The effect of transcutaneous auricular vagus nerve stimulation on HRV in healthy young people

Fig 1. Ear stimulation targets used of taVNS or staVNS. The stimulation target of taVNS was at the A site (tragus, active site), which is rich in vagus nerve distribution. Whereas the stimulation target of staVNS was at the C site (earlobe, control site), where was no distribution of vagus nerve [25].

Fig 2. Procedure for study 1. Fourteen participants were randomly assigned to receive either taVNS or staVNS at the beginning and then another opposite stimulation three days later. Each round consisted of three phases: 5min baseline, 5min stimulation, 5min recovery.

Fig 3. Procedure for study 2. Twenty participants were recruited. Each of them was subjected to taVNS with four different parameters. The order of the parameters was randomly customized. Each round of stimulation consisted of four phases: baseline, stimulation, recovery1st and rcovery2nd.

Fig 4. The response of autonomic function measured by HRV in taVNS or staVNS conditions over time: (A) HF, (B) RMSSD, (C) pRR50, (D) SDRR. In taVNS, the measurements of HF, RMSSD, PRR50 and SDRR were significantly higher than those in staVNS.

Fig 5. Time course analysis for HF (A) and RMSSD (B) by the taVNS or staVNS conditions. Measurements of vagal tone: RMSSD and HF significantly increased from baseline to stimulation and recovery period in taVNS condition, whereas HF significantly decreased from baseline to stimulation period in staVNS condition. Post hoc simple effect analysis of time with Bonferroni corrected, *p<0.0017.

Fig 6. The baseline LF/HF ratio and TP can predict response to taVNS: (A) Higher baseline LF/HF ratio was associated with greater LF/HF ratio decreases (R2 = 0.691, p = 0.0001), (B) lower baseline TP was associated with greater increases in TP (R 2 = 0.459, p = 0.004).

Fig 7. Differences in stimulation intensity between various parameters. The 20Hz 250us parameter was significant weaker than the 5Hz 50us parameter (p = 0.0001), the 5Hz 250us parameter (p = 0.0017), as well as the 20Hz 50us parameter (p = 0.0009). In addition, the 5Hz 250us parameter was significant weaker than the 5Hz 50us parameter (p = 0002). Statistical analysis by Friedman test with bonferroni corrected, *p<0.0083.

Fig 8. Stimulation response by parameter over time. TaVNS significantly increased the measurement of HF during stimulation and recovery 1st than baseline period, but there was not significantly different between various parameters. Statistical analysis by two-way repeated measures ANOVA.