Article Figures & Data
Figures
- Figure 1.
Experimental design and the pSFT model. A, Within each scan (9 in total), participants completed two blocks of each condition in a pseudorandom order. B, The visual display was split into an attended hemifield (task-relevant) and a probe hemifield (task-irrelevant). C, For each voxel, condition-specific measured BOLD response was compared with a synthesized BOLD response (green trace) produced by the best prediction of pSFT parameters. D, Estimated pSFT peak (top row) and bandwidth (bottom row) follow expected trends in every condition: decreased pSFT peak and increased bandwidth with pRF eccentricity and visual area. Each point represents a voxel that survived selection criteria. Voxels from all subjects are presented. Each subject has a unique shading in the scatter plots. Green represents estimates from the “Attend Fixation” condition, red from the “Attend HSF” condition, and blue from the “Attend LSF” condition. V1, n = 802; V2, n = 735; V3, n = 398. See also Extended Data Figure S1A and Extended Data Table S1.
- Figure 2.
Attentional modulation of pSFT. A, The key predictions afforded by the pSFT model are an increase/decrease in the pSFT peak and/or pSFT bandwidth. To facilitate the interpretation of attentional modulation of pSFT, the legend in B demonstrates the hypothesized modulatory strategies when the change in peak and bandwidth are plotted against one another. To the right are caricatures of each strategy (Quadrants I–IV). In C, each dot represents a voxel's change in peak (x-axis) and bandwidth (y-axis) in the “Attend LSF” condition (blue, top row) and in the “Attend HSF” condition (red, bottom row). Each subject has a unique shading in the scatter plots. Histograms appended to the north and east walls of each plot reveal the distribution of AMI across all voxels in an ROI. Vertical and horizontal arrows above the histograms represent the group mean (n = 8) for change in peak and bandwidth, respectively. A dot at the intersection of these arrows is included for visibility. The repeated measures correlation coefficient and p-value are reported in the top right of each scatter plot. At the group level, the change in peak significantly differed between conditions in every ROI (V1 and V2 ps < 0.05; V3 p < 0.01). V1, n = 802; V2, n = 735; V3, n = 398. See also Table 1.
- Figure 3.
Attentional modulation of pSFT is preference-dependent. Voxel-wise changes in peak (A) and bandwidth (B) are reported as a percentage and as a function of the octave distance (i.e., dissimilarity) between the voxel's baseline pSFT peak and the attended SF (blue, LSF; red, HSF). Each subject has a unique shading in the scatter plots. Dashed lines represent the group mean (n = 8). Bonferroni-corrected *p < 0.05 and **p < 0.01. V1, n = 802; V2, n = 735; V3, n = 398. See also Figure 4 and Table 2.
- Figure 4.
Attentional modulation of pSFT is dependent on the attended SF. A, Box plot for baseline dissimilarity versus change in peak slopes. B, Box plot for baseline dissimilarity versus change in bandwidth slopes. Each box depicts, from bottom to top, the minimum, 25th percentile, median, 75th percentile, and maximum slope across participants. Blue represents values from the “Attend LSF” condition, while red represents values from the “Attend HSF” condition. The “+” symbol represents outliers >1.5 times the interquartile range. Statistical test results reported above each plot were performed at the group level within conditions (blue/red) and between conditions (black). Bonferroni-corrected p-values. nsp > 0.05, *p < 0.05, **p < 0.01. See also Table 2.
Tables
Attend LSF
Attend HSF
Attend LSF versus Attend HSF
Change in peak Change in bandwidth ROI Mean ± 1 SEM P 95% CI Mean ± 1 SEM p 95% CI V1 −5.65 ± 5.22 0.944 −17.99, 6.69 −6.45 ± 3.54 0.334 −14.83, 1.92 2.28 ± 4.13 1.000 −7.50, 12.05 −10.83 ± 3.64 0.062 −19.45, −2.22 — 0.024* −13.05, −2.80 — 0.087 0.60, 8.16 V2 −8.58 ± 4.16 0.234 −18.41, 1.25 −2.44 ± 3.07 1.000 −9.71, 4.82 5.76 ± 4.97 0.853 −5.99, 17.51 −7.96 ± 2.96 0.093 −14.95, −0.97 — 0.025* −23.67, −5.02 — 0.101 0.56, 10.47 V3 −7.21 ± 4.86 0.544 −18.70, 4.28 −1.21 ± 3.56 1.000 −18.30, 1.31 9.54 ± 6.65 0.583 −6.17, 25.26 −8.50 ± 4.15 0.239 −17.31, 1.20 — 0.004** −24.54, −8.96 — 0.162 −0.17, 14.73 - Table 2.
Group-level results for linear slopes between dissimilarity from attended SF and change in peak and bandwidth (AMI/oct)
Attend LSF
Attend HSF
Attend LSF versus Attend HSF
Change in peak Change in bandwidth ROI Mean ± 1 SEM P 95% CI Mean ± 1 SEM p 95% CI V1 −9.27 ± 2.27 0.014* −14.65, −3.89 3.31 ± 0.68 0.005** 1.71, 4.91 −10.89 ± 2.46 0.009** −16.70, −5.08 2.53 ± 1.18 0.211 −0.27, 5.33 — 0.041* 0.44, 2.79 — 1.000 −2.02, 3.58 V2 −6.62 ± 2.15 0.053 −11.71, −1.54 3.13 ± 0.69 0.008** 1.50, 4.76 −7.81 ± 2.84 0.085 −14.51, −1.10 2.21 ± 1.84 0.810 −2.15, 6.56 — 1.000 −4.05, 6.43 — 1.000 −2.91, 4.76 V3 −7.76 ± 1.94 0.016* −12.34, −3.17 2.75 ± 1.03 0.097 0.31, 5.19 −12.86 ± 2.14 0.002** −17.91, −7.81 3.70 ± 1.26 0.065 0.73, 6.67 — 0.028* 1.71, 8.5 — 0.784 −2.79, 0.89
Supplemental Material
