Memories are created through a highly selective filter: While mundane experiences easily fade away, events laden with emotional salience are etched deeply into our remembrance (Dolan, 2002; LaBar and Cabeza, 2006; Wang and Bukuan, 2015; Rouhani et al., 2023). Emotions are typically elicited during events that are critical to our survival and can therefore functionally be conceived as tags emphasizing the motivational significance of stimuli and stimulus associations (Bradley et al., 2001b). Thus, memory prioritization for emotional material is highly adaptive, as it ensures that significant events (e.g., receiving rewards or facing threats) can be recalled to guide our future behavior (e.g., approach or escape) (McIntyre et al., 2012; Dunsmoor and Kroes, 2019; Szeska et al., 2022). Given its relevance, it may therefore be somewhat surprising that the memory-enhancing effect of emotion has been recognized for a long time (James, 1890), yet the underlying mechanisms of action have only begun to be uncovered in the second half of the last century (Easterbrook, 1959).

More than 50 years later, it is now well established that memory consolidation for emotional experiences is prioritized as they are associated with a profound increase in arousal, entailing increased attentional and perceptual processing (Bradley et al., 2001a; LaBar and Cabeza, 2006; Rouhani et al., 2023). For instance, pictures that have been rated as highly arousing are overall better remembered than low arousing ones, as indicated by increased free recall (Bradley et al., 1992; Hamann et al., 1997) and high-confidence (recollection-based) recognition memory (Dolcos et al., 2005, 2020; Weymar et al., 2009). Even non-emotional pictures of tools or animals have found to be preferably consolidated if they have been associated with physiologically arousing stimuli, e.g., an aversive electric shock or a monetary reward (Dunsmoor et al., 2015a; Patil et al., 2017). Accordingly, dedicated strategies that elevate arousal around the time of stimulus encoding, e.g., by stressful tasks, have found to additionally foster memory enhancement for emotional material (Nielson et al., 1996; Nater et al., 2007; Schwabe et al., 2008; Weymar et al., 2012), while strategies that decrease arousal, e.g., by relaxing music, have shown to attenuate such effect (Rickard et al., 2012; for a review see McGaugh, 2018).

Extensive research in animals and humans unveiled that arousal-based memory enhancement is hinged upon the adrenal glands’ immediate adrenaline and delayed glucocorticoids release (Roozendaal et al., 2009; McIntyre et al., 2012; McGaugh, 2018). For instance, aversive emotionally arousing pictures (e.g., of a snake or a gun pointing towards the participant; see Figure 1) lead to a profound increase in peripheral adrenergic activation, as indexed by endocrine markers such as salivary alpha amylase (sAA), and cortisol release, both of which positively covarying with enhanced memory retention (Abercrombie et al., 2003; Codispoti et al., 2003; Van Stegeren et al., 2006; Segal and Cahill, 2009). Highlighting the particular role of adrenergic activation in memory consolidation, depleted levels of peripheral adrenaline–e.g., due to adrenalectomy–in contrast impair memory performance for emotional stimuli (Borrell et al., 1983). Providing even stronger mechanistic evidence, small doses of exogeneous adrenaline foster memory enhancement for emotional material and even reverse the effects of adrenalectomy (Borrell et al., 1983; Cahill and Alkire, 2003), while beta-blockers that attenuate adrenergic transmission prevent memory enhancement from unfolding (Van Stegeren et al., 1998; for electrocortical evidence see e.g., Weymar et al., 2010).

Pharmacological and imaging studies have demonstrated, that increased levels of peripheral adrenaline invoke such memory enhancement by impacting on the neural transmission in the amygdala–a central hub organizing the establishment of emotional memories via its projections to the hippocampus and cortical regions (Dolcos et al., 2004; Kensinger, 2004; Ritchey et al., 2008; McIntyre et al., 2012; McGaugh, 2018). Accordingly, the activity of this region during encoding of emotionally arousing material increases and positively correlates with memory performance (Canli et al., 2000; Figure 1), while bilateral amygdala lesions have found to prevent memory enhancement for emotionally salient stimuli (Adolphs et al., 1997; Phelps et al., 1998). At this, increases in amygdala activity are indeed strikingly concomitant to increases in adrenergic activity (Van Stegeren et al., 2006; van Stegeren et al., 2007), suggesting that the release of adrenaline might stimulate this region to ultimately invoke memory enhancement: In line with this view, direct infusions of adrenaline into the amygdala foster, while direct infusions of beta-blockers attenuated emotional memory enhancement in animals (Liang et al., 1995).

However, as peripheral adrenaline is unable to cross the blood-brain barrier (Weil-Malherbe et al., 1959), a neural axis–the vagus nerve-has been presumed to convey information about elevated adrenergic levels from the body to the brain, thereby indirectly increasing neural transmission in memory-relevant brain regions. The vagus nerve, as a cranial nerve consisting of 80% afferent fibers, had long been considered a major autonomic communication route by which the brain receives information about the state of the inner body (Foley and DuBois, 1937; Berthoud and Neuhuber, 2000; McIntyre et al., 2012). Importantly, vagal afferents indeed innervate the adrenal glands and are highly responsive to the release of peripheral adrenaline due to a high number of beta-adrenergic receptors (Coupland et al., 1989; Niijima, 1992; Miyashita and Williams, 2006; Figure 1). Thus, the release of peripheral adrenaline is able to activate vagal afferents, which then project to the nucleus of the solitary tract (NTS) in the brainstem where adrenergic activation is finally synapsed onto the main hub of noradrenaline in the brain: the locus coeruleus (LC) (Miyashita and Williams, 2004; McIntyre et al., 2012; McGaugh, 2018; Figure 1). By way of the vagus nerve, a peripheral release of adrenaline is thus translated into increased noradrenergic release via the LC-brainstem arousal system. This system ultimately projects to memory-relevant brain areas including the basolateral amygdala (AMY), hippocampus (HC) and cortex (e.g., the medial prefrontal cortex; mPFC), increasing noradrenergic transmission and, thus, promoting plasticity in these areas to eventually foster memory establishment (McIntyre et al., 2012; Mather et al., 2016; McGaugh, 2018; Figure 1). Accordingly, the vagus nerve was presumed to constitute the first relay of a neural circuit mediating the memory enhancing effects of emotional arousal. Indeed, animal research supports this concept by showing that an invasive stimulation of the vagus nerve (iVNS) elicits firing in the NTS and LC and eventually increases noradrenergic transmission in the amygdala (Hassert et al., 2004; Hulsey et al., 2017; Cooper et al., 2021). Consequently, iVNS in fact promotes the memory formation in animals (Clark et al., 1998) while such memory enhancement is prohibited from unfolding, when the LC is optogenetically silenced (Calderon-Williams et al., 2024).

While a major role of the vagus nerve in the formation of emotional memories is thus supported by the animal model (see for a review Olsen et al., 2023), a comparable role has long been elusive in humans. In 1999, however, Clark and colleagues demonstrated that iVNS in fact leads to an enhancement in word recognition memory (Clark et al., 1999). Despite such promising preliminary evidence, though, subsequent stimulation studies that further explored a vagal route to memory in humans remained scarce due to the requirement of invasive surgery. This changed at the beginning of the millennium, when Peuker and Filler (2002) discovered that vagal afferents reach the body’s surface at the Cymba Conchae of the human auricle – a skin area that is exclusively innervated by the auricular vagal branch (Figure 1), allowing non-invasive vagal stimulation (transcutaneous auricular vagus nerve stimulation; taVNS) with minimal side effects (Giraudier et al., 2025). TaVNS leads to activation in the vagal afferent network important for memory formation including the NTS, LC, amygdala, hippocampus and mPFC, and also results in an increase in central noradrenergic transmission (Frangos et al., 2015; Yakunina et al., 2017; Ventura-Bort et al., 2018; Sclocco et al., 2019; Borgmann et al., 2021; Teckentrup et al., 2021; Giraudier et al., 2022; Figure 1). The availability of this neuromodulatory strategy has opened up new possibilities to test whether the vagus nerve fulfils a similar role for emotional memory in humans as it does in animals. In the following sections, we will provide a brief overview of recent stimulation studies, which support the view that vagal firing promotes episodic and associative emotional memory establishment. Based on this body of evidence, we will highlight potential areas of clinical application where VNS may be utilized to facilitate the long-term consolidation of learning experiences.

Initial evidence for the causal influence of ascending vagal fibers in emotional memory came from animal research (Clark et al., 1995, 1998). In one study, Clark et al. (1998) trained rats in an inhibitory avoidance task, where animals were to learn to avoid an electric shock. Directly after, animals received either 30 s of VNS or sham stimulation (see Table 1). The authors observed that VNS compared to sham stimulation improved memory performance, assessed 24 h after stimulation, especially for intermediate stimulation intensities (Clark et al., 1998; see also Sanders et al., 2019; Olsen et al., 2022; for a review see Olsen et al., 2023; also see Table 1). Subsequent studies have shown that VNS modulates molecular mechanisms within the HC, suggesting that VNS-induced memory improvements are likely related to hippocampal activity (c.f., Olsen et al., 2023).

In a follow up study, Clark et al. (1999) attempted to extend their animal findings to humans. In this study, epileptic patients with an implanted VN stimulator underwent verbal learning before receiving either active VNS at varying intensities (Table 1) or no stimulation in a control condition. Memory performance was tested in an immediate word-recognition test. Results showed that when the stimulation device was turned on after encoding, participants exhibited better memory performance, compared to when the stimulator remained off, particularly at intermediate stimulation intensities (Clark et al., 1999). Subsequent investigations on the effects of iVNS on memory performance have, however, yielded mixed findings. While some studies also reported positive effects of iVNS on memory performance (Ghacibeh et al., 2006), others failed to observed memory improvements after iVNS (Mertens et al., 2022; for a review see Olsen et al., 2023; see Table 1).

Unlike animal studies, which focused on emotional memories, human studies using iVNS did not assess episodic memory for emotional information, limiting the generalization of the same neural path from animals to humans. Ventura-Bort et al. (2021) therefore investigated the role of the VN on the formation and consolidation of emotional episodic memories in humans using taVNS. In this study, participants underwent two encoding sessions in which unpleasant and neutral images were encoded while receiving taVNS or sham stimulation in a counterbalanced, within-subjects design. One week later, recognition memory was tested by also assessing the contribution of familiarity vs. recollection-based remembering (i.e., low vs. high confidence; Wixted and Stretch, 2004), with the latter representing a more elaborate mnemonic process that particularly reflects increased amygdala and hippocampal activity (Dolcos et al., 2005, 2020).

Although no overall effects of VNS on memory performance were found (see also Ludwig et al., 2025), unpleasant images encoded under taVNS were more often retrieved with high confidence (Ventura-Bort et al., 2021), indicating a recollection-driven increase for emotional but not neutral images (see also Ludwig et al., 2025 for general memory improvements for unpleasant images encoded under taVNS). These findings were also accompanied by larger recollection-sensitive brain potentials (late ERP Old/New effect) during retrieval of emotional scenes encoded under taVNS, compared to sham stimulation (for a recent conceptual replication of the electrocortical findings using different stimulation protocols, see Ventura-Bort et al., 2025). Similar recollection-related results were also obtained in a behavioral study investigating the effects of taVNS on memory for emotional and neutral words (Giraudier et al., 2020). Although no overall effects of taVNS were found, participants receiving taVNS during the word encoding task showed a recollection-driven advantage (i.e., for words with the highest confidence) 1 day later (but see for no effects when taVNS was applied offline using a same day memory paradigm, Mertens et al., 2020). In contrast to Ventura-Bort et al. (2021), however, no emotion-specific memory enhancement was observed after taVNS (Giraudier et al., 2020), which may be partly explained by the use of the less arousing emotional material (words compared to high arousing pictures, c.f., Ventura-Bort et al. (2021)).

Altogether, these results suggest that stimulation of vagal afferents improves the formation of episodic, particularly recollection-based (i.e., hippocampal-mediated), memories, as indicated by behavioral and electrophysiological measures.

Importantly, it is not only vital for survival to remember distinct stimuli, but also to remember associations between them, ultimately allowing to anticipate upcoming events based on past experiences. In both animals and humans, such associative emotional memory is predominantly investigated by means of Pavlovian conditioning and extinction protocols. During conditioning, an inherently neutral conditioned stimulus (CS; e.g., a light in animal research; a geometrical figure in human research) repeatedly predicts the occurrence of an emotionally salient, i.e., unpleasant or pleasant, unconditioned stimulus (US; e.g., electric shock or food incentive) (Lonsdorf et al., 2017). As a result, the CS gains the capacity to elicit conditioned emotional responses (e.g., fear in case of a highly aversive US), reflecting a learned CS–US association (Lonsdorf et al., 2017). In contrast, during subsequent extinction protocols, the CS is no longer paired with any US (Lonsdorf et al., 2017). Thus, a novel association (CS-No US) is established, that inhibits the activation of an originally conditioned memory trace and thus reduces conditioned emotional responses (Dunsmoor et al., 2015b). Importantly, vagal projection targets (see Figure 1) play pivotal roles in both conditioning and extinction: Plasticity in the basolateral amygdala underlies initial associative learning (i.e., initial conditioning and extinction), while the medial prefrontal cortex is particularly involved in the consolidation and recall of extinction memory (Herry et al., 2008; Senn et al., 2014; Tovote et al., 2015; Szeska et al., 2022). In addition, noradrenaline heavily impacts on the plasticity in these areas, and consequently increased activity of the LC-NA system has found to promote associative memory processes (Uematsu et al., 2017; Giustino and Maren, 2018).

Accordingly, animal studies found promoted associative emotional memory by iVNS: Rats, that underwent an extinction protocol under iVNS, consistently show promoted extinction of previously conditioned fear–an effect, that maintains even for 10 days and may even reverse experimentally induced extinction impairments (Peña et al., 2013, 2014; Alvarez-Dieppa et al., 2016; Noble et al., 2017, 2019; Souza et al., 2019, 2021, 2022; see Table 1). Importantly, this extinction enhancement by iVNS is abolished if the LC is optogenetically silenced (Calderon-Williams et al., 2024), providing strong mechanistic evidence that respective memory improvements are dependent to vagal projections to the LC-NA system. Notably, the enhancing effects of iVNS are not limited to the extinction of fear, but also apply to the extinction of conditioned appetitive responses (e.g., cocaine-induced place preference) (Childs et al., 2017, 2019). Animal research therefore strongly suggests, that the vagus nerve is involved in guiding the establishment of associative emotional memory, primarily through its projections to the LC-NA system.

Human research utilizing taVNS has largely translated these findings from the animal model, although results have been mixed (see e.g., Burger et al., 2018 and D’Agostini et al., 2025; Szeska et al., 2020, 2021; see Table 1): Using a differential-cue Pavlovian fear conditioning paradigm, requiring discriminative learning between a threat and a safety cue, Burger and colleagues showed that taVNS accelerates the extinction of previously conditioned fear (Burger et al., 2016, 2017). Notably, though, these extinction enhancements were limited to verbal report measures of fear (i.e., threat expectancy). Yet, in a simpler single-cue Pavlovian fear conditioning paradigm, which is more closely adapted to animal research and requires simpler associative learning (either threat learning or not in a between-group design), taVNS has found to accelerate the extinction of verbal report, behavioral and physiological components of the fear response (Szeska et al., 2020, 2021). Consistent with prolonged extinction enhancements observed in animals, such beneficial effect maintained for even 4 weeks. However, the extinction enhancements by taVNS may be limited to the learning of associations between distinct cues: In a differential-context Pavlovian fear conditioning protocol, requiring discriminative learning of a threat-signaling vs. safety-signaling environment in virtual reality, taVNS failed to promote the extinction of contextually-related conditioned responses (Genheimer et al., 2017).

As for episodic memory, the current body of evidence therefore also generally supports a pivotal role of the vagus nerve in the establishment of human associative emotional memory, although beneficial effects primarily unfold when simple cue-outcome relationships are to be learned.

Based on the beneficial effects of non-invasive VNS on memory formation, this method was rendered a potential adjunct to the treatment of mental conditions, that are either marked by impairments in memory performance or where treatments are hinged upon successful learning of new information. In the following sections we will highlight such clinical areas, where taVNS may therefore be utilized to aid treatment strategies, including neurocognitive, anxiety and trauma-related disorders (see Figure 2).

A prioritized storage of emotionally salient stimuli into long-term memory warrants that significant cues, people, places and events can be remembered in the future, thus acting as a prerequisite for behavioral adaptivity (Dunsmoor and Kroes, 2019). Animal research has long suggested the vagus nerve to be a central relay driving this effect, since adrenergic actions on this nerve in emotionally arousing situations elicit central noradrenergic release, which promotes plasticity in memory-relevant brain areas (McIntyre et al., 2012). The availability of electrical invasive and non-invasive vagus nerve stimulation (VNS) has recently opened up the possibility to test a comparable vagal route to memory in humans. In this review, we synthesized animal and human studies utilizing VNS, which altogether suggested that (1) the vagus nerve constitutes an evolutionarily preserved brain-body axis driving emotional memory formation and (2) VNS may consequently be used to promote emotional episodic and associative memory consolidation. Following up on the latter notion, we further discussed VNS as a tool to combat memory decline in neurocognitive disorders and as an adjuvant to facilitate learning processes underlying exposure-based treatment of anxiety and trauma-related disorders. In fact, preliminary evidence suggests that electrical VNS improves mnemonic performance in mild cognitive impairment and Alzheimer’s disease (Wang et al., 2022), just as it promotes responding to exposure-based treatment (Powers et al., 2025; Szeska et al., 2025), rendering electrical VNS as a promising adjunct to a variety of therapeutic strategies. Interestingly, this might also apply to non-electrical VNS: The afferent vagus nerve is a critical component of the microbiota-gut-brain axis and thus it may be possible to invoke VNS by manipulations of the microbiome to achieve similar beneficial effects (Kuijer and Steenbergen, 2023; Faraji et al., 2025).

Nevertheless, more research into the mechanisms of electrical VNS is necessary to fully exhaust its memory-enhancing potential and utilize this stimulation technique in clinical areas: Although there is abundant evidence that VNS increases noradrenergic transmission by projections to the locus coeruleus (Frangos et al., 2015; Yakunina et al., 2017; Ventura-Bort et al., 2018; Sclocco et al., 2019; Borgmann et al., 2021; Teckentrup et al., 2021; Giraudier et al., 2022), VNS also targets centers of other transmitter systems (Frangos et al., 2015) and thus modulates cholinergic (Bowles et al., 2022), dopaminergic (Manta et al., 2009; Brougher et al., 2021), serotonergic (Manta et al., 2009, 2013) as well as glutamatergic and GABA-ergic neural transmission (Ben-Menachem et al., 1995). As each of these systems is involved in memory formation (Myhrer, 2003), the exact mechanism that underlies VNS-driven memory enhancement is yet to disentangle. Likewise, it needs to be determined under which conditions VNS yields the strongest memory-enhancing effects: Of the 40 articles included in our mini-review (see Table 1), only four have reported null-effects by VNS, with all using taVNS. Thus, we may preliminarily conclude that iVNS produces more consistent effects, possibly as it exhibits more direct action on the vagus. However, future studies need to additionally test the role of different stimulation parameters (frequency, intensity, duration; see D’Agostini et al., 2023) in VNS effects: Indeed, it has been shown that adrenergic agents affect memory retention in an inverted-U-shaped function, that depends upon the arousal level of the organism, where exceeding levels of systemic adrenaline may even impair memory performance (Gold and Korol, 2012). As VNS is suggested to tap into similar mechanisms, it is tempting to speculate that its memory enhancing effects are similarly shaped, i.e., being strongest at intensities that – when acting upon the current arousal state – invoke moderate levels of adrenergic activation (for preliminary evidence, see Souza et al., 2021). However, besides the current arousal state, further individual characteristics may also shape VNS effects and need to be systematically investigated (e.g., age, genetic factors, baseline cognitive functions, but also baseline vagal tone). Furthermore, it needs to be tested how (1) acute vs. multiple stimulation, (2) online (during task) vs. offline stimulation (before and/or after task), (3) type of task (e.g., relying on HC function), and (4) immediate vs. delayed testing impact on the effects of VNS on learning and memory. This also applies to the affective valence of the encoded material: While our review showed that VNS enhances memory for unpleasant material, to best of our knowledge there is no study that investigated whether the same holds true for pleasant stimuli, despite (mnemonic) processing similarly taps into (nor-) adrenergic mechanisms (Sternberg et al., 1985; Codispoti et al., 2003). Hence, we hope that our mini-review strongly encourages specific experimental designs or meta-and mega-analytic approaches (see Giraudier et al., 2022) to answer these open research questions. While VNS is already an FDA-approved clinical treatment of epilepsy and drug-resistant major depressive disorder (and applied in further clinical trials focusing e.g., on Alzheimer’s disease, mild cognitive impairment, PTSD, alcohol use disorder or stroke rehabilitation; see (Herr et al., 2024)), addressing these gaps will help to integrate it even more effectively into therapeutic strategies and tailor this stimulation technique for specific patient populations.

Altogether, this mini-review revealed that the vagus nerve constitutes a major communication route between the body’s periphery and the brain, which is critically involved in the formation of emotional memories. Vagus nerve stimulation can therefore be considered as one of the most promising neuromodulation techniques to combat mental disorders, and its full potential at this is yet to unfold.