Table 4.

Simulation results for a 3D/1U staircase procedure with no vestibular bias (μ) and a physiological noise standard deviation (σ) of 1 for four alternative assumptions: 1) an initial stimulus of 1σ instead of 8 σ, 2) a psychometric function and fitting procedure based on a logistic distribution instead of a Gaussian distribution, 3) an alternate ending criteria that terminated the staircase procedure after 5 minima, and 4) another ending criteria that terminated the adaptive sampling procedure when the CV (see text) on the σ parameter estimate reached 0.25.

	GLM: μ̂	BRGLM: μ	GLM: σ̂	BRGLM: σ̂
1) Initial stimulus at 1σ (n = 50)	0.00±(0.21)	0.00±(0.21)	0.94±(0.24)	1.00±(0.25)
2) Logistic distribution (n = 50)	0.00±(0.23)	0.00±(0.22)	0.93±(0.27)	1.02±(0.29)
3) End criteria: 5th minima (n = 52±9)	0.00±(0.25)	0.00±(0.23)	0.93±(0.26)	1.01±(0.27)
4) End criteria: CV (n = 57±10)	0.00±(0.19)	0.00±(0.19)	0.94±(0.23)	0.99±(0.23)

Category I Bias (<10% of stdev)