Skip to main content
. 2025 Aug 12;19:1639864. doi: 10.3389/fnins.2025.1639864

Figure 2.

Four subject-specific graph pairs labeled A-H. Box plots A, C, E, and G show error ẽ(θ) across θ for subjects S5, S8, S6, and S13, respectively; bar plots B, D, F, and H show corresponding bias b(θ). S5 shows symmetrical errors matching frame direction and mostly positive bias. S8 also shows symmetry, but errors reverse sign near ±30°, with bias shifting from positive to negative at large angles. S6 shows asymmetry: errors follow θ for positive values, but reverse for negative; bias is positive for θ > 0, switching negative for θ < 0. S13 shows the opposite pattern of S6.

Examples of distribution of RFT alignment error (A, C, E, G) and the corresponding bias (B, D, F, H) as a function of frame tilt angle θ. The data is organized by subject: the first row shows Subject S5, the second row Subject S8, the third row Subject S6, and the fourth row Subject S13. Panels (A) and (C) show rather symmetric error curves, in contrast to the asymmetric angular dependence seen in (E) and (G). Panel (B) shows the bias for the subject who predominantly exhibits only the direct RFT effect. In contrast, panel (D) shows that Subject S8 exhibits both direct and indirect effects. For Subjects S6 and S13, the indirect effect is observed only for counterclockwise (F) and clockwise (H) rotations, respectively.