From the oVEMP graphs, unrectified signals from 100 trials were averaged. The first negative and positive responses were designated as n1 and p1 waves, respectively. The oVEMP response was only considered reliable if the n1 and p1 peaks were reproducible in two consecutive trace runs. Additionally, the cVEMP response was only considered reliable if the p13 and n23 peaks were reproducible in two consecutive runs of the unrectified trace. The p13-n23 responses were observed best in the unrectified trace. Initial positive and negative polarities of the waveform with peaks were termed p13 and n23 on the basis of their respective latencies. Rectified values were used since the VEMP response amplitude is significantly affected by the force of muscular contraction or stimulus intensity. After rectification (Synergy Reader software, version 20.1), peak latencies of p13 and n23 and amplitude parameters p13 and n23 were measured. The results of both runs were averaged, providing the final response from which the peak-to-peak amplitude (n1-p1) and absolute latencies (n1, p1) in oVEMP and rectified amplitude and absolute latencies (p13, n23) in cVEMP were derived. Interside differences of n1 and p1 latencies in oVEMP and p13 and n23 latencies in cVEMP were calculated. Amplitude asymmetry ratio (AR) was calculated in oVEMP and cVEMP as follows: (larger response - smaller response) / (larger response + smaller response) × 100 [4].