Our laboratory has recently demonstrated that VNS can be used to induce a rapid, general improvement of thalamic sensory processing (Figure 1). This is a continuation of our team’s studies investigating the effects of the LC-NE system on thalamocortical circuitry (Rodenkirch et al., 2019), a critical stage for sensory processing and perception (Saalmann and Kastner, 2011; Stanley et al., 2012; Wang et al., 2012; Millard et al., 2013; Kelly et al., 2014; Ollerenshaw et al., 2014; Wimmer et al., 2015; Rikhye et al., 2018). These studies found that direct activation of the LC-NE system (electrical or optogenetic), in a continuous tonic fashion, optimized intrathalamic dynamics for sensory processing. Specifically, tonic LC stimulation (continuous, 5 Hz, 60 μA, 500 μs biphasic pulses) increased the efficiency and rate of sensory-related information transmitted by thalamocortical neurons (Rodenkirch et al., 2019). Further, the observed NE-enhancement of sensory processing resulted in a significant improvement in perceptual sensitivity for rats tasked with discriminating between whisker stimuli of different frequencies. Through pharmacological manipulation it was determined that tonic LC activation improved thalamic sensory processing because a steady increase in NE concentration precludes priming, and in turn activation, of thalamic T-type calcium channels. When active, T-type calcium channels introduced a non-linear bursting response that degraded transmission of detailed sensory information.

Vagus nerve stimulation has been shown to activate the LC-NE system (Hulsey et al., 2017) and is accessible in a non-invasive manner, unlike the LC deep in the brainstem. Therefore, our team next investigated whether tonic VNS would drive similar rapid beneficial effects on sensory processing. Through testing the effects of multiple patterns of VNS on sensory processing, the beneficial effect was found to be highly transient (i.e., benefit begins to dissipate within seconds of ceasing VNS) (Rodenkirch and Wang, 2020). For example, duty-cycled VNS (30 s on/60 s off duty cycle, 30 Hz, 500 μs biphasic pulses) enhanced tactile sensory processing during the on cycle, but this enhancement rapidly dissipated during the off cycle, suggesting that cycling VNS on and off creates fluctuations in sensory processing that would likely be sub-optimal for discrimination. This indicated that an uninterrupted pattern is required to produce a stable benefit. Indeed, continuous tonic VNS (continuous, 30 Hz, 500 μs biphasic pulses) induced a steady enhancement of sensory processing similar to that observed with direct tonic LC stimulation. This immediate enhancement of sensory processing during continuous, tonic VNS was found to be reliably present across recorded neurons. As each recorded neuron encoded for a unique kinetic feature of the whisker stimuli, this suggests the tonic VNS modulation provided a general enhancement of sensory processing regardless of stimulus input. This effect is distinct relative to the selective facilitation of responses to a specific sensory stimulus found after repeatedly pairing VNS bursts with that sensory stimulus.

Further, testing of various tonic VNS current levels and frequencies showed the beneficial effect of tonic VNS on sensory processing increased with intensity and frequency (10 vs. 30 Hz, 0.4 vs. 1 and 1.6 mA) and did not exhibit the inverted U-shape function of effect strength that has been observed with other types of VNS modulation (Morrison et al., 2019) (at least within the parameter ranges tested).

Other research groups working with human subjects have published findings that suggest VNS has immediate beneficial effects on auditory processing. One study in humans who had been receiving chronic VNS (via implanted cuffs as a treatment for epilepsy), found VNS enhanced performance on a standard auditory oddball task when compared to performance after their VNS device was turned off (De Taeye et al., 2014). Specifically, during VNS (7 s on/18 s off duty cycle, 20–30 Hz, 0.75–3 mA, 250–500 μs pulses) both accuracy and response time were improved for participants tasked with responding to low frequency target audio tones while ignoring high frequency non-target tones. This same study analyzed auditory event-related potentials (AERP), measured via EEG, and found that during VNS, AERP amplitude was also increased. However, the effect on AERP was only significant in individuals whose epilepsy symptoms had positively responded to VNS treatment. A separate study investigating transcutaneous auricular vagus nerve stimulation (taVNS) (30 s on/30 s off duty cycle, 25 Hz, 250 μs pulses) in healthy adults found similar results. Specifically, taVNS increased the strength of AERPs during an oddball auditory task (Rufener et al., 2018). As this study used low frequency tones as non-targets and high frequency tones as targets, a reversal of the prior discussed oddball auditory task, taken together they suggest immediate VNS modulation of auditory response is not specific to low or high frequency audio tones. Another study delivering continuous taVNS (25 Hz, 500 μs biphasic pulses) to healthy adults analyzed the neural response to auditory tones using magnetoencephalography (MEG) instead of EEG and found taVNS altered synchrony of brain activity (Hyvarinen et al., 2015). Further, recent studies using fMRI to monitor neural activity have shown taVNS rapidly affects auditory processing pathways. When taVNS (25 Hz, 0.1 to 1.8 mA, 500 μs monophasic pulses) was delivered to male adults with chronic tinnitus, fMRI recordings exhibited altered activity of multiple brain regions involved with auditory processing (Yakunina et al., 2018). More recently, analysis of fMRI data from human subjects receiving taVNS indicated increased activity in the thalamus and auditory cortex (Peng et al., 2018), suggesting VNS rapidly modulates central auditory sensory processing in humans.

These findings in humans are further supported by multiple electrophysiological and behavioral work in animals that found VNS rapidly affects the response properties of neurons of the auditory pathway. In isoflurane-anesthetized rats, the responses of neurons along the auditory pathway were compared with and without VNS delivered via an implanted VNS cuff (30 s on/5 min off duty cycle, 10 Hz, 0.5 mA, 130 μs pulses). The baseline condition was recorded without any ongoing VNS. The VNS condition consisted of discontinuous duty-cycled VNS where auditory testing was performed only during the off periods of the VNS duty cycle. Here they found duty-cycled VNS weakened stimulus-specific adaptation in the cortex but not the thalamus (Shiramatsu et al., 2016), suggesting VNS may modulate thalamocortical transmission but not earlier stages of the auditory pathway. Further work by the same group, using the same paradigm, found VNS predominantly increased the amplitudes of auditory-evoked potentials in the sensory cortex (Takahashi et al., 2020).

The immediate effects of VNS on olfactory processing had been demonstrated as early as the 1980s. Specifically, a study in rats found that a single pulse of VNS from an implanted cuff (0.8–1.5 mA, 200 μs monophasic pulses) reliably evoked firing in the homolateral olfactory bulb (HOB) (Garcia-Diaz et al., 1984). Further evidence that VNS affects olfactory processing was found in more recent studies that used positron emission tomography (PET) to analyze the effects of VNS in awake rats. A PET scan conducted during the time period when the VNS cuff was switched on for the first time (30 s on/5 min off duty cycle, 30 Hz, 1.5 mA, 500 μs pulses) found VNS induced a significant increase in glucose metabolism in both olfactory bulbs (Dedeurwaerdere et al., 2005). However, another study in humans with implanted VNS cuffs for treatment of depression found that whether VNS (30 s on/5 min off duty cycle, 20 Hz, 1.25 mA) was on or off had no effect on subjects’ ability to discriminate or detect olfactory stimuli (Sperling et al., 2011). Yet that same study did find that VNS significantly increased the intensity of the taste of sweet and bitter, suggesting that VNS may rapidly affect gustatory processing as well.

The LC is the primary source of NE in the forebrain (Sara, 2009). The LC exhibits constant tonic firing (1–5 Hz) that regulates brain state (e.g., arousal) as well as intermediate phasic burst spiking events (2–5 spikes at 10–20 Hz per burst) that occur in response to salient sensory stimuli as well as when decisions or responses are made (Devilbiss and Waterhouse, 2011). These two firing modes have been shown to produce distinctly different modulations of the response properties of sensory neurons (Devilbiss and Waterhouse, 2011). The LC innervates multiple regions along the sensory pathway, including the sensory thalamus and cortex (Morrison and Foote, 1986; Simpson et al., 1997).

There is a large body of evidence showing that the LC-NE system modulates sensory processing and perceptual learning (Manunta and Edeline, 1997; Ego-Stengel et al., 2002; Hirata et al., 2006; Doucette et al., 2007; Martins and Froemke, 2015; McBurney-Lin et al., 2019; Waterhouse and Navarra, 2019). Moreover, it is well documented that activation of the LC-NE system immediately modulates the response of sensory neurons. In vitro, NE has a depolarizing effect on auditory and visual thalamic relay neurons that coincides with a suppression of burst spiking (McCormick and Prince, 1988). This likely occurs because NE depolarization prevents the extended hyperpolarized periods needed to prime the calcium T-type channels responsible for bursts (Sherman, 2001a). In vivo, tonic LC activation has been found to reduce spontaneous activity of the somatosensory thalamus, while facilitating sensory evoked activity, resulting in an increase in signal to noise ratio (Hirata et al., 2006). Our team has shown how tonic LC-NE activation enhances the accuracy of encoded stimuli in the somatosensory thalamus by reducing the fluctuating influence of the calcium T-type channels responsible for bursting. Within the cortex, the LC-NE system can cause either facilitation or inhibition with resulting effect specific to the sensory modality, cell, and stimulation pattern (Devilbiss and Waterhouse, 2004; Videen et al., 1984; Sato et al., 1989; Vazey et al., 2018).

Vagus nerve stimulation’s ability to activate the LC-NE system has long been hypothesized to underlie, in part, the clinical benefits of VNS (Slater and Wang, 2021). VNS is thought to activate the LC via the vagus nerve’s afferent projections to the nucleus tractus solitarius (NTS) (Van Bockstaele et al., 1999; Ruffoli et al., 2011). The NTS then sends an excitatory signal to the LC, likely via the nucleus paragigantocellularis (Ennis and Aston-Jones, 1988; Reyes and Van Bockstaele, 2006). Indeed, multiple studies have confirmed VNS readily activates the LC-NE system in both animals and humans. In rats, VNS delivered via an implanted cuff has been shown to increase the activity of LC neurons as confirmed by electrophysiological recordings under halothane (Groves et al., 2005), chloral hydrate (Dorr and Debonnel, 2006), equithesin (Manta et al., 2009a), and ketamine (Hulsey et al., 2017) as well as by immunohistochemical biomarkers of short-term neuronal activation (Cunningham et al., 2008). Similarly, multiple studies have found that microdialysis samples taken from rats receiving VNS exhibited increased NE concentration in the primary hippocampus (Raedt et al., 2011), basolateral amygdala (Hassert et al., 2004), and cortex (Roosevelt et al., 2006; Follesa et al., 2007; Manta et al., 2013). Finally, the findings in animals seem to be conserved in humans, as fMRI data from a study of adult males with tinnitus indicated taVNS activates the NTS and LC (Yakunina et al., 2018). However, variations in VNS parameters may affect how reliably VNS drives the LC-NE system, as one study measuring NE concentration in the CSF of patients receiving VNS as a treatment for depression failed to detect a significant change (Carpenter et al., 2004).

In addition to direct evidence VNS activates the LC-NE system, many effects of VNS are blocked if the LC-NE system is impaired through either LC lesion or adrenergic receptor blockers. For example, the anticonvulsive effect of VNS is abrogated when hippocampal adrenergic receptors are blocked (Krahl et al., 1998; Raedt et al., 2011). Further, VNS enhancement of perforant path-CA3 synaptic transmission is blocked by either electrical lesions of the LC or an adrenergic receptor antagonist (timolol) (Shen et al., 2012). The antidepressant-like effects of VNS in rats, as measured by feeding and swim tests, have been shown to be blocked by lesion of noradrenergic neurons (Furmaga et al., 2011; Grimonprez et al., 2015). Immunotoxin depletion of norepinephrine was also found to prevent VNS-driven enhancement of motor cortex neuroplasticity (Hulsey et al., 2019).

Cholinergic nuclei of the basal forebrain (BF) project to the sensory processing regions of the thalamus (Kolmac and Mitrofanis, 1999) and cortex (Ballinger et al., 2016; Jimenez-Martin et al., 2021). Additionally, cholinergic nuclei of the pontomesencephalic area, including the laterodorsal tegmental nucleus (LDT) and pedunculopontine tegmental nucleus (PPT), are a major source of ACh to the thalamus (Schofield et al., 2011; Huerta-Ocampo et al., 2020). There are two distinct neuron populations of the BF that differentiate in exhibiting either a tonic (10–15 Hz) or a bursting (2–6 spikes/burst with bursting events occurring at 0.3–2 Hz) firing pattern (Nuñez, 1996) which influences arousal and attention. The response timing of both types of BF neurons is influenced by sensory stimuli (Laszlovszky et al., 2020) and linked with novelty, salience, and surprise (Zhang et al., 2019).

Extensive work has shown the cholinergic system strongly influences both sensory processing and perceptual learning across multiple sensory modalities (Murphy and Sillito, 1991; Kilgard and Merzenich, 1998; Verdier and Dykes, 2001; Linster and Cleland, 2002; Bentley et al., 2004; Wilson et al., 2004; Furey et al., 2008; Herrero et al., 2008; Pinto et al., 2013; Zhan et al., 2013; Rothermel et al., 2014; Kim et al., 2016; Gratton et al., 2017). Like the noradrenergic system, it is well documented that activation of the cholinergic systems has immediate effects on sensory processing. ACh applied in vitro to neurons of the thalamic reticular nucleus, a subthalamic region involved in sensory processing, causes hyperpolarization and induces burst spiking (McCormick and Prince, 1986), likely due to extended hyperpolarized periods priming the calcium T-type channels responsible for burst spiking (Sherman, 2001a). ACh applied to thalamic neurons of the primary visual and auditory pathways was found to increase firing rate (Sillito et al., 1983; McCormick and Prince, 1987), although a hyperpolarization effect has been observed in thalamic neurons of the secondary (non-lemniscal) auditory pathway (Mooney et al., 2004). Cholinergic modulation of the sensory cortex can cause either facilitation or inhibition with the resulting effect specific to the sensory modality, cell, and stimulation pattern (Donoghue and Carroll, 1987; Metherate and Weinberger, 1989; Metherate and Ashe, 1991; Jimenez-Martin et al., 2021). In the visual cortex, BF stimulation has been shown to enhance accurate encoding by inducing decorrelation and increased reliability (Goard and Dan, 2009).

It has long been hypothesized that VNS activates the BF–ACh system (Detari et al., 1983). VNS innervates the NTS (Ruffoli et al., 2011) and projections from the NTS activate the BF (Martin et al., 2022) in addition to the NTS projections that activate the LC (Ennis and Aston-Jones, 1988; Van Bockstaele et al., 1999; Reyes and Van Bockstaele, 2006). The LC also projects to the BF (Berridge et al., 2003), suggesting VNS activates the BF both directly through the NTS as well as indirectly through the LC. Indeed, two separate studies investigating the potential of VNS for inducing neuroprotection from cerebral ischemia found that VNS enhanced protein levels of the nicotinic acetylcholine receptor alpha7 subunit (a7nAchR) in the ischemic penumbra (Jiang et al., 2014; Lu et al., 2017). Recently, researchers performed in vivo calcium imaging of the auditory cortex and found VNS evoked activity of cholinergic axons innervating the region (Mridha et al., 2021). Further, they found the intensity of the evoked activity covaried with VNS intensity. In addition to this direct evidence that VNS rapidly activates the cholinergic system, multiple studies have shown ACh modulation of sensory pathways is a critical component underlying the plasticity effect induced by repeatedly pairing a burst of VNS with a sensory stimulus. For example, the effects of VNS on sensory processing in the auditory cortex were found to be blocked by a muscarinic antagonist (Nichols et al., 2011). Further, lesioning the NB in rats was shown to abrogate the well-documented ability of VNS pulses repeatedly paired with a movement to enhance motor cortex plasticity (Hulsey et al., 2016).

The dorsal raphe nucleus (DRN) is a major source of serotonin (5-HT) to the forebrain (Jacobs and Fornal, 1999). Neurons of the DRN consistently exhibit a continuous slow tonic firing rate (1–2 Hz) with little variation in inter-spike-interval (Trulson and Jacobs, 1979; Mlinar et al., 2016). Response of the DRN is related to both reward and punishment (Ranade and Mainen, 2009; Li et al., 2016; Ren et al., 2018) as well as linked to sensory input (Rasmussen et al., 1984; Waterhouse et al., 2004). The DRN innervates both cortical and subcortical regions of the sensory processing pathways (Kirifides et al., 2001). There is also a large body of work suggesting DRN activity modulates sensory processing and perception (Hurley and Pollak, 1999; Kähkönen et al., 2002; Dacks et al., 2009; Hurley and Hall, 2011; Jaber et al., 2014; Kapoor et al., 2016; Seillier et al., 2017). 5-HT has been shown to have instant effects on neurons of the sensory pathways. For example, 5-HT has been shown to cause excitation of thalamic perigeniculate and reticular nucleus neurons (McCormick and Wang, 1991; Funke and Eysel, 1993). In the inferior colliculus, an auditory region of the midbrain, 5-HT was found to modulate responses in both a cell and auditory stimulus specific manner (Hurley and Pollak, 1999). In the primary visual and auditory relay neurons of the visual and auditory pathways, 5-HT has been shown to have an inhibitory effect (Marks et al., 1987; Kayama et al., 1989; Monckton and McCormick, 2002). Additionally, activation of the DRN has been found to increase signal to noise ratio of the olfactory cortex (Lottem et al., 2016).

Vagus nerve stimulation may activate the DRN indirectly by first activating the LC which then projects to the DRN (Kim et al., 2004). This hypothesis is supported by a study in rats anesthetized with sodium pentobarbital that found VNS increased DRN neurons’ firing rates, but this causal relationship was lost once the LC was lesioned (Manta et al., 2009a). Multiple studies have also shown that VNS increases DRN firing rate as measured via extracellular electrophysiological recordings (Dorr and Debonnel, 2006; Manta et al., 2009b). However, one study found only a subset of VNS patterns they tested increased DRN activity suggesting VNS activation of the DRN may be dependent on VNS parameters (Manta et al., 2012). In a follow-up work, the same group performed in vivo microdialysis in rats following chronic duty-cycled VNS and found increased 5-HT concentration in the DRN but not the hippocampus nor prefrontal cortex (PFC) (Manta et al., 2013). In contrast to these studies supporting VNS’ ability to activate the DRN, another study analyzing microdialysis measurements in different brain regions of rats reported that neither vagotomy or chronic unilateral VNS had an effect on 5-HT levels in the ventral tegmental area (VTA), nucleus accumbens (NAc), PFC, and striatum (Ziomber et al., 2012). These conflicting findings could potentially be related to the fact that electrical stimulation was delivered to an abdominal branch of the vagus nerve in this study. Further suggesting a more complex interplay between the VNS and DRN, a study analyzing immunohistochemical biomarkers of both short-term and long-term neuronal activation suggests chronic VNS does not induce DRN activation until stimulation has occurred across multiple days (Cunningham et al., 2008).

In addition to direct evidence that VNS increases activity of the serotonergic system, functionality of serotonergic neurons has been shown to be critical for multiple documented effects of VNS. For example, the earlier-mentioned study on the antidepressant-like effects of VNS in rats, which used feeding and swim tests as indexes of depression, found the beneficial effects of VNS were also precluded by administration of a neurotoxin for serotonergic neurons (Furmaga et al., 2011). Additionally, a separate study found immunotoxin depletion of serotonin prevented the well-researched ability of repeatedly pairing a VNS burst with a movement to enhance motor cortex neuroplasticity (Hulsey et al., 2019).

The VTA and Substantia Nigra pars Compacta (SNc) are primary sources of dopamine (DA) to the forebrain (Poulin et al., 2018) and, respectively, they modulate cognition and movement (Mercuri et al., 1992). The VTA has been shown to innervate the sensory cortices (Hosp et al., 2019). The VTA exhibits both tonic (1–8 Hz) and burst firing (2–5 spike bursts with bursting events occurring at 0.1–1 Hz) with firing rates varying across cell types (Kiyatkin and Rebec, 1998; Hyland et al., 2002; Lodge and Grace, 2006). Tonic firing rate likely modulates brain state (e.g., motivation and arousal) and bursting events likely encode for salient stimuli (e.g., reward and sensory stimuli) (Dahan et al., 2007). Although the body of work investigating the effects of DA on sensory processing is limited, there is evidence it rapidly modulates sensory processing and response (Ungless, 2004; Govindaiah et al., 2010; Woolrych et al., 2021).

Although previous work demonstrated the LC projects to the VTA (Mejías-Aponte et al., 2009), many studies also suggest VNS effects on DA circuitry may be dependent on other factors besides VNS directly increasing VTA firing rates. For example, one study that performed in vivo microdialysis of rats following chronic duty-cycled VNS found an increase in DA in the PFC and NAc but a decrease in VTA neurons’ firing rates as measured with electrophysiological recordings (Manta et al., 2013). A lack of VNS-induced changes in VTA firing and bursting rates was also reported in a separate study (Perez et al., 2014). Studies analyzing brain sections from rats that received chronic VNS have also reported varied results. One such study found decreased DA levels in the VTA, NAc, PFC, and striatum (Ziomber et al., 2012); however, to properly interpret these results it should be mentioned that electrical stimulation was delivered to an abdominal branch of the vagus nerve in this study. Two other studies performing a similar analysis found VNS induced changes to the elemental composition of dopamine-related brain structures (Szczerbowska-Boruchowska et al., 2012) and to the lipids and proteins within the VTA, NAc, SNc, striatum, dorsal motor nucleus of vagus, and the motor cortex (Surowka et al., 2015). A more recent study in awake rats found optogenetic VNS, which carries no risk of unintentional activation of surrounding nerves, increased the firing rate of dopaminergic VTA neurons as measured via in vivo imaging (Fernandes et al., 2020). This same study also found lesioning the hepatic branch of the vagus nerve abrogated the increase in VTA neuron activity usually observed following ingestion.