The study investigated how a non-invasive neurostimulation tool, transcutaneous auricular vagus nerve stimulation (taVNS), affects autonomic function during visually-induced motion sickness. Participants were given taVNS and sham, in separate instances, while motion sickness was visually-induced, simultaneous with electrocardiogram (ECG) acquisition. ECG morphology and variability were quantified using symmetric projection attractor reconstruction (SPAR). It was found that taVNS-induced reductions in peak theta density correlated with SSQ nausea scores when compared to sham, while taVNS-induced reductions in maximal density was linked with motion sickness susceptibility. Additionally, machine learning models trained on ECG-derived attractor features were able to differentiate between taVNS or sham administration with an area under the receiver operating characteristic curve of 0.81. These findings indicate that there is much potential for taVNS therapy response detection, as well as adaptive taVNS, using machine learning models. The main limitations of the study were that only a small amount of attractor measures were utilized, interpretations may be skewed due to a female-dominant participant pool, and that optimal taVNS dosages were not identified.