Article Figures & Data
Figures
- Figure 1.
Schematic illustration of the three trial types from experiment 1 (1O = one orientation; 3O = three orientations; 1O1C1L = one orientation, one color, one luminance). Dotted circles indicate the other possible stimulus presentation locations, but were not presented during experiment. Experiment 2 did not include 1O1C1L trials.
- Figure 2.
BOLD signal intensity results from three ROIs in experiment 1. A, Trial-averaged BOLD signal. Dots below the x-axis indicate significance versus baseline; dots above the plots indicate significant differences between trial types. B, Within-subject correlations (ANCOVAs) between delay-period BOLD signal intensity and behavioral precision of recall. In each plot, data from each subject are portrayed in a different color. The “1,” “3,” and “1-1-1” symbols indicate individual values in the 1O, 3O, and 1O1C1L tasks, respectively. Thus, for example, the left-most plot illustrates that on 3O trials, delay-period BOLD signal in IPS is high and precision is low, whereas on 1O trials, delay-period BOLD signal in IPS is low and precision is high. The slope of the lines indicates the tendency at the group level, and the offsets between lines indicate where different individuals sit within this space.
- Figure 3.
Swap error rates for each subject in the fMRI session of experiment 2. Low swap-error subjects (dark gray) and high swap-error subjects (light gray) were defined via median split.
- Figure 4.
Orientation recall specificity effect in the occipital ROI, at TR10 on 3O trials (experiment 2). A, Neural reconstructions for probed and non-probed orientations in the two groups. Asterisks indicate significant reconstructions at p < 0.05. B, Distributions of bootstrapping estimates of orientation recall-specificity effects. The recall specificity effects were identified by the recall specificity difference between low and high swap error group. Asterisk indicates significant difference between the two distributions (p < 0.05).
- Figure 5.
Location recall specificity effect in the occipital ROI, at TR10 on 3O trials (experiment 2). A, Neural reconstructions for probed and non-probed locations in the two groups. Asterisks indicate significant reconstructions (p < 0.05), and triangle indicates trend-level evidence for a significantly negative-going reconstruction (0.05 < p < 0.1). B, Distributions of bootstrapping estimates of location recall-specificity effects in the two groups. Triangle indicates trend-level evidence for a difference between the two distributions (0.05 < p < 0.1).
- Figure 6.
Time course of IEM reconstructions of unshifted data from 3O trials. A, Reconstructions of the orientation of the probed stimulus broken out by swap-error group, in the three ROIs. B, Reconstruction of the location of the probed stimulus broken out by swap-error group, in the three ROIs. Bars along the top indicate values statistically different from baseline, with color corresponding to swap-error group; * indicates statistical difference between swap-error group at that TR (not FDR-corrected). Bars along the bottom indicate the timing of trial epochs.
Tables
Descriptive data Three-factor mixture model (parameter estimates) Trial type RT(s) Response error (degree) pT pN pU κ(rad−1) Behavior-only session 1O 2.825 (0.247) 7.395 (2.360) 0.987 (0.019) N/A 0.013 (0.019) 12.899 (7.706) 3O 2.941 (0.231) 16.214 (5.889) 0.828 (0.120) 0.120 (0.116) 0.052 (0.064) 4.166 (2.230) 1O1C1L 2.687 (0.336) 10.764 (5.076) 0.895 (0.114) N/A 0.105 (0.114) 10.718 (7.004) fMRI session 1O 2.912 (0.149) 7.765 (4.219) 0.946 (0.001) N/A 0.054 (0.001) 13.748 (6.881) 3O 3.029 (0.223) 16.510 (7.308) -- -- -- -- 1O1C1L 3.009 (0.248) 12.284 (4.028) -- -- -- --
In this issue
