Stress can be broadly defined as any situation associated with augmented energy consumption that is beyond expected or physiological range. Alternatively, stress can be explained through the central concept of homeostasis. Physiological and biochemical processes in the organism maintain equilibrium; an ideal steady state that is hardly achievable. Environmental factors, internal or external stimuli, continually disrupt this equilibrium. Therefore, an ideal harmony or a homeostatic point can be considered as an attractor and the organism’s current state as dynamic fluctuations around this attractor. Factors pushing the organism’s functions to diverge too far from homeostasis are defined as stressors (Robert and Labat-Robert 2015; Chrousos 2009; Chrousos and Gold 1992).

To maintain homeostasis, organisms have developed an integrated response managed by the stress system. The system is equipped with central and peripheral neuroendocrine mechanisms to cope with challenges that threaten or are perceived as threatening for homeostasis (Selye 1936). Importantly, not all stressful conditions had adverse effects on human health. On the contrary, ‘eustress’ represents those states of stress that are associated with pleasant feelings, enhanced human growth and development at the emotional and intellectual levels. On the other hand, distress consists of stressful conditions that trigger pathologic alterations. Stress response depends not only on catecholamines originating from the adrenal medulla and the systemic sympathetic nervous system, but also on the adrenal cortex-derived glucocorticoids that exert a strong anti-inflammatory effect (Nicolaides et al. 2015; Szabo et al. 2012).

Nowadays, it is recognized that activation of the hypothalamic–pituitary–adrenal (HPA) axis in response to stress is highly dependent on the characteristics of individual stressors (Herman et al. 2016). Information about pathological stressors representing a straightforward threat to homeostasis, e.g., inflammation, hypoxia, hypoglycemia, or blood loss, is transmitted directly to the paraventricular hypothalamic nucleus (PVN) whose neurons release corticotropin-releasing hormone (CRH) via monosynaptic fibres originating in sensory organs (Herman et al. 2003). Therefore, blood volume or oxygenation changes are communicated via baro- and chemoreceptors to the solitary tract nucleus (Accorsi-Mendonca and Machado 2013) that sends noradrenergic projections to the PVN, triggering an immediate HPA axis response (Ulrich-Lai and Herman 2009).

In contrast, psychological stimuli are typically transmitted to the PVN via more complex circuitry including one or more limbic structures. These structures process polymodal sensory information in the context of available mnemonic information regarding a potential threat, and generate an anticipatory response to manage the real or perceived threat to health or well-being (Herman et al. 2003). Inputs from multiple limbic regions converge to provide direct projections to the PVN (Radley and Sawchenko 2011), so that the information from stress-excitatory and stress-inhibitory parts of limbic structures is reconfigured and streamlined aiming at the optimization of a net stress response (Ulrich-Lai and Herman 2009). Different types of psychological stimuli may affect different limbic stress-regulatory pathways to an extent. Therefore, stressors has been classified by numerous investigators as physical, psychological or emotional (Jacobson 2005; Dayas et al. 2001), interoceptive vs. exteroceptive (Sawchenko et al. 2000), systemic vs. processive (Herman and Cullinan 1997), and more recently as reactive vs. anticipatory (Herman et al. 2003).

Components of the central stress system consist of the following: 1/ parvocellular neurons of the PVN, which secrete CRH; 2/ PVN neurons that secrete arginine vasopressin (AVP); 3/ CRH neurons that form the paragigantocellular and parabrachial nuclei of the medulla and the LC; and 4/ other neuronal groups in the medulla and pons forming the LC/NE system mostly secreting NE (Nicolaides et al. 2015). Stimulation of CRH neurons activates the LC/NE system and vice versa. In addition, the stress system communicates with the mesocortical and mesolimbic dopaminergic reward systems, receiving inhibitory afferent input from either system (Charmandari et al. 2003; Chrousos and Gold 1992). The stress system is linked to the central nucleus of the amygdala, a structure of the limbic system involved in the generation of fear and anger; both forming a positive regulatory feedback loop upon activation. CRH actions also modulate the transmission of sensory information by the thalamic and cortical neurons. In particular, during stress of high intensity, CRH acts at the LC to suppress neuronal processing of weak afferent stimuli within sensory systems (Devilbiss et al. 2012).

In the awake state, as long as homeostasis is maintained, LC exhibits low tonic discharge of 1–2 Hz (Aston-Jones and Bloom 1981a, b). Stressful events shift LC activity toward a high tonic mode of firing (3–8 Hz). The CRH is believed to drive the high tonic state while simultaneously decreasing phasic firing events, thus decreasing the signal-to-noise ratio (Page and Abercrombie 1999; Valentino and Foote 1988; Curtis et al. 1997). In normal conditions, PFC exerts an inhibitory control over the amygdala, HPA axis, and the LC-NE system. After activation by a highly stressful event, amygdala inhibits the PFC, which relieves both HPA axis and LC-NE system from PFC-controlled inhibition. The HPA and LC-NE activate the amygdala, which further mitigates the PFC. Taking together, PFC is inhibited under highly stressful stimulation, while the HPA axis, LC-NE, and amygdala create a positive feedback loop amplifying the effect and potentially leading to pathologic conditions (Gold and Chrousos 2002). It has been recently shown that release of amygdalar CRH into the LC promotes anxiety and aversive behavior; the processes in which LC and its afferent circuitry are critical for encoding and producing stress-induced anxiety (McCall et al. 2015). These findings are strengthened by the experiments in which the loss of hypothalamic CRH markedly reduces anxiety behavior (Zhang et al. 2016).

The current knowledge on the LC-PFC axis is related to the research focused on the psychostimulant-enhanced PFC cognitive performance. Methylphenidate and amphetamine, the psychostimulants used in treatment of attention deficit hyperactivity disorder (ADHD), potently block NE and dopamine (DA) reuptake in the brain and consequently increase the extracellular level of these neurotransmitters (Arnsten et al. 1996; Kuczenski et al. 1995; Kuczenski and Segal 1992). DA and NE exert an inverted-U action on the PFC dependent working memory, PFC neuronal signaling, and in the case of DA on D1 receptors (Vijayraghavan et al. 2007). At low, clinically-relevant doses of these psychostimulants, elevated concentration of NE and DA improves the PFC cognitive function in both ADHD and healthy subjects. In contrast, systemically administered four to eight-fold higher doses impair working memory and cognition (Spencer et al. 2012). Importantly, low-doses of these psychostimulants exert very modest effect on extracellular catecholamines in the subcortical regions associated with motor activation (e.g., nucleus accumbens) and arousal (e.g., medial septum). Therefore, the prototypical behavior-activating and arousal-enhancing effects of high-doses of psychostimulants are absent (Spencer et al. 2015).

Postsynaptic α₂-adrenoceptors have high-affinity to NE and are engaged at lower rates in NE release (Arnsten 2000). Consistently, a beneficial working memory effect of low-dose methylphenidate and atomoxetine is blocked with systemic treatment with α₂ and D₁ antagonists (Gamo et al. 2010; Arnsten and Dudley 2005). Likewise, the cognition-enhancing effects of intra-PFC infusion of methylphenidate are abolished by co-infusion of α₂ and D₁ antagonists (Berridge and Spencer 2016). Interestingly, methylphenidate injected intra-PFC does not impair cognitive performance, even at concentrations 16-fold and 32-fold higher than a clinically relevant dose (Spencer et al. 2012). It seems that PFC pathways remain operational at values close to zero of the descending limb of the inverted U-shaped curve. The mechanisms of cognition-impairing actions of high doses of systemically administered psychostimulants should thus be located outside of PFC, with HPA axis being the most likely candidate (Barsegyan et al. 2010).

Lower affinity α₁-adrenoceptors, responsible for impairment of working memory and PFC neuronal signaling, become involved at higher rates of NE release (Arnsten 2000). Consequently, α₁-adrenoceptors are activated by acute restraint stress (Alves et al. 2014), predatory stress (Rajbhandari et al. 2015), and maternal separation (Coccurello et al. 2014). Stimulation of α₁-adrenoceptors reinforces the stress response, while their blockade diminishes the HPA activation (Yang et al. 2012). Likewise, behavior-activating high doses of psychostimulants activate the HPA axis, and impair spatial working memory through activation of glucocorticoid receptors present in the PFC (Barsegyan et al. 2010). In line with these findings, exposure to prolonged stress, resulting in PFC hyperactivity and a loss of function in the nucleus accumbens, leads to behavioral inflexibility. An impairment in set-shifting may be responsible for the disengagement of attention in individuals who are exposed to traumatic events (Piao et al. 2016). Furthermore, high intensity stress profoundly suppresses the activity of neurons that are strongly tuned to key task events and, at the same time, it activates neurons displaying relatively weak task-related tuning. The end result is a profound collapse in the fidelity of goal-related coding across the PFC output neurons (Devilbiss et al. 2016).

Modulation of the LC-NE and HPA axes plays an essential role in cognitive function and task performance modifications. Nonetheless, a number of issues remain that need to be resolved. In particular, the interference of the LC-NE and HPA axes with other systems responsible for attention and alertness, such as serotoninergic, histaminergic, or cholinergic pathways needs to be further clarified. Likewise, the role of microglia and interactions among inflammatory agents, NE, and CRH merit detailed attention and research.

This review focused almost exclusively on rodent and macaque research. Recent studies have revealed that nonluminance-mediated changes in pupil diameter might provide a moment-to-moment index that reflects the LC-mediated coordination of noradrenergic neuronal activity. The exact mechanisms underlying the relation between pupil changes and NE-driven neural and behavioral phenomena still remain conjectural. Nonetheless, an appealing possibility arises of future non-invasive investigations in humans based on the pupil distention.

We briefly summarized the current knowledge on the neuroendocrine mechanisms that may account for the relation between stress and cognition. This knowledge is in line with the notion that moderate stress might be quite stimulating, while the effect of exposure to intense, and prolonged hostile psychological environment is devastating regarding cognitive abilities and human general well-being. It is worth mentioning that this widely accepted popular notion is increasingly justified in light of recent research discoveries.