The Effect of Training on Whole-Body Seated Vertical Vibration Threshold Detection Testing Using the Levitt Algorithm

Seated vertical vibration thresholds were tested using an adaptive Levitt algorithm. All such testing raises issues concerning potential shifting of thresholds during testing as subjects improve at the task. Additional testing was done at 4 and 16 Hz to quantify the adequacy of training within the algorithm. A 3-down 1-up algorithm starting at 8 mG descended in 3 dB steps until the first error, then switched to 1 dB steps and continued for 9 more reversals, with the last 6 averaged for threshold. Stimuli were paired with intervals containing no vibration in random order. Subjects closed their eyes and were presented with sounds in earphones to indicate the stimulus intervals, and chose the interval they thought contained the stimulus. A combination of eyes closed for concentration, gradual approach to the threshold, 4 reversals before data was used, and feedback on each trial provided built-in training to avoid threshold shift. Six subjects were tested with a pre-training run, a data collection run, and a post testing confirmation to detect shifts. An additional 30 subjects were given a single test each at 4 and 16 Hz. Median thresholds for the 6 subjects were 1.5 and 1.1 mG for 4 and 16 Hz, and 1.6 and 1.2 mG for the 30 subjects, indicating no significant shift with the additional training the 6 subjects received. There was no pre-post shift for the latter subjects either.