Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 May 15.
Published in final edited form as: Neuroscience. 2019 Mar 21;406:290–299. doi: 10.1016/j.neuroscience.2019.03.019

Vagus nerve stimulation rate and duration determine whether sensory pairing produces neural plasticity

Elizabeth P Buell 1,2, Michael S Borland 1,2, Kristofer W Loerwald 1,2, Collin Chandler 2, Seth A Hays 1,2,3, Crystal T Engineer 1,2, Michael P Kilgard 1,2
PMCID: PMC6511481  NIHMSID: NIHMS1021496  PMID: 30904665

Abstract

Repeatedly pairing a brief train of vagus nerve stimulation (VNS) with an auditory stimulus drives reorganization of primary auditory cortex (A1). A number of studies document that the stimulation parameters of VNS, including intensity and pulse rate, regulate the magnitude of VNS-dependent cortical plasticity. However, there is currently little data to guide the selection of VNS train durations, an easily adjusted parameter that could influence the effect of VNS-based therapies. Here, we tested the effect of varying the duration of the VNS train on the extent of VNS-dependent cortical plasticity. Rats were exposed to a 9 kHz tone 300 times per day for 20 days. Coincident with tone presentation, groups received trains of 4, 16, or 64 pulses of VNS delivered at 30 Hz, corresponding to train durations of 0.125 s, 0.5 s, and 2.0 s, respectively. High-density microelectrode mapping of A1 revealed that 0.5 s duration VNS trains significantly increased the number of neurons in A1 that responded to tones near the paired tone frequency. Trains lasting 0.125 or 2.0 s failed to alter A1 responses, indicating that both shorter and longer stimulation durations are less effective at enhancing plasticity. A second set of experiments evaluating the effect of delivering 4 or 64 pulses in a fixed 0.5 s VNS train duration paired with tone presentation reveal that both slower and faster stimulation rates are less effective at enhancing plasticity. We incorporated these results with previous findings describing the effect of stimulation parameters on VNS-dependent plasticity and activation of neuromodulatory networks to generate a model of synaptic activation by VNS. The findings from the present study may facilitate determination of optimal parameters for clinical implementation of VNS therapies.

Keywords: plasticity, auditory cortex, vagal nerve stimulation, duration, pulses

Introduction

Precisely-timed vagus nerve stimulation (VNS) has emerged as a strategy to treat a wide range of neurological disorders, including tinnitus and stroke (Ridder et al., 2013; Dawson et al., 2016; Hays, 2016; Engineer et al., 2017). VNS engages pro-plasticity neuromodulatory networks, including the noradrenergic and cholinergic systems, during auditory or motor rehabilitative training to drive robust, training-specific neural plasticity (Engineer et al., 2011; Porter et al., 2012; Hulsey et al., 2016a, 2016b). This enhancement of plasticity is believed to underlie the therapeutic benefits of VNS (Hays et al., 2013). Therefore, identifying stimulation paradigms that maximize plasticity and thus yield greater recovery is a critical step in the effective translation of VNS therapy.

A number of previous papers have systematically evaluated how the stimulation rate, intensity, spacing between pairings, pulse width, and number of stimulation-tone pairings influence the degree of VNS-dependent plasticity (Borland et al., 2016, 2018; Loerwald et al., 2017; Buell et al., 2018). The effect of inter-stimulation timing on plasticity reveals a monotonic relationship, such that the magnitude of plasticity increases with the interval between stimulations (Borland et al., 2018). Other stimulation parameters exhibit a non-monotonic inverted-U relationship. For instance, moderate stimulation intensities drive robust plasticity, while both low and high stimulation intensities fail to enhance cortical reorganization (Borland et al., 2016). Similarly, at a fixed stimulation intensity and inter-stimulation interval, plasticity is not observed when the pulse rate of VNS trains is too distributed or compressed, while moderate pulse rates significantly enhance plasticity (Buell et al., 2018). Together, these studies highlight the complex effects of timing and the amount of stimulation on the degree of VNS-dependent plasticity.

In this study, we evaluated the effect of modulating the duration of the stimulation train on the magnitude of VNS-directed plasticity in auditory cortex. First, we examined VNS-dependent plasticity in primary auditory cortex after pairing short (125 ms), standard (500 ms), or long (2000 ms) trains of 30 Hz pulse rate VNS with a 9 kHz tone. Because changing the train duration at a constant pulse rate produces stimulation trains that deliver differing amounts of total charge in each group, we performed a second experiment to match charge delivery and determine whether the train duration or the total charge delivery was responsible for effects on plasticity. The results from these empirical studies were combined with previously published data to develop a model describing the action of VNS engagement of neuromodulator networks and the resultant magnitude of plasticity. Together, this study provides a characterization of the effect of stimulation duration on VNS-dependent plasticity and develops a testable framework to probe the interaction of stimulation parameters on the magnitude of plasticity in response to paired VNS.

Methods

All handling, housing, stimulation, and surgical procedures were approved by The University of Texas at Dallas Institutional Animal Care and Use Committee and by the Animal Care and Use Review Office of the United States Army Medical Research and Materiel Command Office of Research Protections. 86 adult female Sprague–Dawley rats (250–400 g) were used in this study. Data was analyzed and reported for 60 rats. The remaining 26 animals were excluded based on the following predefined criteria, as in previous studies: incomplete cortex map (n = 12), failure of head mount (n = 8), or failure of VNS lead (n = 6, confirmed by an absence of VNS-elicited reduction in blood oxygen saturation in anesthetized rats) (Borland et al., 2016, 2018, Loerwald et al., 2017, 2018; Buell et al., 2018). In experiment 1, rats were interleaved and randomly assigned to receive short VNS (n = 9; 4 pulses, 125 ms), standard VNS (n = 10; 16 pulses, 500 ms), or long VNS (n = 10; 64 pulses, 2000 ms). In experiment 2, rats were interleaved and randomly assigned to receive slow VNS (n = 7; 4 pulses at 7.5 Hz), standard VNS (n = 10; 16 pulses at 30 Hz), or fast VNS (n = 7; 64 pulses at 120 Hz). Animals from the Standard VNS and naïve groups were temporally interleaved with their respective experimental groups for experiments 1 and 2. Thus, of the 14 animals in the naïve group, 6 were included in the analysis for both experiment 1 and 2 and 8 were included only in one experiment, and of the 13 animals in the Standard VNS group, 7 were included in the analysis for both experiments and 6 were included only in one experiment. All rats were housed in a 12:12 hour reversed light–dark cycle with ad libitum access to food and water.

Vagus nerve surgery

VNS cuff implantation followed previously described procedures (Engineer et al., 2011, 2015; Porter et al., 2012; Borland et al., 2016; Buell et al., 2018). Animals were deeply anesthetized using ketamine hydrochloride (80 mg/kg, intraperitoneal (IP) injection) and xylazine (10 mg/kg IP) and given supplemental doses as needed. A Ringer’s lactate and dextrose solution (5 mL) was given to the rats to prevent dehydration. Cefotaxime sodium (2 × 10 mg, subcutaneous (SC) injection) solution was administered to animals after the surgery to prevent infection. After surgical preparation, all animals were implanted with a skull mounted connector. Marcaine (1 mL, SC) was injected into the scalp at the incision site. Bregma and lambda landmarks on the skull were exposed by an incision of the scalp running anterior to posterior. Bone screws were placed near the bregma suture, near the sagittal suture, near the lambda suture, and over the cerebellum. The connector headcap was secured to the screws with acrylic.

After implantation of the skull mount, animals were placed in a supine position for implantation of the stimulating cuff. Lidocaine (2%, 0.5 mL SC) was injected in the neck at the incision site, the left cervical vagus nerve was exposed through blunt dissection of the neck. The cuff electrode was secured around the nerve. Leads from the electrode were tunneled subcutaneously from the implant site in the neck between the eye and the ear and connected to the headcap on top of the skull. The headcap and cuff pins were secured with acrylic. After connection, confirmation of cuff function was obtained by observation of a blood oxygen drop while stimulating the nerve, as described in previous studies (Borland et al., 2018; Buell et al., 2018). Upon confirmation that the cuff was successfully stimulating the nerve, the neck was sutured closed and a topical antibiotic cream was applied to incision sites on the neck and head. Animals were given amoxicillin (5 mg) and carprofen (1 mg) for 2 days after surgery.

Vagus nerve stimulation

After 5–7 days of recovery from surgery, animals began the VNS tone pairing procedure, which consisted of 300 presentations of a 9 kHz tone each day for 20 days with VNS delivered as appropriate for each group, as in previous studies (Engineer et al., 2011; Borland et al., 2016; Loerwald et al., 2017; Buell et al., 2018). For each daily session, animals were placed in a 25 cm × 25 cm × 25 cm wire cage within a 50 cm × 60 cm × 70 cm chamber lined with acoustic insulating foam. A speaker suspended above the cage presented a 500 ms 9 kHz tone at 50 dB SPL with a 50% probability every 15 s, such that the average interval between tone presentations was 30 s. A VNS train consisting of biphasic 100 μs pulses at 0.8 mA was delivered 150 ms before the onset of each tone. The duration and pulse rate of the train for experimental groups were as follows: Short VNS, 125 ms duration at 30 Hz, n = 9; Standard VNS, 500 ms duration at 30 Hz, n = 10; Long VNS, 2000 ms duration at 30 Hz, n = 10; Slow VNS, 500 ms duration at 7.5 Hz, n = 7; Fast VNS, 500 ms duration at 120 Hz, n = 7. The impedance of the cuff electrode was verified daily and animals were excluded if measured cuff impedances were greater than 10 kΩ, as in previous studies.

Auditory cortex recordings

Using standard procedures, multi-unit responses were collected from the auditory cortex twenty-four hours after the final VNS-tone pairing session (Polley et al., 2007; Engineer et al., 2011; Borland et al., 2016, 2018; Buell et al., 2018). Rats were anesthetized with sodium pentobarbital (50 mg/kg) and anesthesia was maintained using supplemental doses of diluted pentobarbital by evaluating anesthesia levels between cortex recordings every 30–60 minutes (0.2–0.4 ml, 8 mg/ml). One mL of a one to one ratio of dextrose (5%) and standard Ringer’s lactate solution was administered between every recording to prevent dehydration. To minimize respiratory problems under pentobarbital, a tracheotomy was performed. In addition, cerebral edema was minimized by opening a cisternal drain. A craniotomy exposed the right primary auditory cortex. A durotomy over this exposed section of cortex further exposed the cortex and allowed electrodes to be placed easily. A thin film of silicone oil was placed over the exposed cortex to maintain hydration. Four parylene coated tungsten microelectrodes (1.5–2.5 MΩ, FHC, Bowdoin, ME) were lowered to layer IV/V of the primary auditory cortex. Pure tones were delivered from a speaker positioned 10 cm from the animal’s left ear in a foam-shielded double-walled sound-attenuated chamber. Neuronal responses were recorded using Brainware software (Tucker-Davis Technologies, Alachua, FL). Recording sites were tracked using a photo of the cortex in Canvas 16 software. Various tones were presented at 81 logarithmically spaced frequencies spanning 1–32 kHz in 0.0625 octave steps at 16 intensities from 0 to 75 dB SPL in 5 dB steps. The tones (25-ms duration, 5-ms rise–fall time) were randomly interleaved and separated by 500 ms. These tones were used to determine auditory tuning curves at each of the evaluated sites. The recording window for action potentials is 8–40 ms. For each auditory map, experimenters were blind to the group. At the conclusion of the map, cuff function was confirmed by testing an oxygen saturation drop in response to 10 seconds of continuous vagus nerve stimulation. If a stimulation-dependent O2 saturation drop could not be confirmed, the animal was excluded from analysis.

Data analysis

All groups were analyzed using an automated MATLAB program. This program determines receptive fields based on the characteristics of responses at each site as determined by previous work (Polley et al., 2007). The characteristic frequency (CF) was defined at the frequency at which the lowest intensity evokes a response at a particular site. Response threshold was defined as the lowest intensity capable of evoking a response. Spontaneous firing rate was the rate of firing evoked across all tone frequencies when presented at an amplitude of 0 dB. The time it takes for maximum neural responses to occur was defined as the peak latency. The percentage of A1 responding analysis was calculated as the percentage of A1 receptive fields that evoke a response for each tone frequency-intensity combination. For percentage of A1 responding analysis, experimental and control groups were compared using a one-way ANOVA with post-hoc Bonferroni correction for multiple comparisons. For response strength analysis, a mixed-effects model using SPSS software was used to account for the different number of sites recorded for each animal. The fixed factor was the experimental group and the random factor was the individual animals. Simple contrast analysis was used to determine whether there were statistically significant differences in response strength after VNS-tone pairing.

Modeling Neuromodulator Levels

We developed a mathematical model of neuromodulatory levels in response to trains of VNS with differing stimulation parameters. The level of neuromodulator in arbitrary units as a function of time was described by the following equation:

Levelt=∑j=1NI×Ht−j−1f×e−ln2t1/2×t−j−1f

where I is intensity in mA, N is the number of pulses in the VNS train, f is pulse frequency, H(t) is the Heaviside step function, and t1/2 is half-life. Each individual pulse during a train of VNS produces a step function increase in neuromodulatory levels, modeled by the Heaviside unit step function, which was scaled by stimulation intensity. Levels decline following exponential decay with a half-life of 200 ms. The time between pulses and length of the VNS train is determined by the pulse frequency and number of pulses per train.

Results

The VNS-tone pairing train duration affects A1 plasticity

For the first experiment, a 9 kHz tone was paired with VNS presented at short, standard, or long train durations (short: 4 pulses, 125 ms; standard: 16 pulses, 500 ms; long: 64 pulses, 2000 ms, Figure 1). In experimentally naïve rats, approximately 33% of A1 responds to 50 dB tones with frequencies between 8 – 16 kHz (Figures 2a & 3). VNS-tone pairing significantly altered the percentage of A1 responding to the paired frequency range (F(3,38) = 6.81, p = 0.001, Figures 2 & 3). There was a significant increase in the percent of A1 responding to 8 – 16 kHz tones in the standard VNS-tone pairing group (p = 0.01, Figures 2f & 3). In contrast, there was no significant alteration in the percent of A1 responding in either the short or long VNS-tone pairing groups compared to the naïve control group (p > 0.05, Figures 2 & 3).

Figure 1.

Figure 1

Open in a new tab

Schematic diagram of the VNS-tone train pairing procedure for Experiment 1. A train of biphasic pulses were delivered to the left vagus nerve via a cuff electrode. Each group received 30 Hz VNS at one of three different train durations: Short VNS (4 pulses over 0.125 s), Standard (16 pulses over 0.5 s), or Long VNS (64 pulses over 2 s). Rats received VNS paired with a 9 kHz tone every 30 s, 300 times during each 2.5 hour session for 20 days.

Figure 2.

Figure 2

Open in a new tab

Percent of A1 responding to each tone frequency intensity combination for naïve control rats (A), Short 0.125 s VNS (B), Standard 0.5 s VNS (C), and Long 2.0 s VNS (D). (E-G) Plots illustrating the difference between the percent of A1 responding for naïve controls and each VNS condition. Red regions indicate a greater percent of A1 neurons that respond in VNS-tone paired rats compared to naïve control rats, while blue regions indicate a decrease in the percent of A1 neurons that respond to that tone frequency intensity combination. Note that 0.5 s Standard VNS-alters the cortical response more than Short or Long VNS-tone pairing. White lines delineate the frequency intensity combinations which activate significantly more or fewer neurons after VNS pairing (p < 0.05).

Figure 3.

Figure 3

Open in a new tab

VNS-tone pairing reorganizes the auditory cortex frequency map as a non-monotonic function of VNS duration. The 0.5 s Standard VNS group exhibits a significant expansion of the percentage of A1 responding to 8–16 kHz frequency tones at 50 dB SPL after 20 days of VNS tone-pairing. The 0.125 s Short VNS and 2.0 s Long VNS groups do not demonstrate a significant increase in neurons responding to 8–16 kHz (p > 0.05).

This increase in the percent of A1 responding to tones neighboring the paired tone frequency was accompanied by alterations in the A1 response strength to tones (F(7, 86.69 = 16.27, p < 0.0001). In addition to an increase in the percent of A1 responding, the standard VNS-tone pairing group exhibited a 58% increase in the number of spikes evoked for tones between 8 – 16 kHz compared to the control group (t(46.27) = −2.2, p = 0.04, Figure 4). The short and long VNS-tone pairing groups did not have alterations in the number of spikes evoked for tones between 8 – 16 kHz; however, both groups exhibited a significant decrease in the number of spikes evoked for tones between 1 – 2 kHz (short: t(48.51) = 2.1, p = 0.04; long: t(48.8) = 2.7, p = 0.009, Figure 4). The standard VNS-tone pairing group that experienced 16 VNS pulses at 30 Hz for 500 ms paired with tone presentation exhibited A1 plasticity specific to the paired tone frequency, while the short and long VNS-tone pairing groups failed to alter A1 responses to the paired sound.

Figure 4.

Figure 4

Open in a new tab

VNS-tone pairing reduces the number of spikes evoked by low-frequency tones (1– 2 kHz, blue), and increases the number of spikes evoked by high-frequency tones (8–16 kHz, red). Standard VNS significantly increases the number of spikes in evoked by 8–16 kHz 50 dB SPL. This change is not observed with Short or Long VNS. Error bars indicate standard error of the mean across rats. Asterisks indicate experimental groups that were statistically significant from the naïve control group (p < 0.05).

The VNS-tone pairing train speed affects A1 plasticity

The previous experiment determined that a 500 ms VNS train duration resulted in A1 plasticity that was specific to the paired sound. In a second experiment, we sought to hold the stimulation train duration constant and vary the rate of the pulses to evaluate an interaction between duration and rate. To do so, a 9 kHz tone was paired with a 500 ms train of VNS presented at slow, standard, or fast train durations (slow: 4 pulses at 7.5 Hz; standard: 16 pulses at 30 Hz; fast: 64 pulses at 120 Hz, Figure 5). Across groups, VNS-tone pairing significantly altered the percentage of A1 responding to the paired frequency range (F(3,33) = 3.09, p = 0.04, Figures 6 & 7). There was a significant increase in the percent of A1 responding to 8 – 16 kHz tones in the standard VNS-tone pairing group (p = 0.003, Figures 6f & 7). In contrast, there was no significant alteration in the percent of A1 responding in either the slow or fast VNS-tone pairing groups compared to the naïve control group (p > 0.05, Figures 6 & 7).

Figure 5.

Figure 5

Open in a new tab

Schematic diagram of the VNS-tone train pairing procedure for Experiment 2. Each group received VNS at one of three different rates: Slow VNS (4 pulses at 7.5 Hz ), Standard VNS (16 pulses at 30 Hz), or Fast VNS (64 pulses at 120 Hz). Each train of VNS was paired with a 9 kHz tone 300 times during each 2.5 hour session for 20 days.

Figure 6.

Figure 6

Open in a new tab

Percent of A1 responding to each tone frequency intensity combination for naïve controls (A), 7.5 Hz Slow VNS (B), 30 Hz Standard VNS (C), and 120 Hz Fast VNS (D). Plots illustrating the difference between the percent of A1 responding for naïve controls and each VNS condition. Red regions indicate a greater percent of A1 neurons that respond in VNS-tone paired rats compared to naïve control rats, while blue regions indicate a decrease in the percent of A1 neurons that respond to that tone frequency intensity combination. Note that 0.5 s Standard VNS-alters the cortical response more than Slow or Fast rate VNS-tone pairing. White lines delineate the frequency intensity combinations which activate significantly more or fewer neurons after VNS pairing (p < 0.05).

Figure 7.

Figure 7

Open in a new tab

VNS-tone pairing reorganizes the auditory cortex frequency map as a non-monotonic function of VNS rate. 30 Hz rate Standard VNS results in a significant expansion of the percentage of A1 responding to 8–16 kHz frequency tones at 50 dB SPL after 20 days of VNS tone-pairing (p < 0.005). 7.5 Hz Slow VNS rate and 120 Hz Fast VNS rate groups do not exhibit a significant increase in neurons responding to 8–16 kHz (p > 0.05).

Similar to experiment 1, there were significant alterations in the A1 response strength to tones following VNS-tone pairing (F(7, 74.91 = 10.88, p < 0.0001). In addition to an increase in the percent of A1 responding, the standard VNS-tone pairing group exhibited a 60% increase in the number of spikes evoked for tones between 8 – 16 kHz compared to the control group (t(38.95) = −2.4, p = 0.02, Figure 8). The slow and fast VNS-tone pairing groups did not have alterations in the number of spikes evoked for tones between 8 – 16 kHz; however, the slow VNS-tone pairing group exhibited a significant decrease in the number of spikes evoked for tones between 1 – 2 kHz compared to the naïve control group (t(40.73) = 2.9, p = 0.006, Figure 8). The standard VNS-tone pairing group that experienced 16 VNS pulses at 30 Hz for 500 ms paired with tone presentation exhibited A1 plasticity specific to the paired tone frequency, while the slow and fast VNS-tone pairing groups failed to alter A1 responses to the paired sound.

Figure 8.

Figure 8

Open in a new tab

VNS-tone pairing reduces the number of spikes evoked by low-frequency tones (1– 2 kHz, blue), and increases the number of spikes evoked by high-frequency tones (8–16 kHz, red). Standard VNS increases the number of action potentials evoked by 8–16 kHz 50 dB SPL tones. This change is not observed with Slow or Fast VNS rates. Error bars indicate standard error of the mean across rats. Asterisks indicate experimental groups that were statistically significant from the naïve control group (p < 0.05).

Discussion

More than a dozen studies have confirmed that repeatedly pairing VNS with an auditory or motor event can drive highly specific and long-lasting plasticity in primary auditory or motor cortex, respectively (Shetake et al., 2011; Engineer et al., 2011, 2015, 2017; Porter et al., 2012; Borland et al., 2016, 2018, Loerwald et al., 2017, 2018; Morrison et al., 2018; Buell et al., 2018; Ganzer et al., 2018; Meyers et al., 2018). Since VNS-directed neural plasticity appears to enhance recovery of auditory and motor function following injury (Engineer et al., 2011; Hays et al., 2014; Khodaparast et al., 2014; Dawson et al., 2016; Pruitt et al., 2016; Tyler et al., 2017; Vanneste et al., 2017; Meyers et al., 2018), it is critical to understand the optimal conditions to drive plasticity in order to maximize therapeutic benefits. Previous papers have systematically evaluated how the stimulation rate, intensity, spacing between pairings, pulse width, and number of VNS-tone pairings influence the degree of neural plasticity (Borland et al., 2016, 2018; Loerwald et al., 2017; Buell et al., 2018). This study was designed to determine the influence of VNS pulse rate and train duration. Repeatedly pairing a 500 ms train of sixteen 0.8 mA pulses of VNS delivered at 30 Hz with a tone effectively drives map expansion in the auditory cortex. Shorter and longer stimulation durations as well as slower and faster stimulation rates are less effective at enhancing plasticity.

Numerous experiments have documented that VNS alters cortical function through activation of neuromodulatory networks (Nichols et al., 2011; Hulsey et al., 2016a, 2016b). VNS drives rapid, phasic activation of the noradrenergic locus coeruleus (LC) and increases levels of norepinephrine in multiple brain regions (Roosevelt et al., 2006; Hulsey et al., 2016b). Moreover, the rapid action of VNS on cortical activity requires acetylcholine (Nichols et al., 2011). VNS-dependent activation of these neuromodulatory networks coincident with neural activity evoked by an event, such as an auditory stimulus in the present study, enhances event-specific plasticity (Hays et al., 2013). Either depletion of neuromodulators or temporal dissociation of VNS and the event prevents VNS-dependent enhancement of plasticity, suggesting the rapid activation of these neuromodulatory networks is critical to enhanced plasticity (Engineer et al., 2011; Hulsey et al., 2016a). Despite the link between activation of these neuromodulatory systems and plasticity, the amount of activity in neuromodulatory neurons fails to be a good predictor of the magnitude of plasticity. Longer VNS train durations result in a linear increase in firing rate in the LC, which would lead to consequently higher concentrations of norepinephrine release (Hulsey et al., 2016b). However, a 2000 ms train of 30 Hz VNS paired with a 9 kHz tone was not capable of driving cortical plasticity, whereas a 500 ms train of 30 Hz VNS significantly enhanced plasticity while driving fewer spikes in the LC (Hulsey et al., 2016b). A similar dissociation between LC activity and plasticity has been described for VNS intensity (Borland et al., 2016). Thus, while activation of the noradrenergic network is critical for VNS-dependent plasticity, there is not a simple, direct relationship between LC activity and VNS-directed plasticity.

To reconcile these findings, we developed a model describing neuromodulatory activation in response to a train of VNS based on the experimental data from this study and on existing literature characterizing the effect of varying stimulation parameters on cortical plasticity (Engineer et al., 2011; Borland et al., 2016, 2018; Loerwald et al., 2017; Buell et al., 2018; Morrison et al., 2018). The model incorporates neuromodulatory activation in response to individual pulses of VNS in a train scaled by the intensity of each pulse. Pulses generate additive activation, such that longer trains and faster pulse rates summate into greater neuromodulatory activation. Levels decline following exponential decay. The model illustrates a bounded range of neuromodulatory activation within which stimulation parameters will effectively enhance plasticity (Fig. 9). VNS parameters that yield activation that falls below the lower bound of this range fail to enhance plasticity, likely due to insufficient engagement of signaling pathways. Alternatively, VNS parameters that yield activation exceeding the upper bound of the range result in over-activation that fails to enhance plasticity, presumably due to desensitization or engagement of an overriding anti-plasticity system.

Figure 9.

Figure 9

Open in a new tab

Model of neuromodulatory activation in response to VNS across a range of parameters. (A) We developed a model that predicted VNS-dependent neuromodulatory activation and incorporated plasticity outcomes from previous studies using a variety of stimulation parameters (Engineer et al., 2011; Borland et al., 2016, 2018; Loerwald et al., 2017; Buell et al., 2018). Each individual pulse during a train of VNS produces a step function increase in neuromodulatory levels that scales with stimulation intensity. Pulses generate additive activation, such that longer trains and faster pulse rates summate into greater neuromodulatory activation. Levels decline following exponential decay. Black lines denote VNS parameters the significantly enhanced plasticity in their respective studies, and gray lines indicate parameters that failed to enhance plasticity. A single VNS train for each parameter set is depicted in the lower panel. Standard VNS was a 500 ms train of 16 × 0.8 mA, 100 µs biphasic pulses delivered at 30 Hz. Other parameters are named by the feature that differs from Standard VNS. The green area represents a bounded region such that parameters that drive maximal neuromodulatory activation within this range result in significantly enhanced plasticity. Parameters that drive neuromodulatory activation that falls below the effective range are predicted to fail to enhance plasticity due to insufficient activation. Parameters that drive neuromodulatory activation that exceeds the upper bound of the effective region are predicted to fail to enhance plasticity due to over-activation of neuromodulatory signaling. (B) Depiction of predicted neuromodulatory activation in response to the parameters used in experiment 1 of the present study. The duration of Standard VNS drives neuromodulatory levels into the effective range and significantly enhances plasticity. Short VNS produces insufficient activation and Long VNS drives over-activation, and both parameters fail to significantly enhance plasticity. (C) Depiction of predicted neuromodulatory activation in response to the parameters used in experiment 2 of the present study. The moderate stimulation rate of Standard VNS drives predicted neuromodulatory activation into the effective range and results in significantly enhanced plasticity. Alternatively, Slow and Fast VNS generate insufficient and over-activation of neuromodulator levels respectively, and fail to enhance plasticity. Together, this model combines VNS-dependent plasticity outcomes and predicted neuromodulator levels to provide a framework to rationally define testable parameter sets. Neuromodulatory levels are in arbitrary units.

The bounded range of activation illustrates the inverted-U relationship observed between several stimulation parameters and the degree of VNS-dependent plasticity. In both auditory and motor cortex, moderate intensity stimulation significantly enhances plasticity, while low and high stimulation intensities fail to yield significant increases in cortical reorganization (Borland et al., 2016; Morrison et al., 2018) (Fig. 9A). Similar results in which moderate stimulation intensities are most efficacious have been described in studies evaluating VNS-dependent enhancement of memory and hippocampal plasticity (Clark et al., 1995, 1998, 1999; Zuo et al., 2007). Moreover, increasing pulse width can compensate for reductions in current amplitude, suggesting that total stimulation intensity as a function of charge per pulse is the main predictor of activation (Loerwald et al., 2017).

In addition to VNS intensity, the timing of pulses during a train influences the degree of neuromodulatory activation. A previous study demonstrated that pulse rate during a VNS train exhibits an inverted-U relationship with plasticity (Buell et al., 2018) (Fig. 9A). Moderate pulse rates drove maximal plasticity while high and low rates failed to significantly increase plasticity, even though total amount of stimulation was balanced across conditions. Findings from the present study corroborate the necessity of moderate stimulation rates and reveal additional insight into the importance of timing of neuromodulatory activation on VNS-directed-plasticity. In experiment 1, we found that a moderate duration train significantly enhances plasticity in A1, whereas shorter or longer stimulation trains do not drive significant plasticity (Fig. 9B). However, while the effect failed to reach significance (p = 0.06), Short VNS demonstrated a trend toward enhancing plasticity and the predicted neuromodulator levels for this stimulation paradigm fall very near the lower bound for the effective range. Additionally, in experiment 2, we found that slow and fast rates of VNS pulse delivery fail to enhance plasticity, while moderate rates generate robust plasticity (Fig. 9C). While the absence of an effect with longer and faster stimulation paradigms is generally counterintuitive, these results align with the inverted-U relationship observed with other stimulation parameters. These findings demonstrate the necessity of an upper bound on the model that incorporates not only degree of activation, but also accounts for the timing of activation.

A number of neural mechanisms can explain the behavior of the model. The most likely explanation that could account for this inverted-U relationship is a two-component system consisting of a low-threshold pro-plasticity component and an overriding high-threshold anti-plasticity component. Moderate activation of the system optimally enhances plasticity by maximizing engagement of the pro-plasticity component while concurrently minimizing activation of the high-threshold anti-plasticity component. Too little activation fails to provide sufficient engagement to drive plasticity, while too much activation generates strong engagement of the overriding anti-plasticity component and prevents plasticity. One likely candidate that could provide the biological manifestation of this system is adrenergic receptors. The noradrenergic system is rapidly activated by VNS and is required for VNS-dependent cortical plasticity (Hulsey et al., 2016b). Different classes of adrenergic receptors display a range of affinities for norepinephrine (Atzori et al., 2016). Moderate levels of norepinephrine release driven by VNS would preferentially activate higher-affinity α2-receptors, while higher levels would also activate lower-affinity β-receptors. Different classes of adrenergic receptors are coupled to distinct downstream signaling cascades, and existing evidence indicates that activation of these different classes of adrenergic receptors can indeed produce opposing actions on synaptic plasticity (Ramos and Arnsten, 2007; Salgado et al., 2012). Alternatively, a single desensitizing system could explain the inverted-U relationship. In this case, moderate activation by VNS would produce maximal plasticity, while higher activation would induce desensitization and blunt the response of repeated VNS pairing. G-protein-coupled receptors, which likely mediate the neuromodulatory response to VNS, are known to demonstrate desensitization (Gainetdinov et al., 2004). Although norepinephrine is most widely associated with the action of VNS, other neuromodulatory networks may contribute to the inverted-U response. Dopamine is known to influence plasticity and emerging evidence suggests it is co-released with norepinephrine from the LC (Kempadoo et al., 2016; Takeuchi et al., 2016). Beyond potential single-system actions of dopamine on plasticity, the interaction of varying levels of norepinephrine and dopamine release from the LC at different stimulation parameters could potentially represent an alternative two-component opposing system to generate the inverted-U. The serotonergic and cholinergic systems are also activated by VNS and known to be involved in cortical plasticity, thus may also contribute to VNS-dependent plasticity (Manta et al., 2009; Nichols et al., 2011; Hulsey et al., 2016a).

A clear delineation of the neuronal mechanisms that determine the inverted-U response may allow for rational selection and evaluation of stimulation parameters to maximize VNS-dependent plasticity. Future studies could leverage pharmacological manipulations to probe the contributions of adrenergic receptor types. If high-affinity α2-receptors mediate the pro-plasticity actions of VNS, then blockade of these receptors should prevent any VNS paradigm from enhancing plasticity. In the context of the present study, we predict antagonism of α2-receptors would prevent 0.5 s trains of VNS from enhancing plasticity. Additionally, if low-affinity β-receptors are responsible for the anti-plasticity actions observed at higher stimulation intensities, with longer trains, or with faster pulse rates, then antagonism of these receptors should convert parameters that previously exceeded the effective range to now enhance plasticity. For instance, we predict that antagonism of β-receptors would allow the 2.0 s VNS trains used in this study to enhance plasticity. Finally, because both the timing and amount of activation influence the degree of plasticity, manipulating neuromodulator levels with a reuptake inhibitor should impact the range of parameters which enhance plasticity. Blocking reuptake of norepinephrine would increase the dwell time of the neuromodulator, thus increasing its action. As a result, parameter sets that previously fell below the effective range, such as 0.125 s VNS trains or 0.2 mA VNS, may now provide sufficient activation to enhance plasticity. Additionally, some parameters that previously drove effective plasticity, like 0.5 s trains, would be expected to now generate overactivation and thus fail to enhance plasticity. Thus, the model described here generates testable predictions about ranges of parameters that enhance plasticity and may provide a framework to elucidate the neural mechanisms that underlie VNS-dependent plasticity.

In summary, this study demonstrates that the rate and duration of VNS when paired with a tone influence the extent of cortical change. Repeatedly pairing sixteen 0.8 mA pulses of VNS delivered at 30 Hz with a tone effectively drives map expansion in the auditory cortex. Four pulses of VNS at 7.5 Hz and 30 Hz, as well as sixty-four pulses at 30 Hz and 120 Hz, fail to drive plasticity. These findings likely arise from the interaction between norepinephrine and specific adrenergic receptors that regulate plasticity. Additional parametric evaluation of VNS would be beneficial in the development of this therapy for sensory, emotional, or motor disorders.

Acknowledgements

We would like to thank Alan Carroll, Emily Jensen, Madelyne Frech, Natasha Houshmand, Pryanka Sharma, Shen Xian, Mark Lane, Aisha Khan, Lena Sadler, Eric Meyers, Camilo Sanchez, Jayant Kurvari, Jordan Chen, Daniel Kaminski, Son Pham, and Samuel Yang for their contributions to this project. Special thanks to Victor Varner for input on mathematical modeling.

Funding

This work was supported by the National Institutes of Health R01NS085167 (MPK), R01NS085167 (MPK), R01DC017480 (CTE), and R01NS094384 (SAH) and by the Defense Advanced Research Projects Agency (DARPA) Biological Technologies Office (BTO) Electrical Prescriptions (ElectRx) program under the auspices of Dr. Eric Van Gieson through the Space and Naval Warfare Systems Center, Pacific Cooperative Agreement No. N66001-15-2-4057 and the DARPA BTO Targeted Neuroplasticity Training (TNT) program under the auspices of Dr. Tristan McClure-Begley through the Space and Naval Warfare Systems Center, Pacific Grant/Contract No. N66001-17-2-4011.

Footnotes

Financial Disclosure

MPK is a consultant for, and has a financial interest in, MicroTransponder, Inc., which is developing therapies using VNS. CTE is married to an employee of MicroTransponder Inc. All other authors report no conflicts of interest.

Publisher's Disclaimer: Disclaimer

Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of the Defense Advanced Research Projects Agency (DARPA) Biological Technologies Office (BTO).

Data Statement

All data from this study will be made freely available upon request to the corresponding author.

References

  1. Atzori M, Cuevas-Olguin R, Esquivel-Rendon E, Garcia-Oscos F, Salgado-Delgado RC, Saderi N, Miranda-Morales M, Treviño M, Pineda JC, Salgado H (2016) Locus ceruleus norepinephrine release: A central regulator of cns spatio-temporal activation? Front Synaptic Neurosci 8:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Borland MS, Engineer CT, Vrana WA, Moreno NA, Engineer ND, Vanneste S, Sharma P, Pantalia MC, Lane MC, Rennaker RL, Kilgard MP (2018) The Interval Between VNS-Tone Pairings Determines the Extent of Cortical Map Plasticity. Neuroscience 369:76–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Borland MS, Vrana WA, Moreno NA, Fogarty EA, Buell EP, Sharma P, Engineer CT, Kilgard MP (2016) Cortical Map Plasticity as a Function of Vagus Nerve Stimulation Intensity. Brain Stimul 9:117–123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Buell EP, Loerwald KW, Engineer CT, Borland MS, Buell JM, Kelly CA, Khan II, Hays SA, Kilgard MP (2018) Cortical map plasticity as a function of vagus nerve stimulation rate. Brain Stimul [DOI] [PMC free article] [PubMed]
  5. Clark KB, Krahl SE, Smith DC, Jensen RA (1995) Post-training unilateral vagal stimulation enhances retention performance in the rat. Neurobiol Learn Mem 63:213–216. [DOI] [PubMed] [Google Scholar]
  6. Clark KB, Naritoku DK, Smith DC, Browning RA, Jensen RA (1999) Enhanced recognition memory following vagus nerve stimulation in human subjects. Nat Neurosci 2:94–98. [DOI] [PubMed] [Google Scholar]
  7. Clark KB, Smith DC, Hassert DL, Browning RA, Naritoku DK, Jensen RA (1998) Posttraining electrical stimulation of vagal afferents with concomitant vagal efferent inactivation enhances memory storage processes in the rat. Neurobiol Learn Mem 70:364–373. [DOI] [PubMed] [Google Scholar]
  8. Dawson J, Pierce D, Dixit A, Kimberley TJ, Robertson M, Tarver B, Hilmi O, McLean J, Forbes K, Kilgard MP, Rennaker RL, Cramer SC, Walters M, Engineer N (2016) Safety, Feasibility, and Efficacy of Vagus Nerve Stimulation Paired with Upper-Limb Rehabilitation after Ischemic Stroke. Stroke 47:143–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Engineer CT, Engineer ND, Riley JR, Seale JD, Kilgard MP (2015) Pairing Speech Sounds With Vagus Nerve Stimulation Drives Stimulus-specific Cortical Plasticity. Brain Stimul 8(3):637–644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Engineer CT, Hays SA, Kilgard MP (2017) Vagus nerve stimulation as a potential adjuvant to behavioral therapy for autism and other neurodevelopmental disorders. J Neurodev Disord 9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Engineer ND, Riley JR, Seale JD, Vrana WA, Shetake JA, Sudanagunta SP, Borland MS, Kilgard MP (2011) Reversing pathological neural activity using targeted plasticity. Nature 470:101–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Gainetdinov RR, Premont RT, Bohn LM, Lefkowitz RJ, Caron MG (2004) Desensitization of G Protein–Coupled Receptors and Neuronal Functions. Annu Rev Neurosci 27:107–144. [DOI] [PubMed] [Google Scholar]
  13. Ganzer PD, Darrow MJ, Meyers EC, Solorzano BR, Ruiz AD, Robertson NM, Adcock KS, James JT, Jeong HS, Becker A, Goldberg MP, Pruitt DT, Hays SA, Kilgard MP, Rennaker RL (2018) Closed-loop Neuroprosthesis Restores Network Connectivity and Motor Control after Spinal Cord Injury. eLife:1–19. [DOI] [PMC free article] [PubMed]
  14. Hays SA (2016) Enhancing Rehabilitative Therapies with Vagus Nerve Stimulation. Neurotherapeutics 13:382–394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hays SA, Khodaparast N, Hulsey DR, Ruiz A, Sloan AM, Ii RLR, Kilgard MP (2014) Training Improves Functional Recovery After Intracerebral Hemorrhage. Stroke 45:97–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hays SA, Rennaker RL, Kilgard MP (2013) Targeting plasticity with vagus nerve stimulation to treat neurological disease [DOI] [PMC free article] [PubMed]
  17. Hulsey DR, Hays SA, Khodaparast N, Ruiz A, Das P, Rennaker RL, Kilgard MP (2016a) Reorganization of Motor Cortex by Vagus Nerve Stimulation Requires Cholinergic Innervation. Brain Stimul 9:174–181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hulsey DR, Riley JR, Loerwald KW, Rennaker RL, Kilgard MP, Hays SA (2016b) Parametric characterization of neural activity in the locus coeruleus in response to vagus nerve stimulation. Exp Neurol 298:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kempadoo KA, Mosharov EV, Choi SJ, Sulzer D, Kandel ER (2016) Dopamine release from the locus coeruleus to the dorsal hippocampus promotes spatial learning and memory. Proc Natl Acad Sci 113:14835–14840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Khodaparast N, Hays SA, Sloan AM, Fayyaz T, Hulsey DR, Rennaker RL, Kilgard MP (2014) Vagus nerve stimulation delivered during motor rehabilitation improves recovery in a rat model of stroke. Neurorehabil Neural Repair 28:698–706. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Loerwald KW, Borland MS, Rennaker RL, Hays SA, Kilgard MP (2017) The interaction of pulse width and current intensity on the extent of cortical plasticity evoked by vagus nerve stimulation. Brain Stimul [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Loerwald KW, Buell EP, Borland MS, Rennaker RL, Hays SA, Kilgard MP (2018) Varying Stimulation Parameters to Improve Cortical Plasticity Generated by VNS-tone Pairing. Neuroscience 388:239–247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Manta S, Dong J, Debonnel G, Blier P (2009) Enhancement of the function of rat serotonin and norepinephrine neurons by sustained vagus nerve stimulation. J psychiatry Neurosci JPN 34:272. [PMC free article] [PubMed] [Google Scholar]
  24. Meyers EC, Solorzano BR, James J, Ganzer PD, Lai ES, Rennaker RL, Kilgard MP, Hays SA (2018) Vagus Nerve Stimulation Enhances Stable Plasticity and Generalization of Stroke Recovery. Stroke 49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Morrison RA, Hulsey DR, Adcock KS, Rennaker RL, Kilgard MP, Hays SA (2018) Vagus nerve stimulation intensity influences motor cortex plasticity. Brain Stimul [DOI] [PMC free article] [PubMed]
  26. Nichols JA, Nichols AR, Smirnakis SM, Engineer ND, Kilgard MP, Atzori M (2011) Vagus nerve stimulation modulates cortical synchrony and excitability through the activation of muscarinic receptors. Neuroscience 189:207–214. [DOI] [PubMed] [Google Scholar]
  27. Polley DB, Read HL, Storace DA, Merzenich MM (2007) Multiparametric auditory receptive field organization across five cortical fields in the albino rat. J Neurophysiol 97:3621–3638. [DOI] [PubMed] [Google Scholar]
  28. Porter BA, Khodaparast N, Fayyaz T, Cheung RJ, Ahmed SS, Vrana WA, Rennaker RL, Kilgard MP (2012) Repeatedly pairing vagus nerve stimulation with a movement reorganizes primary motor cortex. Cereb Cortex 22:2365–2374. [DOI] [PubMed] [Google Scholar]
  29. Pruitt DT, Schmid AN, Kim LJ, Abe CM, Trieu JL, Choua C, Hays SA, Kilgard MP, Rennaker RL (2016) Vagus nerve stimulation delivered with motor training enhances recovery of function after traumatic brain injury. J Neurotrauma 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ramos BP, Arnsten AFT (2007) Adrenergic pharmacology and cognition: Focus on the prefrontal cortex. Pharmacol Ther 113:523–536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ridder D De, Vanneste S, Engineer ND, Kilgard MPMP (2013) Safety and efficacy of vagus nerve stimulation paired with tones for the treatment of tinnitus: a case series. Neuromodulation Technol Neural Interface 17(2):170–179. [DOI] [PubMed] [Google Scholar]
  32. Roosevelt RW, Smith DC, Clough RW, Jensen RA, Browning RA (2006) Increased extracellular concentrations of norepinephrine in cortex and hippocampus following vagus nerve stimulation in the rat. Brain Res 1119:124–132. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Salgado H, Köhr G, Treviño M (2012) Noradrenergic “Tone” Determines Dichotomous Control of Cortical Spike-Timing-Dependent Plasticity. Sci Rep 2:417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shetake JA, Engineer ND, Vrana WA, Wolf JT, Kilgard MP (2011) Pairing tone trains with vagus nerve stimulation induces temporal plasticity in auditory cortex. Exp Neurol 233:342–349. [DOI] [PubMed] [Google Scholar]
  35. Takeuchi T, Duszkiewicz AJ, Sonneborn A, Spooner PA, Yamasaki M, Watanabe M, Smith CC, Fernández G, Deisseroth K, Greene RW, Morris RGM (2016) Locus coeruleus and dopaminergic consolidation of everyday memory. Nature 537:357–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tyler R, Cacace A, Stocking C, Tarver B, Engineer N, Martin J, Deshpande A, Stecker N, Pereira M, Kilgard M, Burress C, Pierce D, Rennaker R, Vanneste S (2017) Vagus Nerve Stimulation Paired with Tones for the Treatment of Tinnitus: A Prospective Randomized Double-blind Controlled Pilot Study in Humans. Sci Rep 7:11960. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Vanneste S, Martin J, Rennaker RL, Kilgard MP (2017) Pairing sound with vagus nerve stimulation modulates cortical synchrony and phase coherence in tinnitus: An exploratory retrospective study. Sci Rep 7:17345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zuo Y, Smith DC, Jensen RA (2007) Vagus nerve stimulation potentiates hippocampal LTP in freely-moving rats. Physiol Behav 90:583–589. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES