involves a cascade of autonomic and fine motor coordination. The activation of the sympathetic nervous system is a common consequence observed in sharpshooters under stressful conditions. Despite, sympathoexcitation is an undesirable effect as it may inhibit the PFC and thus restrict the ability to make strategic decisions as well as compromise fine motor skills (Arnsten, 2009; Buschman and Miller, 2007; Lane et al., 2009; Thayer et al., 2009; Thayer and Lane, 2009). Moreover, a loss of parasympathetic activation augments the PFC inhibition and activates the amygdala (Thayer et al., 2009; Thayer and Lane, 2009). Thus, such cognitive and autonomic alterations due to stress can be directly associated with motor performance via the medial visceromotor pathway (Lane et al., 2009). Interestingly, a study confirmed that in a competitive setting individuals with less stress related sympathetic response as measured by heart rate variability changes (HRV) showed a more accurate shooting performance (Thompson et al., 2014). In line with this finding another study found that also in untrained teams the ability to identify and target armed vs. non-armed subjects depends on the physiological compliance within teams as measured by the HRV (Elkins et al., 2009). Also in police academy cadets, higher marksman and identification scores were associated with less decreased HRV during the simulation (Saus et al., 2006). Other than that, elite soldiers exhibited a more pronounced daily drop of heart rate as compared to their non-elite equivalent as well as reported less experienced stress during the maneuver, which points to an increased parasympathetic tone (Taylor et al., 2009). The underlying neurobiological mechanisms are not fully understood, however there are various explanations for the observed autonomic differences between experienced and novice shooters. For example, learned cortical plasticity changes could lower HRV. In line with such consideration is a functional MRI study investigating brain activity changes between novice and expert archers (Kim et al., 2014b). Interestingly, results of the frontal cortical lobe indicated only an activation of the superior frontal area in elites whereas the novice showed activation in the superior frontal area, inferior frontal area and ventral prefrontal cortex. These results may add to learned cortical modulation, which may explain variance of autonomic function i.e., via activity changes in infralimbic vs. prelimbic PFC areas.

However, it remains unclear how ANS function can be actively modulated as proposed in the example above as compared to selection criteria that favor those with distinct baseline physiology profiles. Thus, another explanation for a more successful shooting performance may pertain to an inherently subdued sympathetic response. For example, carriers of melanocortin-4 receptor gene mutations have been suggested a reduced central sympathetic outflow and thus reduced sympathoexcitability (Copperi et al., 2021; Kuo et al., 2004; Sayk et al., 2010). Other than that, a genetic study linked 25 independent SNPs (single nucleotide polymorphisms) in 23 loci to either heart rate recovery or increase (Verweij et al., 2018). Those genetically inherited traits certainly add up to a more or less successful shooting performance.

5.1.1. Application to integrative model

How is this example integrated into the Model shown in Fig. 8? There are hyperresponsive as well as hyporesponsive activity changes in the ANS, however such control can be learned and trained (Kox et al., 2014; McDougall et al., 2015). In the sharpshooter condition, primary control is in the CNS (#7). Areas of the brain that can consciously control heart rate and blood pressure are prefrontal and forebrain regions. Interestingly, the infralimbic and prelimbic cortices, which are subregions of the ventromedial PFC, modulate opposite cardiovagal responses which has been demonstrated by either selectively lesioning/blocking (Tavares et al., 2009) or stimulating (Napadow et al., 2008; Powell et al., 1996) the infralimbic or prelimbic subregions. These regions relay to various subcortical regions (#6). Due to their distinct function within the cascade of competitive or stressful shooting, crucial connections from the ventromedial PFC to subcortical regions may pertain to the amygdala, the nucleus accumbens and the hypothalamus (Ghashghaei and Barbas, 2002; Price and Amaral, 1981; Price et al., 1996), which regulate emotional responses and habits (Arnsten, 2009). Other than that, the dorsolateral cortex extensively relays to sensory and motor cortices and regulates adaptive responses to acute stress; namely attention, thought and action (Goldman-Rakic, 1987). However, the dorsal PFC seems to modulate autonomic output differently and thus has been rather associated with neuroendocrine responses (Radley et al., 2006). Although, function of these prefrontal subregions might differ substantially, they are closely interconnected and crucial for higher-order decision making (Arnsten, 2009). Other than that, the PFC projects to brainstem regions such as the locus coeruleus, the substantia nigra and the ventral tegmental area which regulate noradrenaline and dopamine release. High levels of catecholamines however may prevent a balanced stress response e.g. due to PFC impairment and fear response conditioning of the amygdala due to high catecholamine levels (Arnsten, 2009; Debiec and LeDoux, 2006). However, this response is another key feature which may be harnessed and trained to perfection in competitive rifle shooters as well as elite soldiers. Additionally, monoamine cells of the brainstem control sympathetic outflow to skeletal muscle, and vagal outflow to the heart (Konttinen et al., 1998).

In this condition, the ability to restrain sympathetic hyperfunction (e.g. via infralimbic vMPFC deactivation (Frysztak and Neafsey, 1991; Tavares et al., 2009)) and simultaneously promote parasympathetic hyperfunction (i.e. via prelimbic vMPFC activation (Powell et al., 1996)) in high-performance shooters is required. We suggest that within our engineering model these findings translate to a negative feedforward mechanism of sympathetic (#9c) activation and positive feedforward mechanism of parasympathetic (#9d) activation which can be conditioned via training but to a certain extent also belongs to a trait characteristic of the individual. We propose that the cortical M4 module (#7) which can be trained and conditioned is of central importance in the evaluation and execution of successful shooting tasks under stressful conditions (Arnsten, 2009; Buschman and Miller, 2007; Kox et al., 2014; Lane et al., 2009; McDougall et al., 2015; Thayer et al., 2009; Thayer and Lane, 2009).

5.2. The hypofunctional pathological ANS states

Examples of clinical states which may produce a hypofunctional ANS state are shown in Table 1. Here we use one example in which hypofunctional autonomic responses are seen – baroreceptor irradiation.

A baroreflex failure, may develop as a late sequela of head or neck irradiation. This may disrupt baroreflex regulation of cardiovagal and sympathetic outflows (Sharabi et al., 2003) and in response to stress lead to severe hypertension and elevated heart rate (Heusser et al., 2005). Mostly case reports have contributed to the current state of knowledge (Fagius et al., 1985; Heusser et al., 2005; Timmers et al., 1999). Patients often present blood pressure fluctuations (labile hypertension and hypotension). Other than that, symptoms such as orthostatic intolerance, headache, flushing, elevated norepinephrine levels, diaphoresis and palpitations were frequently reported (Aksamit et al., 1987; Robertson et al., 1993; Timmers et al., 1999).

The underlying pathomechanism of patients after neck and head irradiation points to blocking of carotid sinus mechanoreceptors, which may be the consequence of an inflammatory vessel wall remodeling leading to carotid stenosis and thus to stiffening of carotid arterial walls (Elerding et al., 1981; Monahan et al., 2001). Moreover, a decreased afferent baroreceptor input to the brain rather than loss of efferent effector function was suggested, whereas efferent parasympathetic cardiovagal and sympathetic neurocirculatory function remains intact (Sharabi et al., 2003). A physiological baroreceptor function is crucial for the blood pressure buffering capacity in hypo- and hypertensive states (Timmers et al., 2004). Therefore, such an underlying mechanism nicely fits into the heterogenic clinical observations with regard to fluctuating hypo- and hypertensive blood pressure conditions. Other than that most patients report an exacerbation of symptoms (i.e. hypertension, palpation) when experiencing emotional or physical stress (Heusser et al., 2005; Sharabi et al., 2003; Timmers et al., 2004), in combination with elevated plasma catecholamine levels this suggests a loss of baroreflex-inhibition of sympathetic nerve fiber activity (Heusser et al., 2005). These findings point to an unbuffered sympathetic outflow which may present as a sympathetic hyperfunction.

5.2.1. Application to integrative model

The pathomechanism of baroreceptor irradiation points to a causative dysfunction at the transducer/sensor level (#1), namely a loss of high pressure baroreceptors within the carotid sinuses which prevents an adequate routing via peripherocentral relays (i.e. the glossopharyngeal nerve) (#3) and the processing of signals at brainstem module M2 (#5). The first synapse of the glossopharyngeal nerve is located in the nucleus of the solitary tract (NTS) and under physiological conditions afferents of the carotid sinus modulate hypertension via (i) an inhibitory pathway that inhibits sympathetic outflow to vessels via interneurons of the caudal ventrolateral medulla that project to pacemaker neurons in the rostral ventrolateral medulla, which are further connected to muscle ergoreceptors and renal afferents. (ii) Second a connection between the NTS and the supraoptic nucleus and the paraventricular nucleus of the hypothalamus regulates arginine vasopressin release through the pituitary gland (Kaufmann et al., 2020). As a consequence, disruption of afferent input, may lead to a negative sympathetic feedback (#9a) mechanism which may result in an unbuffered positive sympathetic feedforward (#9d). Within the context of disease exacerbation (i.e. as indicated by signs such as hypertension) in stressful conditions in patients with baroreflex failure (Norcliffe-Kaufmann et al., 2010) the direct connection to subcortical brain regions (#3) that process stressful or fearful events such as the paraventricular nucleus of the hypothalamus and the central nucleus of the amygdala (Herman et al., 2008; McDougall et al., 2017; Ricardo and Koh, 1978; van der Kooy et al., 1984; Viltart et al., 2006) to the NTS may be one key element of this observation. In line with this notion a study showed that an activation of corticotropin-releasing hormone type 2 receptors in the NTS via projections from the paraventricular nucleus of the hypothalamus increased blood pressure (Wang et al., 2019). Additionally, it was demonstrated that the central nucleus of the amygdala directly projects to baroreceptive neurons in the NTS and rostral ventrolateral medulla that are particularly activated due to blood pressure changes. Moreover, an inhibitory GABAergic pathway suggested central nucleus of the amygdala-activation mediated blood pressure changes via an attenuation of baroreceptor reflexes (Saha, 2005). Although in some patients a diminished baroreflex response via vagal afferents of the aortic arche baroreceptors may still be intact (which reduce heart rate via preganglionic parasympathetic neurons of the nucl. ambiguus, that in turn activate postganglionic parasympathetic sinoatrial neurons), this implicates a vulnerable baroreflex response system that might be completely dysfunctional (i.e. shifted towards sympathetic hyperfunction) due to additional subcortical inputs. Moreover, such afferent hypofunction at sensor/transducer level, especially in stressful conditions, may facilitate a shift of homeostasis threshold towards a sympathetic hyperfunction.

5.3. The hyperfunctional pathological ANS states

We provide examples of hyperfunctional states following acquired brain injury, which we have designated as a neurological condition (Type 1), and PTSD and Panic Disorder, examples of psychiatric conditions (Type 2) (see Table 1).

5.3.1. Acquired brain injury (type 1)

An acquired brain injury can be the result of a concussion (traumatic brain injury), hypoxia or a subarrachnoidal hemorrhage. Such an event may affect the brain temporarily or permanently and lead to excessive autonomic hyperfunction (Fernandez-Ortega et al., 2012; Perkes et al., 2011). Subsequent to an acquired brain injury some patients develop autonomic hyperactivity which frequently has been reported in combination with hypersensitivity towards noxious and non-noxious stimuli (Baguley et al., 2009a, b). Interestingly a review identified two types of patients with paroxysmal autonomic hyperfunction after acquired brain injury (Perkes et al., 2011). One group predominantly experiences sympathetic hyperfunction whereas the second group exhibited a combination of sympathetic and parasympathetic hyperfunction. Interestingly, the mixed type autonomic hyperfunction predominantly occurred in congenital or non-traumatic cases of brain injury, such as agnesis of the corpus callosum (Boeve et al., 1998; Diamond et al., 2005; Do et al., 2000; Dutau et al., 1975; Hirayama et al., 1994) or tumors (Carr-Locke and Millac, 1977; Fox et al., 1973; Giroud et al., 1988; LeWitt et al., 1983; Penfield, 1929; Solomon, 1973), whereas hypoxic brain injury seemed mostly associated with the sympathetic hyperfunction type. Interestingly, although traumatic brain injury is most frequently documented with regard to paroxysmal sympathetic hyperactivity the actual incidence seems much higher in hypoxia (9/31 cases, 29 %) then traumatic brain injury (20/146 cases, 14 %) (Krach et al., 1997). This finding is corroborated by hypoxia prior to admission in 22/35 cases of patients with paroxysmal sympathetic hyperactivity following traumatic brain injury (Baguley et al., 1999). In the context of stress, it has been reported that stress may severely alter recovery after acquired brain injury and patients with persistently activated stress reactions are more vulnerable to develop other diseases such as depression (Bay et al., 2002, 2004; McEwen, 1998; Silverberg et al., 2018).

Although, the prior emphasized relevance of a hypoxic metabolic state may add to the understanding of paroxysmal sympathetic hyper-activity development the exact underlying pathomechanism is not understood. In the past different hypothesis have been put forward such as the presence of epilepsy or seizures which have been discounted due to lack of seizure activity (Boeve et al., 1998) and treatment ineffectiveness with anticonvulsants (Perkes et al., 2010). Also, different disconnection hypotheses were suggested, which assume a sympathoexcitatory state due to disinhibition of higher control brain areas from brainstem excitatory centers (Baguley et al., 2008). Alternatively, another model postulates a dysbalance of the excitatory/inhibitory ratio of inhibitory brainstem regions which leads to spinal cord mediated sympathetic hyperfunction similar to autonomic dysreflexia following spinal cord injury (Baguley, 2008).

Additionally, concussion or traumatic brain injury (TBI) produces an inflammatory response acutely, which may persist (Rathbone et al., 2015). Acute inflammatory response may be found peripherally presumably due to soft tissue injury that induces cytokines (Begum et al., 2020; Meier et al., 2020; Nitta et al., 2019). Within the brain, inflammatory molecules following TBI have been reported using PET imaging in preclinical models (Missault et al., 2019) and which is supported by microglial activation in humans (Ramlackhansingh et al., 2011). Damage of the peripheral nervous system entails inflammatory responses such as cytokine release presumably via terminals of sympathetic fibers (Jänig, 2009; Jänig and Levine, 2006) or/and the sympatho-adrenal system (Jänig et al., 2000; Khasar et al., 2003). The release of cytokines is not only involved in nociceptor sensitization it also alters autonomic function as indicated by a decrease of heart rate and blood pressure (Jänig, 2014). Moreover, the autonomic nervous system is involved in immune regulation (Elkhatib and Case, 2019; Suzuki, 2016). It is thought that both the sympathetic nervous system and the parasympathetic nervous system (Mirakaj et al., 2014) controls the resolution of inflammation (Körner et al., 2019). Conversely, cytokines affect the responsivity of the ANS (Kenney and Ganta, 2014). Thus, ANS modulation of the immune system may enhance or suppress the immune system. This notion has implication in disease states (Kin and Sanders, 2006) that may be influenced by and from the CNS (Rabin et al., 1989).

5.3.1.1. Application to integrative model.

Based on the above review of the underlying pathophysiology, we propose two different mechanisms of action based on the disinhibition/disruption hypothesis: (i) a higher order disintegration between the M3/M4 (#6/7) and the M2 (#5) module and (ii) a lower level disintegration between the M2 (#5) and M1 (#4) module. In line with both hypotheses it was shown that hypoxia may induce an upregulation of corticotropin-releasing hormone receptor mRNA levels in the nucleus tractus solitarius which contributes to hypoxia induced hypertension (Wang et al., 2018). Additionally, severe brain parenchymal or regional injuries are associated with an activation of subcortical, hypothalamic endocrine pathways, an increase of sympathoadrenal tone and intracranial pressure (Krishnamoorthy et al., 2016; Nguyen and Zaroff, 2009; Oppenheimer, 1994). In fact, the elevation of intracranial pressure might be a vicious key element as it was suggested to particularly activate the autonomic response system and thus lead to a systemic catecholamine release and hypertension (Krishnamoorthy et al., 2016; Nguyen and Zaroff, 2009). In line with the first hypothesis Kishner et al. suggested an involvement of several cortical (orbitofrontal, anterior temporal and insular cortex), subcortical (amygdala) and brainstem (periaqueductal gray, nucleus tractus solitarius, cerebellar uvula and vermis) regions that are integrated in the hypothalamic ANS response (Baguley et al., 2008). The second hypothesis specifically suggested anterior hypothalamic or medullary lesions that activate central sympathoexcitatory regions due to a diminished inhibition (Baguley et al., 2008; Boeve et al., 1998; Hackl et al., 1991; Meythaler and Stinson, 1994; Silver and Lux, 1994). Despite therapeutic consequences of each pathomechanistic etiology may differ fundamentally, both hypothesis as well as a combination seem possible as corroborated by previously suggested sites of dysfunction (Baguley et al., 2008, 2006; Baguley et al., 2007; Boeve et al., 1998; Bullard, 1987; Hackl et al., 1991; Hörtnagl et al., 1980; Klug et al., 1984; Pranzatelli et al., 1991; Rossitch and Bullard, 1988; Sandel et al., 1986; Silver and Lux, 1994; Thorley et al., 2001).

5.3.2. Panic disorder (PD, type 2)

A panic disorder is an anxiety disorder characterized by recurrent unexpected panic attacks followed by a month or more of worry about their recurrence, implications, or consequences or by a change in behavior related to the panic attacks (2020). Panic attacks occur without advance warning, despite the fact that a panic attack is a very conscious event. This suggests a merely unconscious processing of some pathomechanistic components in between panic attacks, which in turn engender failure of defense. Other than that panic may not only be an emotional thread it is often accompanied by an autonomic response such as tachycardia or hyperventilation, which are signs of ANS activation (Cohen et al., 2000). Similar to inter-event related unconsciously processed PD mechanisms a large portion of the ANS processes are unconscious. The combination of two unconsciously processed conditions and the failure to separate the emotional from the autonomic response may lead to the aggravation of a panic disorder. This notion is further corroborated by the PD’s reciprocal nature as the trigger of a panic attack may not only be of emotional nature. Thus, an independent activation of the sympathetic nervous system may also trigger a panic attack. This is reinforced by a study demonstrating that a catheter ablation in patients with paroxysmal supraventricular tachycardia cured panic disorders (Frommeyer et al., 2013). Thus, tachycardia may also be the cause and not only the consequence of a panic disorder. Based on this evidence, we propose that either emotional or autonomic dysfunctions may alter (ie. lower) the threshold necessary to trigger a panic attack.

In line with these clinical observations, the emotional component may be processed via the lateral amygdala which relays to the basal nucleus and further to the nucleus accumbens and the central nucleus of the amygdala (AMYce) (Elman et al., 2018; Neugebauer, 2015; Simons et al., 2014). The fact that the central nucleus of the amygdala is predominantly involved in the activation of the sympathetic response system may foster the importance of shared subcortical circuits for the maintenance of a panic disorder. This was further corroborated by a study exploring the heart rate variability (HRV) which is a measure of the cardiovascular autonomic response function in patients with panic disorders (PD) as well as PTSD, at baseline and after experimental recall of psychological stress. At baseline both patient groups showed a higher heart rate as well as low frequency components of the HRV, which indicate a higher sympathetic tone. Interestingly, after the experimental recall of either a circumstance of severe panic in PD or a traumatic experience in PTSD only in the PD group responded with an increase of heart rate and low frequency components, whereas the PTSD group failed to respond (Cohen et al., 2000). These results, suggest different underlying neuropathological mechanisms with regard to ANS response and integration.

5.3.2.1. Application to integrative model.

The basis for dysregulation in PD may be the unconscious activation or alteration of shared subcortical circuits at the transducer level; ie. within the M3 (#6) module. Key to these considerations are the connections of the central nucleus of the amygdala, which maintain direct connections to key autonomic regions such as the nucleus tractus solitarius, caudal ventrolateral medulla, motor nucleus of the vagus nerve, etc. (Bienkowski and Rinaman, 2013; Hopkins and Holstege, 1978; McDougall et al., 2017; van der Kooy et al., 1984; Viltart et al., 2006; Zhang et al., 2003); key stress processing regions such as the paraventricular hypothalamic nucleus and supraoptic hypothalamic nucleus (Paredes et al., 2000; Pittman et al., 1981; Silverman et al., 1981) as well as infralimbic and prelimbic PFC areas (Bienkowski and Rinaman, 2013; Hoover and Vertes, 2007; Vertes, 2004).

Moreover, based on the observations made by Cohen et al., 2000 we suggest that in PDs the servo feedback system is intact, whereas in PTSD this feedback mechanisms may be disrupted. Additionally, sympathoexcitation in PD may alter an adequate stress response due to the previously described fear response conditioning of the amygdala due to high catecholamine levels (Arnsten, 2009; Debiec and LeDoux, 2006). We suggest that such a vicious cycle of sympathoexcitation may sensitize the feedback threshold which may lead to a positive sympathetic feedback (#9b) and thus facilitate a panic attack more easily.

5.3.3. Post-Traumatic stress disorder (PTSD, type 2)

Post-Traumatic Stress Disorder (PTSD, Type 2): is a stressor-related condition that may develop after extremely traumatic events, such as combat, crime, an accident or natural disaster. People with PTSD may relive the event via intrusive memories, flashbacks and nightmares; avoid anything that reminds them of the trauma; and thus develop anxious feelings, that disrupt their lives (Kazdin, 2000). Heightened responsiveness to aversive stimuli including reminders of the trauma is one of the most consistent findings in psychophysiology (Blechert et al., 2007; Orr et al., 2002) and functional brain imaging PTSD studies (Elman et al., 2018; Rauch et al., 2006). However, the pathophysiologic significance of these abnormal findings is unclear.

One interpretation posits that PTSD is associated with enhanced conditionability (Blechert et al., 2007; Pitman et al., 2012) that subserves stronger conditioned response and/or failure to extinguish the response to the cues associated with the traumatic experience. For example, studies reported skin conductance response (SCR) evidence consistent with stronger conditioning and delayed extinction following aversive electrical stimulus (Blechert et al., 2007; Maples-Keller et al., 2019). These deficits along with other autonomic, sensory, and cognitive abnormalities (Orr et al., 2002) found in PTSD clinical studies would be expected to underlie the phenomenon of “re-experiencing” (Blechert et al., 2007; Milad et al., 2005), which is encoded in the DSM-5 PTSD “B” diagnostic criteria viz., re-experiencing the traumatic event in thoughts, dreams, and in intrusive reminiscences, hallucinations and flashbacks in conjunction with psychological and physiological distress caused by trauma cues exposure (American Psychiatric Association, 2013). Another interpretation of the exaggerated fear and anxiety symptomatology findings relates to a non-associative sensitization, namely escalating neurochemical and behavioral effects of any stimulus repeatedly administered in an aversive context. This is a multifactorial process that can be dissected into distinct, but interacting neurochemical, neuroanatomical and functional systems, including (1) mesolimbic dopaminergic pathways, (2) norepinephrine and CRF neurotransmission within the extended amygdala structures, namely the central nucleus of the amygdala (Shepard et al., 2000, 2003) and the bed nucleus of the stria terminalis (Rasmussen et al., 2000; Southwick et al., 1999, 1993) and (3) the hypothalamic-pituitary-adrenal axis (Elman and Lukas, 2005; Piazza et al., 1996).

The ANS and HPA axis are functionally integrated in a number of ways. For example, both are activated by inflammation (Karrow, 2006), glucoprivic (Elman et al., 1998; Elman et al., 2004), and psychosocial stress (Kemeny and Schedlowski, 2007). As noted above, there are important interactions between the ANS and the immune system. Segregation of the interaction of the PNS and the SNS in their interactions on the HPA axis have been reported. The PNS interacts with the HPA axis, for example - stimulation of the HPA axis, but not suppression, is associated with decreased vagal tone (Agorastos et al., 2019). For the SNS, interactions with the HPA axis may counterbalance their responses to alter a stress response, for example following stroke (Mracsko et al., 2014). Importantly sex hormones may influence the interaction between these two systems (Barel et al., 2018; Stephens et al., 2016). Such integration and balance of these systems has implications in healthy states in terms of perceived stress (Rotenberg and McGrath, 2016) and in disease states, for example fibromyalgia where inflammation produces an excessive ANS response (viz heart rate) to injections of IL-6 (Torpy et al., 2000), in childhood anxiety disorders (Dieleman et al., 2015) or in addictive states (Fox et al., 2009).

A way to assess the relative contribution of the above mechanisms is to compare the pattern of fMRI signal changes in the amygdala and medial prefrontal cortex under paired and unpaired conditioned stimuli (i.e., CS + and CS-, respectively) during early and late extinction and during recovery of the extinct memories (Ji and Maren, 2007). If PTSD patients increase amygdala signal in response to CS+, but not to CS-, amygdala overrecruitement probably does not occur secondary to sensitization and the case for primary conditionability may be supported. On the other hand, enhanced amygdala activation during both CS + and CS− conditions in PTSD patients would suggest both conditionability and sensitization mechanisms.

5.3.3.1. Application to integrative model.

Based on the underlying evidence we suggest alternative models. The first theory (i.e. conditioning theory) pertains to enhanced conditioning at the CNS M4 (#7) module and suggests a disrupted negative sympathetic feedback loop (#9a) or an enhanced positive sympathetic feedforward activity (#9d) (Blechert et al., 2007; Maples-Keller et al., 2019; Milad et al., 2005; Orr et al., 2002; Pitman et al., 2012). This hypothesis is corroborated by the connections of cortical regions such as the PFC and the insula (Bienkowski and Rinaman, 2013; Hoover and Vertes, 2007; Vertes, 2004) towards the amygdala which as previously described projects to key autonomic structures such as the nucleus tractus solitarius, caudal ventrolateral medulla, motor nucleus of the vagus nerve, etc. (Bienkowski and Rinaman, 2013; Hopkins and Holstege, 1978; McDougall et al., 2017; van der Kooy et al., 1984; Viltart et al., 2006; Zhang et al., 2003). The second theory (i.e. non-associative sensitization theory) suggests a non-associative sensitization that may mainly function via the subcortical module M3 (#6) (Elman and Lukas, 2005; Piazza et al., 1996; Rasmussen et al., 2000; Shepard et al., 2000, 2003; Southwick et al., 1999, 1993). The basis for this consideration may pertain to the previously described high level catecholamine alteration that may prevent a balanced stress response e.g. due to PFC impairment and fear response conditioning of the amygdala due to high catecholamine levels (Arnsten, 2009; Debiec and LeDoux, 2006).

6. Conclusions and future directions

Demonstration that somatic and neuropsychiatric syndromes share common neurobiological foundation in the form of failure and/or exhaustion of ANS homeostatic mechanisms leading to mounting of allostatic load could have important implications.

First, it may point to the initial stage of a co-morbidity cycle relatively free from potential confounds associated with end stage consequences of somatic psychopathology that might exaggerate organ dysfunction and stress-related negative emotional states (Katz et al., 2015). Second, it could help in understanding and predicting relapse as physical symptoms could trigger emotional angst i.e., psychache in remitted subjects (Elman et al., 2013). In fact, such cross-sensitization may be bi-directional, that is to say, post-traumatic headache triggered by the re-experienced PTSD- or anxiety symptoms (Lambru et al., 2020; Scott et al., 2020) or the catheter ablation in patients with paroxysmal supraventricular tachycardia which cured panic disorders (Frommeyer et al., 2013).

Remarkable advances of central and peripheral ANS research have uncovered fundamental mechanisms at the cellular and molecular levels. Nonetheless we are still searching for theories that will integrate the prevailing models of the corresponding neuropsychiatric- and comorbid syndromes. One approach reflected in the Research Domain Criteria (Nelson et al., 2020) is to divide these multidimensional constructs into whole body domains comprised of central and peripheral components alike (e.g. (Elman et al., 2020)). Each domain can be then studied separately, which may still be a daunting task given the complexity of the systems involved. For example, the pain and analgesia construct encompasses overlapping social (e.g., interpersonal) vs. physical (sensation) and anticipatory vs. consummatory components (Borsook et al., 2013; Elman and Borsook, 2016) pertaining to respective aspects of the perceptual (Baliki and Apkarian, 2015) and motivational (Elman et al., 2011) processing; their discussion is beyond the scope of the present paper that inevitably entailed a degree of oversimplification as it was intentionally limited to key issues of direct relevance to the presented model.

In conclusion, various ANS mechanisms by which somatic and mental conditions may affect and derail each other have been put forward but they do not yet seem to fully explain the complexity of this link. As it becoming clearer that neuropsychiatric disorders may be driven by the allostatic load ensuing due to somatic dysfunction (Elman et al., 2020), an ensuing logical query that we began to address here is whether psychiatric patients e.g., with panic disorder even in the absence of any other risk factors are vulnerable for the development of somatic ailments e.g., coronary artery disease (Katerndahl, 2008) by the reason of similar somatic (Martinez et al., 2001) and neuropathological (Fernandes et al., 2020) alterations inherent in their illness. This concept while strongly supported by theoretical considerations presented here now needs to gain confirmation in clinical research with evolving novel diagnostic, therapeutic and preventive strategies.

Abbreviations:

M₄

cortex

M₃

subcortex

M₂

brainstem

M₁

spinal cord/periphery

A1/C1

A1 and caudal C1 catecholamine neurons, VLM

ACCv

ventral anterior cingulate cortex

ACCd

dorsal anterior cingulate cortex

AMY

amygdala

AMYb

basal nucleus of the amygdala

AMYbl

basolateral nucleus of the amygdala

AMYbm

basomedial nucleus of the amygdala

AMYce

central nucleus of the amygdala

AMYco

cortical nucleus of the amygdala

AMYl

lateral nucleus of the amygdala

AMYm

medial nucleus of the amygdala

AMYp

posterior nucleus of the amygdala

AMYPIR

amygdalo-piriform transition zone

AP

area postrema

BN

basal nucleus

BNST

bed nucleus of the striae terminalis

CAN

caudate nucleus

CLA

claustrum

CVLM

caudal ventrolateral medulla

DMX

dorsal motor nucleus of the X nerve

DBbca

diagonal band of Broca

DB

nucleus of the diagonal band

DLPFC

dorsolateral prefrontal cortex

EdWest

Edinger-Westphal nucleus

EPNBF

endopiriform nucleus of the basal forebrain

EP

entopeduncular nucleus

ENRC

entorhinal cortex

ECRC

ectorhinal cortex

FAC

facial nucleus

FCa

frontal association cortex

FPCmd

frontal pole cortex, medial division

GP

globus pallidus

HAB

habenula

Hd

dorsal horn

Hv

ventral horn

HIPPO

hippocampus

HIPPOah

hippocampus, ammon’s horn

HIPPOpara

parahippocampus

HYPOl

lateral hypothalamus

HYPOlm

lateral mammillary nucleus of the hypothalamus

HYPOa

anterior hypothalamus

HYPOac

arcuate nucleus of the hypothalamus

HYPOvm

ventromedial nucleus of the hypothalamus

HYPOdm

dorsomedial nucleus of the hypothalamus

HYPOlpo

lateral preoptic nucleus of the hypothalamus

HYPOmpo

medial preoptic nucleus of the hypothalamus

HYPOmnpo

median preoptic nucleus of the hypothalamus

HYPOp

posterior hypothalamus

HYPOpf

perifornical area of the hypothalamus

HYPOpvn

paraventricular nucleus of the hypothalamus

HYPOvlpo

ventrolateral preoptic nucleus of the hypothalamus

HYPOso

supraoptic hypothalamus

HYPOsm

supramammillary nucleus of the hypothalamus

II

oculomotor nucleus

ilC

infralimbic cortex

ilMPFC

infralimbic medial prefrontal cortex

IML

intermediolateral nucleus

INC

nucleus incertus

INS

insular cortex

INSant

anterior insular cortex

INSpost

posterior insular cortex

IP

interpeduncular nucleus

IPAC

interstitial nucleus of the posterior limb of the anterior commissure

KF

Kölliker-Fuse nucleus

LFC

lateral frontal cortex

LS

lateral septum

MBN

magnocellular basal nucleus

ME

median eminence

MOC

motor cortex

MPFC

medial prefrontal cortex

MS

medial septum

NA

nucleus ambiguus

NAC

nucleus accumbens

NTS

nucleus of the solitary tract

OB

olfactory bulb

OC

orbital cortex

OCm

medial orbital cortex

OCv

ventral orbital cortex

OFC

orbitofrontal cortex

OT

olfactory tubercle

PAG

periaqueductal gray

PAGvl

ventrolateral periaqueductal gray

PAm

medial preoptic area

PBN

parabrachial complex

PBl

lateral parabrachial nucleus

PCa

parietal association cortex

PFC

prefrontal cortex

PG

pineal gland

preLC

pre-locus coeruleus

plC

prelimbic cortex

plMPFC

prelimbic medial prefrontal cortex

PMC

primary motor cortex

PRC

perirhinal cortex

PSSC

primary somatosensory cortex

preMOC

premotor cortex

preBötz

preBötzinger complex

PSTN

parasubthalamic nucleus

PPL

posterior pituitary lobe

PPTA

postpiriform transition area

RAPHc

caudal raphe nucleus

RAPHcl

central linear nucleus raphe

RAPHd

dorsal raphe nucleus

RAPHm

median raphe nucleus

RAPHmes

mesencephalic raphe

RAPHp

nucleus raphe pallidus

RF

reticular formation

RFc

caudal reticular formation

RFr

rostral reticular formation

RFpo

pontine reticular formation

RNcp

caudal pontine reticular nucleus

RNim

intermediate reticular nucleus

RNgc

gigantocellular reticular nucleus

RNopo

oral pontine reticular nucleus

RNpvc

parvocellular reticular nucleus

RRA

retrorubral area

RSC

retrosplenial cortex

RTRA

retrotrapezoid nucleus

RVLM

rostral ventrolateral medulla

SAm

medial septal area

SAL

salivatory nucleus

SALsup

superior salivatory nucleus

SCa

anterior sigmoidal cortex

SCS

supracommissural septum

SFO

subfornical organ

SGC

subgenual cingulatea area

SNC

sensory cortices

SN

substantia nigra

SNpc

substantia nigra pars compacta

SNpr

substantia nigra pars reticularis

SNpl

substantia nigra pars lateralis

SOC

supraoptic crest

STN

subthalamic nucleus

SUBI

substantia innominata

SUCv

subiculum ventral portion

SYMPPN

sympathetic preganglionic neuron

TTd

taenia tecta, dorsal part

Tcl

central lateral thalamic nucleus

Tcm

central medial thalamic nucleus

Td

dorsal thalamic nuclei

Tiad

interanteriomedial thalamic nucleus

Til

intralaminar thalamic nuclei

Timd

intermediodorsal thalamic nucleus

Tmd

mediodorsal thalamic nucleus

TEGld

laterodorsal tegmental nucleus

TEGpp

pedunculopontine tegmental nucleus

TEGrm

rostromedial tegmental nucleus

TEGv

ventral tegmental area

TCa

temporal association cortex

Treu

reuniens nucleus of the thalamus

Trho

rhomboid nucleus of the thalamus

Tpt

paratenial nucleus of the thalamus

Tpev

periventricular nucleus of the thalamus

Tpav

paraventricular nucleus of the thalamus

Tvpm

ventral posteromedial nucleus of the thalamus

Tvmb

ventromedial basal nucleus of the thalamus

TUBINF

tuberoinfundibular area

TUBMAM

tuberomamillary nucleus

Vmot

trigeminal motor nucleus

vMPFC

ventral medial prefrontal cortex

VP

ventral pallidum

Vsen

trigeminal sensory nucleus

Vsup

supratrigeminal nucleus

Vsp

spinal trigeminal nucleus

Vspdm

spinal trigeminal nucleus, dorsomedial part

Vspc

spinal trigeminal nucleus, caudal part

Vspip

spinal trigeminal nucleus, interpolar part

VIIIco

cochlearis nucleus

XII

hypoglossal nucleus

XIIprep

nucleus prepositus hypoglossi

ZI

zona incerta

Appendix A. The Peer Review Overview and Supplementary data

The Peer Review Overview and Supplementary data associated with this article can be found in the online version, at doi:https://doi.org/10.1016/j.pneurobio.2022.102218.