Acquisition curves of eyelid conditioned responses in delay and trace paradigms when CS shifted from central (Hip–CS) to peripheral (tone-CS) or vice versa.

(A, B) Learning curves of dEBC (A) and tEBC (B) for groups of experiment (square, n = 6, for both dEBC and tEBC) and control (roundness, n = 4, for both dEBC and tEBC) when CS shifted from central to peripheral. Central CS (black in A, B) was presented during first 6 (dEBC) or 12 (tEBC) sessions in stage I and paired (black square in A and B) or pseudo-paired (black roundness in A, B) with US, then CS was switched to peripheral and paired with US (space square and space roundness, A and B) in sessions 7–12 (dEBC) or 13–18 (tEBC) of stage II. (C, D) Learning curves of dEBC (C) and tEBC (D) for groups of experiment (square, n = 7, for dEBC; n = 5, for tEBC) and control (roundness, n = 4, for both dEBC and tEBC) when CS shifted from peripheral to central (C, D). Central CS (space in C, D) was presented during first 6 (dEBC) or 12 (tEBC) sessions in stage I and paired (space square in C and D) or pseudo-paired (space roundness in C, D) with US, then CS was switched to central and paired with US (black square and black roundness, C, D) in sessions 7–12 (dEBC) or 13–18 (tEBC) of stage II. Data represent mean ± SEM. A two-way repeated measures ANOVA followed by the LSD post hoc test showed that there were significant differences in the percentages of the conditioned responses (CR) between groups of experiment and control in stage II in both delay and trace paradigms, either shifting CS from central to peripheral or vice versa. [Fig 3A and 3F (1, 8) = 40.028, *p < 0.05; Fig 3B and 3F (1, 8) = 8.905, *p < 0.05; Fig 3C and 3F (1, 9) = 154.691, *p < 0.05; Fig 3D and 3F (1, 7) = 16.299, *p < 0.05]. In recognition tests of stage III in the above 4 conditions, animals were all able to recall the original CR% to CS1.