Visual experience shapes functional connectivity between occipital and non-visual networks
- Center for Educational Science and Technology, Beijing Normal University, China
- Department of Psychological and Brain Sciences, Johns Hopkins University, United States
- Department of Psychology, Faculty of Art and Science, Beijing Normal University at Zhuhai, China
- Trinity College Institute of Neuroscience and School of Psychology, Trinity College Dublin, Ireland
Figures
Figure 1 with 9 supplements
Functional connectivity of secondary visual cortices.
(A) Violin plots show the distributions of functional connectivity (r) of secondary visual cortices (blue) to non-visual+++ sensory-motor areas (purple) and prefrontal cortices (green), averaged across three occipital, PFC, and sensory-motor regions of interest (ROIs; A1 and S1/M1) in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475). Individual dots denote mean connectivity values per participant. Gray trend lines illustrate within-participant changes across sensory-motor and prefrontal targets. Dark-blue horizontal markers indicate group averages. ROIs displayed on the left. Note that regions extend to ventral surface, not shown. See Figure 1—figure supplement 5 for the full views of three occipital ROIs. (B) Circle plots represent the connectivity of secondary visual cortices to non-visual networks, min–max normalized to [0,1], that is, as a proportion. OC: occipital cortices; MTH: math-responsive region; LG: language-responsive region; EF: executive function-responsive (response-conflict) region. Asterisks (*) denote significant Bonferroni-corrected pairwise comparisons (p < 0.05，see Results section for statistical details). Error bars represent SEM.
Figure 1—figure supplement 1
The resting-state functional connectivity matrices.
The resting-state functional connectivity was normalized to ensure comparability across different groups (sighted adults: n = 50, blind adults: n = 30, and infants: n = 475). MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region; SMC: primary somatosensory and motor cortex; lh: left hemisphere; rh: right hemisphere.
Figure 1—figure supplement 2
Functional connectivity of three secondary visual regions.
The bar graph showed the resting-state functional connectivity of three secondary visual regions to non-visual sensory-motor networks (purple) and prefrontal cortices (green) in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475), averaged across occipital, PFC, and sensory-motor regions of interest (ROIs; A1 and S1/M1). Secondary visual regions (math, language, response-conflict) by ROIs (PFC, non-visual sensory) repeated measure ANOVA were conducted in infants. A significant interaction effect was found between the visual regions and ROIs (F(2, 948) = 136.968, p < 0.001). A post hoc Bonferroni-corrected paired t-test revealed a similar connectivity pattern across the three secondary visual regions, which exhibited stronger connectivity to prefrontal regions than non-visual sensory regions. However, the largest mean difference was observed in the occipital math-responsive region, followed by the language-responsive region, with the smallest difference found in the occipital conflict-responsive region (connectivity to non-visual sensory and to PFC, occipital math: mean difference: –0.209, t(474) = –24.546, p < 0.001; occipital language: mean difference: –0.141, t(474) = –16.674, p < 0.001; occipital conflict: mean difference: –0.114, t(474) = –13.755, p < 0.001). ROIs displayed on the upper left. PFC: prefrontal cortices; OC: occipital cortices; MTH: math-responsive region; LG: language-responsive region; EF: executive function-responsive (response-conflict) region. Error bars represent SEM.
Figure 1—figure supplement 3
Infants split-half results.
We split the infants dataset into two halves (n1 = 238, n2 = 237) and conducted split-half cross-validation. The functional connectivity between the secondary visual cortices (upper left) and V1 (upper right) to non-visual sensory-motor networks (purple) and prefrontal cortices (green) was shown in the upper row for sighted adults (n = 50), blind adults (n = 30), and two independent subgroups of infants (n1 = 238, n2 = 237). The lower row displayed the functional connectivity within the hemisphere (blue) versus between hemispheres (orange) from the secondary visual areas (lower left) and V1 (lower right) to the prefrontal cortices. Error bars represent SEM.
Figure 1—figure supplement 4
Infants split-half results.
Occipito-frontal functional connectivity across different subregions of prefrontal (PFC) and occipital cortex (OCC) in sighted adults (n = 50), blind adults (n = 30), and two independent subgroups of sighted infants (infants were randomly assigned to two subgroups, n1 = 238, n2 = 237). PFC: prefrontal cortices; MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region. Error bars represent SEM.
Figure 1—figure supplement 5
Full views of the occipital regions of interest (ROIs).
Occipital math-responsive regions (red) were more active when solving math equations than comprehending sentences. Occipital language-responsive regions (blue) were more active when comprehending sentences than solving math equations; occipital executive function (response-conflict) regions were more active during response inhibition (no-go) trials than active go trials during an auditory no-go task (Kanjlia et al., 2016; Kanjlia et al., 2021; Lane et al., 2015). The occipital ROIs were defined based on group comparisons blind > sighted in a whole-cortex analysis. All three occipital ROIs were defined in the right hemisphere. Any overlapping voxels between ROIs were removed and not counted toward any ROIs.
Figure 1—figure supplement 6
Age-matched adults subgroup results.
We performed analyses in age-matched subgroups of sighted controls (n = 29, average age across scans: M = 43, SD = 13.04) and blind adults (n = 29, average age across scans: M = 43.24, SD = 15.75). The infant sample included 475 participants (n = 475). The functional connectivity between the secondary visual cortices (upper left) and V1 (upper right) to non-visual sensory-motor networks (purple) and prefrontal cortices (green) is shown in the upper row. The middle row displays the functional connectivity within the hemisphere (blue) versus between hemispheres (orange) from the secondary visual areas (lower left) and V1 (lower right) to the prefrontal cortices. The lower row displays the occipito-frontal functional connectivity across different subregions of the prefrontal (PFC) and occipital cortex (OCC). MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region. Error bars represent SEM.
Figure 1—figure supplement 7
Results of the dataset excluding infants with radiology scores of 4 or 5.
We perform our analysis on the dataset that excluded the infants who had a radiology score of 4 or 5 and found that the results remained the same. Sample sizes were sighted adults (n = 50), blind adults (n = 30), and infants (n = 436) after excluding 39 infants with radiology scores of 4 or 5. The functional connectivity between the secondary visual cortices (upper left) and V1 (upper right) to non-visual sensory-motor networks (purple) and prefrontal cortices (green) is shown in the upper row. The middle row displays the functional connectivity within the hemisphere (blue) versus between hemispheres (orange) from the secondary visual areas (lower left) and V1 (lower right) to the prefrontal cortices. The lower row displays the occipito-frontal functional connectivity across different subregions of the prefrontal (PFC) and occipital cortex (OCC). MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region. Error bars represent SEM.
Figure 1—figure supplement 8
Examples of region of interest (ROI) alignment on individual functional images.
The top row shows two blind adults, and the bottom row shows two infants. ROIs are color-coded as follows: A1 and S1/M1 in red, math-responsive frontal and occipital regions in cyan, language-responsive frontal and occipital regions in pink, and response-conflict responsive frontal and occipital regions in blue.
Figure 1—figure supplement 9
Results of datasets excluding adults with signal dropout.
We performed our analysis on the datasets excluding the adult participants who showed signal dropout in one region of interest (ROI; one sighted adult and two blind adults) and found that the results remained the same. Sample sizes were sighted adults (n = 49), blind adults (n = 28), and infants (n = 475). The functional connectivity between the secondary visual cortices (upper left) and V1 (upper right) to non-visual sensory-motor networks (purple) and prefrontal cortices (green) is shown in the upper row. The middle row displays the functional connectivity within the hemisphere (blue) versus between hemispheres (orange) from the secondary visual areas (lower left) and V1 (lower right) to the prefrontal cortices. The lower row displays the occipito-frontal functional connectivity across different subregions of the prefrontal (PFC) and occipital cortex (OCC). MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region. Error bars represent SEM.
Figure 2 with 2 supplements
Functional connectivity of primary visual cortices (V1).
Violin plots show the distributions of functional connectivity (r) of V1 to non-visual sensory-motor areas (purple) and prefrontal cortices (green), averaged across three PFC regions of interest (ROIs) and sensory-motor ROIs (S1/M1 and A1) in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475). Individual dots denote mean connectivity values per participant. Gray trend lines illustrate within-participant changes across sensory-motor and prefrontal targets. Dark-blue horizontal markers indicate group averages. Asterisks (*) denote significant Bonferroni-corrected pairwise comparisons (p < 0.05). Cross (†) denotes marginal difference in Bonferroni-corrected pairwise comparisons (0.05 < p < 0.1, see Results section for statistical details). Error bars represent SEM.
Figure 2—figure supplement 1
The correlation of discrepancy in connectivity of visual cortex with age.
The scatter plot showed the correlation of discrepancy in connectivity of visual cortex (secondary visual cortices (left) and V1 (right)) to non-visual sensory areas and to prefrontal cortex with age after birth in infants (n = 475). Data points represent individual participants.
Figure 2—figure supplement 2
Preterm and term infant results.
We compared the results for preterm (n = 90) and term infants (n = 385) and found similar outcomes. Sample sizes were sighted adults (n = 50), blind adults (n = 30), and infants (n = 475). The functional connectivity between the secondary visual cortices (upper left) and V1 (upper right) to non-visual sensory-motor networks (purple) and prefrontal cortices (green) is shown in the upper row. The middle row displays the functional connectivity within the hemisphere (blue) versus between hemispheres (orange) from the secondary visual areas (middle left) and V1 (middle right) to the prefrontal cortices. The lower row displays the occipito-frontal functional connectivity across different subregions of the prefrontal (PFC) and occipital cortex (OCC). MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region. Error bars represent SEM.
Figure 3
Within hemisphere versus between hemisphere functional connectivity.
A bar graph shows within hemisphere (blue) and between hemisphere (orange) functional connectivity (r coefficient of resting-state correlations) of secondary visual (left) and V1 (right) to prefrontal cortices in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475). Blind adults show a larger difference than any of the other groups. Asterisks (*) denote significant Bonferroni-corrected pairwise comparisons (p < 0.05, see Results section for statistical details). Error bars represent SEM.
Figure 4 with 2 supplements
Occipito-frontal functional connectivity.
The bar graph shows across-functional connectivity of different subregions of prefrontal (PFC) and occipital cortex (OCC) in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475). Subregions (regions of interest) were defined based on task-based responses in a separate dataset of sighted (frontal) and blind (frontal and occipital) adults (Kanjlia et al., 2016; Kanjlia et al., 2021; Lane et al., 2015). PFC/OCC-MATH: math-responsive regions were more active when solving math equations than comprehending sentences. PFC/OC-LANG: language-responsive regions were more active when comprehending sentences than solving math equations (Kanjlia et al., 2016; Kanjlia et al., 2021; Lane et al., 2015). In blind adults, these regions show biases in connectivity related to their function, that is, language-responsive PFC is more correlated with language-responsive OCC. No such pattern is observed in infants. Asterisks (*) denote significant Bonferroni-corrected pairwise comparisons (p < 0.05, see Results section for statistical details). See Figure 4—figure supplement 2 for connectivity matrix. Error bars represent SEM.
-
Figure 4—source data 1
Post-hoc Bonferroni-corrected paired t-test for the connectivity between occipital regions to prefrontal regions in infants.
- https://cdn.elifesciences.org/articles/93067/elife-93067-fig4-data1-v1.xlsx
Figure 4—figure supplement 1
Occipito-frontal functional connectivity across different subregions of prefrontal (PFC) and occipital cortex (OCC) insighted adults (n = 50), blind adults (n = 30), and infants (n = 475).
Subregions (regions of interest) were defined based on task-based responses in a separate dataset of sighted (frontal) and blind (frontal and occipital) adults (Kanjlia et al., 2016; Kanjlia et al., 2021; Lane et al., 2015). PFC/OC-MTH math-responsive regions were more active when solving math equations than comprehending sentences. PFC/OC-LG language-responsive regions were more active when comprehending sentences than solving math equations; EF: executive function (response-conflict) regions were more active during response inhibition (no-go) trials than active go trials during an auditory no-go task (Kanjlia et al., 2016; Kanjlia et al., 2021; Lane et al., 2015). In blind adults (top right), these regions show biases in connectivity related to their function, i.e., language-responsive PFC is more correlated with language-responsive OCC. No such pattern is observed in infants. Error bars represent SEM.
Figure 4—figure supplement 2
The resting-state functional connectivity matrices between secondary visual areas to prefrontal regions in sighted adults (n = 50), blind adults (n = 30), and infants (n = 475).
PFC: prefrontal cortices; OC: occipital cortices; MTH: math-responsive region; LG: language-responsive region; EF: executive function (response-conflict) region.
Additional files
-
Supplementary file 1
ROI-wise noise ceiling statistics across groups.
- https://cdn.elifesciences.org/articles/93067/elife-93067-supp1-v1.xlsx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/93067/elife-93067-mdarchecklist1-v1.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Visual experience shapes functional connectivity between occipital and non-visual networks
eLife 13:RP93067.
https://doi.org/10.7554/eLife.93067.4
