The PFC is located in the anterior frontal lobe of the brain, anterior to the motor cortex. It is often referred to as the brain’s command and control center, which receives and integrates diverse information from all regions, both internal and external, through tightly connected neural circuits (Stevens, 2010). The PFC is involved in higher-order cognitive functions, and its neural circuits possess a unique ability to enable PFC neurons to co-activate and sustain information even when the stimulating stimuli are no longer present in the immediate environment. However, when confronted with interference, the PFC engages its unique “working memory” (WM) mechanism, activating PFC neurons to direct overt responses (movements), to covert responses (attention), and to suppress inappropriate reactions (Chafee and Heilbronner, 2022). In addition to this, the left anterior cingulate cortex/medial prefrontal cortex is responsible for inhibitory control, and the bilateral middle frontal gyrus/precentral gyrus is responsible for modulating dorsal and ventral attention networks, while playing a key role in the flexible regulation of endogenous and extrinsic attention (Wang et al., 2024). Thus, the honeycomb-like neuronal network specific to the PFC can use information from WM to manage stability and ensure content coherence in pursuit of attentional goals, execute plans, and organize appropriate actions (Arnsten, 2011; Yan and Rein, 2022).

The core symptoms of ADHD stem from disrupted circuits of attention regulation and action inhibition. Functional abnormalities in the neural networks of the PFC in ADHD patients (reduced PFC gray matter, diminished PFC connectivity, and altered PFC function) all impair cognitive control of behavior and attention (Arnsten, 2010). The associations of functional circuits involved in the pathophysiology of ADHD are shown in Figure 1. Thus, the neuronal pathways associated with the PFC in ADHD appear to have a central role, primarily manifesting in the following aspects.

First, dysfunction of the prefrontal-subcortical pathway is recognized as one of the core pathophysiological mechanisms of ADHD (Halperin and Schulz, 2006). Structural and functional imaging studies have demonstrated significant abnormalities in neuronal connectivity, signaling, and metabolic activity in the PFC of ADHD patients (Cubillo et al., 2010; Del Campo et al., 2011). For example, a structural magnetic resonance imaging (MRI) study of ADHD adults demonstrated that these people have volume differences in brain regions involved in attention and executive control, primarily in the overall cortical gray matter, with significantly smaller PFC volumes (Seidman et al., 2006). Additionally, diffusion tensor imaging reveals a decreased anatomical functional connectivity of white matter tracts linking the PFC to other brain regions in ADHD patients (Curatolo et al., 2010). Meanwhile, resting-state fMRI analyses showed a more decentralized pattern of connectivity in the default mode and frontal control network in ADHD children, mainly in terms of increased functional connectivity in the right inferior frontal gyrus and bilateral medial frontal lobes, which may reflect a reduced or delayed functional segregation of the PFC (Bos et al., 2017). Therefore, the inability of the PFC to effectively regulate the functions of posterior cortical and subcortical structures in ADHD patients may underlie their abnormalities in attention, emotional regulation, and behavioral control.

Second, the developmental process of the PFC is also closely related to the onset of ADHD (Halperin and Schulz, 2006). Moreover, the development of the PFC involves neuron generation, migration, differentiation, and synapse formation. The PFC is the last to develop in the central nervous system (CNS) and matures at the latest in individual development. However, developmental impairment of PFC in ADHD patients, including structural and functional abnormalities, is evident early in life (Sullivan and Brake, 2003). In addition, functional near-infrared spectroscopy (fNIRS) has shown that children with ADHD exhibit lower PFC activation in performing functional tasks compared to their peers who have developed normally throughout childhood (Su et al., 2023). Shaw and colleagues (Shaw et al., 2007, 2012) utilized MRI to assess the developmental trajectory of brain morphology in ADHD patients, revealing delayed cortical thickness development in both pediatric and adult patients compared to controls. Notably, this delay was evident as cortical thinning, particularly in the frontal, parietal, temporo-parietal, and occipital regions, highlighting the slowed or impaired development of the right lateral PFC.

In addition, interactions between the PFC and other brain regions also play an important role in the pathogenesis of ADHD. Extensive connections between the PFC and neural circuits in the parietal cortex, basal ganglia, cerebellum, hippocampus, and corpus callosum have been identified through functional imaging studies, which together participate in ADHD-associated functional networks (Curatolo et al., 2009). However, in ADHD patients, the interactions between these brain regions may be abnormal. Significantly reduced connectivity in the bilateral lateral PFC, anterior cingulate cortex, superior parietal lobule, and cerebellum has been reported in adults with ADHD compared with healthy controls (Wolf et al., 2009). ADHD patients exhibit reduced interregional functional connectivity between the right inferior frontal lobe, frontal striatum, and frontal-parietal neural networks in a cessation task (Cubillo et al., 2010). In summary, there is a close correlation between the PFC and ADHD. Therefore, understanding the molecular and cellular regulation that modulates PFC function is crucial for exploring therapeutic approaches to ADHD and developing targeted medications, ultimately leading to better patient outcomes.

The PFC network is capable of guiding action in the absence of bottom-up external stimuli and through the mutual excitation of a network of interconnected pyramidal neurons on dendritic spines to store goals/rules (Arnsten, 2011; Vogel et al., 2022). However, the activity of PFC networks is largely dependent on a correct neurochemical milieu, where arousal pathways coordinate cognitive states with environmental events (Arnsten, 2011).

The most representative arousal pathway, the NE pathway, has neurons that originate in the LC within the brainstem and project terminally to a wide range of brain regions, including the PFC. The operating pathway of the NE system in the brain is shown in Figure 2A. Anatomically, the LC provides the sole NE neural input to the PFC, and the PFC is one of the few cortical sites that provide cortical information to the LC. The NE release levels may rapidly alter the strength of PFC network connections to coordinate cognitive states with physiological demands (Yan and Rein, 2022). Among other things, NE is also released in the PFC in response to our arousal state, with an inverted U-shaped dose effect on PFC cognitive function (Levy, 2009). In this inverted U-shaped effect, either too little NE (such as depletion or fatigue) or too much NE (like uncontrollable stress) impairs PFC function, while intermediate levels of NE, released when subjects are alert or interested, reinforce and shape inputs to optimize PFC function (Arnsten, 2011). Thus, the level of NE released in the PFC can harmonize brain arousal states and PFC function.

However, the effects of NE on arousal, mood, and behavior are mediated through interactions with a wide range of receptors. Based on the pharmacological profile of adrenergic receptors, which include those responsive to NE, they can be classified into two major groups: α and β receptors. These are divided into three major subgroups each: α1 (α1A, α1B, α1D), α2 (α2A, α2B, α2C), and β (β1, β2, β3), due to differences in coupled G proteins, signaling, and effector systems (Perez, 2020).

Norepinephrine has the highest affinity for α2 receptors, followed by α1 receptors, and the lowest affinity for β receptors (Ramos and Arnsten, 2007). Among them, α2A receptors play an important role in PFC cognitive function. Furthermore, robust synaptic connections enable the PFC to more effectively enhance WM, engage in the regulation of attentional control, and modulate behavioral inhibition. The efficacy of PFC network connectivity relies on NE stimulation of α2A receptors located on the spines of PFC pyramidal cells (Arnsten, 2011). α2A receptors are located on dendritic spines in the vicinity of the ion channel and control the effects of synaptic inputs on the spines. When there is no α2A receptor stimulation, cyclic adenosine monophosphate (cAMP) levels are high, potassium channels (HCN channels) open, nearby synaptic inputs are diverted, and incoming information escapes, weakening synaptic connections (Arnsten and Jin, 2012). Conversely, when α2A receptors are stimulated by NE or guanfacine (α2A receptor agonists), activation of Gi leads to a decrease in cAMP levels, which in turn reduces the opening of HCN channels and enhances the efficacy of neural network inputs and synaptic excitatory inputs, thereby promoting neuronal firing activity and PFC function (Arnsten and Jin, 2012; Joyce et al., 2024). The mechanism of action of the NE system at the synapses of PFC pyramidal neurons is illustrated in Figure 2B.

The VN is the 10th pair of cranial nerves, the longest and most widely distributed pair of cerebral nerves. This nerve is a mixed nerve composed of 20% efferent fibers and 80% afferent fibers, encompassing both sensory and motor nerve fibers (Butt et al., 2020). The sensory neurons in this nerve protrude centrally into the brainstem, terminating at the NST (nucleus of the solitary tract) and the nucleus of the spinal tract of the trigeminal nerve. Additionally, the motor neurons originate from the nucleus ambiguus of the medulla oblongata and the dorsal nucleus of the VN. Among them, the auricular branch of the VN is mainly distributed in the tragus, the cymba conche, and the cavum conche. It crosses the jugular foramen via the auricular branch of the VN, then enters the medulla oblongata, and subsequently ascends through the trigeminal nucleus of the spinal cord to connect with the NST (Ellrich, 2019; Kaniusas et al., 2019a). Yakunina et al. (2017) conducted a study using fMRI and showed that the cymba conche has the strongest activating effect on the vagal pathway and is considered to be the optimal therapeutic site for taVNS. The conduction diagram of the NE system in the brain after taVNS stimulation is shown in Figure 3. Furthermore, the left and right VN in humans provide parasympathetic visceromotor innervation to the sinoatrial node and the atrioventricular node, respectively (Kaniusas et al., 2019b). Consequently, to prevent intracardiac conduction abnormalities from inducing arrhythmias, clinical practice often favors the application of taVNS on the left ear (Chen et al., 2023b; Zhang et al., 2023; Zhou et al., 2023). Therefore, the activation of the central nervous system through stimulation of VN sensory fibers distributed in the periphery serves as the theoretical basis for the treatment of neurological disorders using taVNS.

The NST is considered to be the relay station between the afferent fibers of the VN and the associated nuclei in the brain. It receives the majority of afferent fibers from the VN and a minor portion of information from other peripheral regions, subsequently integrating and projecting this sensory information to higher centers (Wang L. et al., 2023). The projection fibers of NST form three main pathways: (i) they form an autonomic feedback loop; (ii) they project directly to the medullary reticular formation; and (iii) they project to the pontine parabrachial nucleus, the LC, and other structures in the brain. Of these, the NST receives approximately 95% of the VN input and projects directly to the LC via either excitatory or inhibitory pathways. The LC plays a significant role as a mediator in the reticular superior activation system. It projects to the cerebral cortex through several key structures, including the nucleus of the median eminence, the amygdala, the hypothalamus, the orbitofrontal cortex, and the cingulate gyrus (Maness et al., 2022). Consequently, NE released from this region exerts an excitatory effect on the cerebral cortex. Furthermore, an elevated NE concentration within the LC activates the PFC and a broad network of cortical and subcortical circuitry projections throughout the brain (Ludwig et al., 2021). This activation leads to extensive remodeling of brain networks in areas densely innervated by the LC, facilitating enhanced task performance, such as vigilance and attention (Briand et al., 2020).

The fMRI scans showed that taVNS promoted sequential activation of the NST and LC, as well as brain regions closely related to cognitive functions with which these structures have fiber contact (Borgmann et al., 2021). Additionally, taVNS produced more activation in brainstem regions, including the LC. In an animal model, a 2-week VNS intervention significantly increased extracellular NE levels in the PFC and hippocampus and enhanced tonic activation of postsynaptic α2A in pyramidal neurons (Manta et al., 2013). And taVNS significantly increased c-Fos protein expression in the NST and LC nuclei (Yue et al., 2023). Taken together, given that both NST and LC are key nodes for the regulatory effects of taVNS, taVNS may be involved in the modulation of cognitive functions such as attention, arousal, behavior, and memory through this pathway.

Relevant research indicates that the activation of the NE system is determined by the combination of stimulus intensity and pulse width, rather than solely by intensity alone (Hulsey et al., 2017). In previous clinical studies, a stimulus intensity of 0.5 mA, frequencies of 25 Hz/20 Hz, and pulse widths ranging from 0.25 to 1 ms have been most commonly employed as stimulation parameters for taVNS (Sun et al., 2023; Yang et al., 2023; Pan et al., 2024). However, it remains uncertain whether the activation of the NST-LC-NE system via taVNS requires the adjustment of these parameters to achieve optimal stimulation effects. Notably, Sclocco et al. (2020) compared four different stimulation frequencies (2, 10, 25, and 100 Hz) and found that taVNS at 100 Hz elicited the highest fMRI response within the NST-LC system. However, further clinical research is needed to determine whether this frequency provides the greatest therapeutic benefit for ADHD and to assess tolerability among pediatric patients.

Attention is a bottom-up process driven by external stimuli (Aniwattanapong et al., 2022). Satterfield et al. (1974) proposed a model of low arousal in the central nervous system. The essence of ADHD was primarily explored at the level of physiological arousal, which is the necessary process for maintaining normal physiological activities and alternates with the diurnal cycle (Satterfield et al., 1974). The low arousal theory suggests that children with ADHD have lower levels of arousal than neurotypical children and that their higher activity levels and stimulus-seeking behaviors may be an attempt to increase their level of arousal, thus leading them to show excessive movement to stay awake (Brennan and Arnsten, 2008). Furthermore, psychology and neuroscience suggest that vigilance and cortical arousal are important components of attentional regulation. Vigilance refers to the ability to maintain performance and stay alert, while cortical arousal represents the level of spontaneous activation in the cerebral cortex. These factors are responsible for maintaining optimal brain arousal and behavioral performance (Sherman and Usrey, 2021; Rueda et al., 2023).

The physiological basis of attention is an extensive neural network, and the neural pathways associated with this network can be further divided into three subnetworks that perform arousal, orienting, and executive control functions (Schindler et al., 2024). Using fMRI to explore the neurophysiological basis of attention, regions in the extended frontoparietal network, including the PFC, anterior cingulate cortex, anterior insular cortex, basal ganglia, and thalamus, were found to be activated in a variety of attentional functions, whereas hypoactivation of neural pathways associated with the aforementioned networks can all contribute to attention deficits (Pezze et al., 2014; Aniwattanapong et al., 2022). Increasing evidence suggests that taVNS may modulate cortical activity, excitability, and plasticity through LC-NE activation of neuromodulatory systems across a wide range of projections, including those in attentional networks (Collins et al., 2021; Chen et al., 2023a).

Furthermore, NE can control the state of cortical networks and can influence information processing (Hays et al., 2013). Salivary alpha-amylase (sAA), pupillary dilation (PD), and alpha cortical oscillations have been identified as NE markers and seem to support the existence of a close link between the taVNS and the activity of the LC-NE system (D’Agostini et al., 2023; Wienke et al., 2023). A recent large pooled analysis confirmed that taVNS stimulates NE release by analyzing 10 taVNS studies and showing that vagal activation by taVNS increased sAA release compared to sham stimulation (Giraudier et al., 2022). Wienke et al. (2023) demonstrated this link in a study of 29 subjects following separate true/false taVNS, showing that phasic taVNS significantly increased pupil dilation and improved performance during the emotional Stroop task; in magnetoencephalography, taVNS increased frontal midline theta and alpha power, reflecting a shift toward more internally oriented attention. Chen et al. (2023a) revealed that the reduction in spontaneous alpha oscillations modulated by taVNS may be associated with a broader release of inhibition and higher levels of arousal in the cortex, particularly in the sensorimotor cortex, which may further increase exogenous stimulus detection and improve executive functioning, by administering taVNS interventions to 30 subjects and no interventions to another 30 subjects, respectively. This suggests that taVNS reliably induces arousal reflected in pupillary and EEG markers beyond the effects of somatosensory stimulation, thus supporting the idea that endogenous NE-mediated arousal is dependent on the potency of exogenous stimulation with taVNS (Lerman et al., 2022).

Hence, it can be inferred that, although further clinical and basic experimental validation is required based on these findings, taVNS appears to ameliorate attentional control in ADHD patients by activating the LC-NE pathway, modulating NE release, and thereby influencing neuroregulatory systems closely associated with changes in arousal states.

Executive function (EF) deficits serve as an endophenotype of ADHD symptoms, and the hyperactive, impulsive, and inattentive behaviors exhibited by ADHD patients are often due to impaired EF (Li et al., 2023). EF encompasses three core components: inhibitory control, WM, and cognitive flexibility. Moreover, the results of a recent meta-analysis suggest that taVNS can enhance cognition, especially executive function (Ridgewell et al., 2021).