Vestibular and dlPFC interactions
It has been previously established in many studies that MSNA undergoes robust modulation during low-frequency sGVS (Grewal et al.
2009; Hammam et al.
2011; Macefield and James
2016), with frequencies below 0.2 Hz eliciting the expression of secondary peaks of modulation (Hammam et al.
2011). The current study has found, however, that tACS of the ipsilateral dlPFC (with regards to the microneurographic recording site) on its own at 0.08 Hz will produce equivalent changes to levels of MSNA as produced by sGVS at the same frequency. Indeed, this stimulation was even shown to elicit distinct secondary peaks of modulation as well. Comparisons of paired as well as unpaired modulation indices between sGVS and tACS yielded no significant differences. This lack of significance, alongside the appearance of secondary peaks, shows that the dlPFC-controlled modulation of MSNA is just as robust as the vestibular modulation. Importantly, the combined stimulation protocol expressed the same trend as seen in stimulation of the vestibular apparatuses and the dlPFC as well. Modulation of MSNA remained the same; the two systems did not have an additive influence on each other (contrary to our initial hypothesis), nor did they cancel each other out. The absence of an increased or decreased modulation index strongly implies a disturbance in the higher order processing of sympathetic outflow.
The results seem to suggest that the vestibular contributions to the regulation of blood flow are being blocked by the dlPFC during the combined stimulation protocol. This is in line with our previous findings that the dlPFC was inhibiting vestibular-induced nausea and perceptions of sway (McCarthy et al.
2023a). While that paper explores possible dlPFC influences on higher cortical regions, these new results, however, suggest that the central sympathetic connectome at the level of the brainstem may be playing a key role in this process. It would appear that the vestibular afferent projections from the caudal vestibular nuclei to the NTS, CVLM, and RVLM in particular, may be inhibited by dlPFC activity.
One possible explanation of this dlPFC inhibitory function lies in upregulation of the gamma-aminobutyric acid (GABA) inhibitory pathway centring around the CVLM and RVLM (Agarwal et al.
1989,
1990). Increasing GABA transmission can impair the ability of the vestibular system to generate MSNA. Due to the aforementioned lack of an additive influence on MSNA modulation during the combined stimulation protocol (with evidence that modulation does, indeed, still take place, but not according to the initially hypothesised additive effect), there is an implication that some form of partial sympathetic inhibition is taking place. Holstein et al. (
2016) provide evidence that the vestibulosympathetic reflexes themselves are influenced by GABA transmission from the vestibular nuclei to the RVLM during sGVS. This is usually vastly outcompeted by the GABA transmission from the same vestibular nuclei to the CVLM, which reduces the CVLM inhibitory properties and allows for increased MSNA modulation. However, if the dlPFC can upregulate this sGVS-induced GABA to RVLM pathway, this may account for the results found in the present study. Through this, the dlPFC may be capable of increasing MSNA when stimulated on its own, and function to offset the vestibulosympathetic reflexes when stimulated in conjunction with the vestibular system – having no effect on upregulating vestibular-induced GABA to RVLM functions if the vestibular system is not concurrently engaged. Electrical stimulation of the dlPFC has, in fact, been demonstrated to increase GABA transmission in the past, albeit to the striatum rather than the brainstem (Bunai et al.
2021). Regardless, this provides evidence that the dlPFC is capable of mediating GABA transmission.
Moreover, the insular cortex (IC) may be playing an important role here. As we have previously discussed (McCarthy et al.
2023a), the insula has interconnections to both the vestibular system (Akbarian et al.
1993; Chen et al.
2010; Guldin and Grüsser
1998) and the dlPFC (Fu et al.
2021; Gao et al.
2021; Steward et al.
2016), which led us to believe that it may be a key structure in the abolition of vestibular-induced nausea and perceptions of sway. Additionally, in the fact that the insula has been well established to have cardiovascular control (Cechetto and Chen
1990; Chouchou et al.
2019), and in light of our newly acquired results, the possibility of its involvement in our stimulation protocol is further justified. In rats, the intermediate portion of the IC has been shown to have direct projections to the RVLM, while both the intermediate as well as the rostral regions have projections towards the CVLM (Marins et al.
2016). By injecting an excitatory neurotransmitter into the rostral region of the IC, Marins et al. (
2016) noted a marked reduction in heart rate and renal sympathetic nerve activity via the CVLM. This sympathetic alteration was seen in the opposite manner with neurotransmitter injection into the intermediate IC; heart rate and renal sympathetic nerve activity increased by means of the RVLM. Here, as with our GABA upregulation hypothesis, we may find evidence as to why there were no additive influences on MSNA between the dlPFC and vestibular system during our combined stimulation protocol. If the dlPFC has an excitatory influence on the rostral IC (and, hence, the CVLM) in a similar fashion to that found in these rats, or an inhibitory influence on the intermediate IC (hence the RVLM), we may expect to see a suppression of the vestibulosympathetic responses. The dlPFC may be barring the vestibular sympathetic pathway in particular – not the generalised MSNA pathway through which the dlPFC is capable of increasing MSNA when stimulated on its own (Sesa-Ashton et al.
2022). An inhibitory response on the RVLM through the IC would be consistent with one of our last hypotheses that downregulation of insula activity could ‘gate’ vestibular signal propagation (Huang et al.
2021; McCarthy et al.
2023a).
The latency data of MSNA peak expression reinforces the notion of a blockade of vestibular signals. Much like the modulation indices data, there were no significant differences between stimulation protocols in the latency of MSNA primary and secondary peaks. This supports our conclusion that the dlPFC does not work with or against the vestibular system, per se, rather that it prevents vestibular signalling altogether. Had the combined stimulation data shown a greater correlation to the tACS data, and by proxy been significantly different to the sGVS data, this would be further supported.
Data on the duration of primary and secondary peaks of MSNA likewise showed a lack of significant differences between the stimulation paradigms, with an exception. Both the paired and unpaired data revealed a significant difference between sGVS and combined stimulation primary peaks. This means that the combined stimulation, at least in this aspect, does indeed more closely resemble the patterns of tACS rather than sGVS – it was not different to dlPFC stimulation, but was to sGVS. Granted, had the sGVS vs. tACS data also reached a statistically significant difference, this claim would be more well-supported. However, the P value of 0.0797 between these two stimulations shows that it was close to doing so. If the vestibular system had been a more dominant force driving the generation of MSNA during combined stimulation, we would expect to see longer peak durations that are more akin to those generated by sGVS.
Patterns of primary and secondary MSNA peak expression relative to the phase of the sinusoidal stimulus were also analysed to support the theory. MSNA peaks appeared to have a greater correlation with the sine trough during sGVS and tACS, while combined stimulation peaks appeared to correlate with the middle of the sine wave. As such, in this analysis parameter, there may be some evidence of the vestibular system carrying out a modulatory effect during combined stimulation after all, albeit a severely diminished one. This coincides with our previous study (McCarthy et al.
2023a) in which some participants still reported mild vestibular sensations during the combined stimulation. It would be expected that MSNA peak patterning, too, would have a higher correlation with the sine trough if vestibular influences were completely abolished during combined stimulation. As this is not the case, there remains a possibility that the vestibular system is at conflict with the dlPFC, resulting in the impairment of MSNA peak timing relative to the sinusoidal stimulus. These findings, though, are somewhat dissimilar to another study which concluded that primary peaks correlated with the sine peak and secondary peaks correlated with the sine trough during sGVS (Hammam et al.
2011). There is the possibility, however, that this may be attributed to a grouping system the study utilised that only accounted for the peak and trough of the sinusoid, with no midway parameter. While the pattern results obtained in our experiments did not reach statistical significance, the absence of significance does, however, further prove that tACS of the dlPFC is equivalent to sGVS in the regulation of MSNA in almost every aspect. Furthermore, the absence of significance also provides more evidence to support the idea that the modulation produced from combined stimulation could be solely derived from the dlPFC, without any vestibular influences. With this being said, the process of determining which category each peak fell into was subjective. Categorisation was determined by visuals alone, and a few peaks proved to be ambiguous as to which group they should be placed in.
It is also worth elaborating upon the fact that our combined stimulation protocol used two sets of stimulations delivered with the sinusoidal currents in-phase. This was the same method as used previously (McCarthy et al.
2023a) to show dlPFC inhibition of sGVS-induced vestibular illusions, however it may prove interesting to conduct further study into the effects of a phase lag stimulation – especially with regards to MSNA. As Alekseichuk et al. (
2019) reported, in-phase currents applied to two brain regions will typically increase coordination, with out-of-phase currents producing opposite effects. Exactly how a current with phase lag applied to the dlPFC may affect the vestibulosympathetic reflexes, and the secondary peaks of MSNA in particular, is grounds for consideration. Should the brain regions begin to act asynchronously, secondary peaks of activity may disappear in favour of primary peaks from each brain region in an interleaved pattern (e.g. a dlPFC-generated primary peak followed by a vestibular-generated primary peak, instead of a secondary peak). This, of course, may also depend on the degree of phase lag and may, indeed, influence vestibular perceptions of sway as well (McCarthy et al.
2023a).
Moreover, in seeking to further determine the similarities of the dlPFC and vestibular system in the generation of MSNA, MSNA bursts (obtained from the RMS of the nerve signal) and heart rate (obtained from the ECG) were analysed, as it is known that the vestibular-mediated control of MSNA is linked to the cardiac cycle (Hammam and Macefield
2017). The three-minute periods before, at the start, and at the end of each stimulation protocol were analysed to generate the number of bursts of MSNA for each period. In a few cases, due to technical issues, entire three-minute periods were unavailable, but calculations were adjusted to account for this. There were no statistically significant changes in MSNA burst frequency or in heart rate across each of the time periods for each stimulation paradigm, supplementing the idea of the equality of the dlPFC to the vestibular system in MSNA modulation. While it may not be statistically significant, a look into the patterns of correlation between bursts of MSNA and heart rate reveals some possible insights. In the paired data set, there appears to be an inverse relationship between MSNA and heart rate during sGVS, which is similar to the findings of previous studies (Charkoudian et al.
2005; Hart et al.
2009) noting that, in men, an inverse relationship exists between MSNA and cardiac output (which, itself, is dependent on heart rate and stroke volume). During tACS of the dlPFC, both the paired and unpaired data sets seem to suggest a less pronounced inverse relationship, but one that exists nonetheless. However, this relationship appears to be even less pronounced during combined stimulation. Here, as with the MSNA peak relationship data, we may find evidence to support the idea of the vestibular system having a minor influence in preventing the dlPFC from having complete control during combined stimulation. Though it might be expected that vestibular influences would increase the inverse relationship, the vestibular system may, in fact, be suppressing the dlPFC to a minimal degree.
Limitations
Just as with our previous paper on vestibular perception interference via the dlPFC (McCarthy et al.
2023a), a possible confounding variable in this work is found in the ventromedial prefrontal cortex (vmPFC). In stimulating the vestibular apparatuses via sGVS, our study circumvents issues of older vestibular stimulation methods known to induce off-target effects such as fluid shifts in the body (Cui et al.
2001; Kaufmann et al.
2002) and reflexive responses from muscles of the neck (Bolton and Ray
2000). These are capable of interfering with vestibular and MSNA recordings. Yet, in stimulating the dlPFC in a similar way via tACS referenced towards the nasion, current travelled through the vmPFC, possibly activating it in the process. This is important with regards to a recent study conducted in our laboratory which showed the vmPFC to be capable of MSNA modulation to a similar extent as the dlPFC (Braun et al.
2024). Whether or not the vmPFC was activated and subsequently played a role in the observed interactions during the present study remain unknown – rendering it something worth bearing in mind.
Another point to consider is the diversity of the sample used in this project. The previously acquired data used to supplement the data set had an even spread of six male and six female subjects, but the newly acquired combined stimulation data was gathered from eight male and three female subjects. Incomplete data were collected from two of the female participants, which resulted in paired data being generated from a sample of eight males and one female. Alongside being a skewed representation of the general population, it has been previously established that arterial pressure linked to MSNA is sex-dependent (Hart et al.
2009). Men generally present with a higher mean resting MSNA and an inverse relationship between MSNA and cardiac output not seen in women (Hart et al.
2009). For these reasons, it is conceivable that the paired data on heart rate and MSNA reflect reduced changes in heart rate and higher changes in MSNA across the stimulation paradigms when compared to the population mean.
Finally, we did not include a placebo stimulation protocol, but refer the reader to our previous work (Sesa-Ashton et al.
2022) in which dlPFC modulation of MSNA was shown to be site-specific – stimulation of the motor cortex (referenced to the vertex) produced no modulation.