Figure 3.

The locations at which the participant stopped relative to the true target position (0,0 in the left panel) was used to derive the individual variable error (the area of the ellipse) and constant error (the distance from the center of the ellipse to the target location) for each task (allocentric and egocentric) and each condition (vOICe, TDU, and vOICeTDU) separately. The unit for the x and y position is in meters. Individual estimates for vOICe, tongue display unit (TDU), and vOICeTDU conditions for the allocentric task are shown in the top panel, while for egocentric task, the estimates are shown in the bottom panel for the same participant. The triangle labeled maximum likelihood estimation (MLE) (white) in the central panels refer to the variability predicted by the MLE model for the combined condition (vOICeTDU). The predicted estimate (MLE) was calculated for each subject, and then averaged, by entering the individual vOICe (${\sigma }_{vOICe}$) and TDU (${\sigma }_{TDU}$) measure of variable error into the equation $\frac{{\sigma }_{vOICeTDU=}^2{\sigma }_{vOICe}^2{\sigma }_{TDU}^2}{{\sigma }_{vOICe}^2+{\sigma }_{TDU}^2}$.