Article Figures & Data
Figures
- Figure 1.
Probabilistic diffusion tractography structural connectome of the human cortex. A, Group average (N = 1065) structural connectivity matrix consisting of the 360 HCP-MMPS1.0 atlas parcels organized into 10 functional networks. Raw streamline counts are fractionally scaled yielding the log probability Fpt. The white arrows highlight the diagonal which contains contralateral homologs. B, The first row of the connectivity matrix, showing connection probabilities from left V1 to all other parcels, projected onto the fsaverage template cortex. C, Single subject (100307) volume ray casting visualization of left V1-originating streamline probabilities within the skull-stripped T1-weighted structural MR volume. D, Ten functional networks, adapted from Ji et al. (2019), within HCP-MMPS1.0 atlas. These are indicated by red boxes in panel A.
- Figure 2.
Connectivity strength exponential decays with fiber tract length. A, B, Connections within the right and left hemispheres, respectively. C, Connections between the right and left hemisphere. D, All connections. Each marker represents a pair of parcels. Red traces show the least-squares exponential fit; inset are the length constant λ and r2 of this fit. Note that Fpt is log-transformed making these axes effectively semi-log.
- Figure 3.
Interindividual variability. Shown are (A) the matrix of connectivity coefficients of variation (CV) across subjects (B) pairwise CV versus fiber tract length, (C) the distribution of CV across all connections, (D) the Fpt versus fiber tract length for the connections in the highest quintile of interindividual consistency, and (E) the Fpt of right hemisphere V1–V2 connection in all subjects versus left hemisphere V1–V2 connection. In panels B, D, each marker represents a sample statistic for a connection between two parcels. E, Each marker represents an individual subject. D, The red trace show the least-squares exponential fit, and inset is the length constant λ and r2 of this fit. Note that Fpt is log-transformed making this panel’s axes effectively semi-log. In panel E, the r2 of the least-squares linear fit is reported.
- Figure 4.
Comparison of human diffusion tractography and macaque retrograde tracing connectomes. Subset of homologous parcels in the human HCP-MMPS1.0 and macaque fv91 atlas. A, Macaque group-average retrograde tracer derived structural connectome, gray indicates missing data. B, Human probabilistic diffusion tractography connectome. C, Pairwise correlation between macaque and human structural connectivity, r = 0.35, p = 0.0013.
- Figure 5.
Interhemispheric connectivity. Differential connectivity between ipsilateral and contralateral connectivity. Greater ipsilateral connectivity dominates and is indicated in red. Parcel-pairs with greater contralateral connectivity than ipsilateral are blue. The green cortical patches show anatomic extent of parcel groups of notable contrast.
- Figure 6.
Contralateral homologs. Differential connectivity between contralateral homologous parcels versus the mean of all other contralateral parcels. Red indicates contralateral homologous connectivity greater than mean contralateral connectivity. Note that many language-implicated regions have relatively weak connectivity with their contralateral homologs.
- Figure 7.
Language/auditory network hyperconnectivity and left-lateralization. A, Distance-binned connectivity within the language and auditory networks compared with connectivity between the language and auditory networks and other networks, separately for the left and right hemispheres. B, Differential trace for the within-connectivity and between-connectivity in both hemispheres. In both panels, gray patches show Bonferroni-corrected bootstrapped 95% confidence intervals across subjects.
- Figure 8.
Connectivity is influenced by the cortical hierarchy. A, B, Connectivity is strongly predicted by hierarchical similarity in some networks and modestly predicted overall. A, All connectivity versus myelination difference, including within-network and across-network connections, for the left, right, and callosal connections. For both panels, each marker represents a parcel pair. B, Within-network connectivity versus myelination difference for 10 functional networks. Linear fits and correlation coefficients computed independently for the left and right hemisphere. A negative correlation indicates that parcels at similar hierarchical levels tend to be more connected. C, D, Higher order prefrontal areas are better connected. C, Histogram of correlation coefficients between areal myelination and Fpt connectivity to each parcel. Only significant coefficients after Bonferroni correction are shown. Most coefficients are negative indicating high connectivity to low-myelination (i.e., higher-order) areas. D, Significant negative coefficients (red) map onto bilateral prefrontal cortex. Only the bilateral DVT and V6A are show positive significant correlations (blue).
- Figure 9.
Probabilistic dMRI more closely resembles CCEPs than rs-fMRI. A, Connectivity matrices for probabilistic dMRI tractography, CCEP, and rs-fMRI. For CCEPs missing data has been colored gray and pre-log zero-strength connections black. B, Correlations among the three modalities. The least-squares linear fit is shown in red. C, Non-zero pairwise connection strength distributions. Note that rs-fMRI connectivity values, which are not log-transformed, display two modes, separated at 0.0014. D, Cortical parcels displaying lower (left) and higher (right) modes of rs-fMRI connectivity.
- Figure 10.
Network theoretic differences between the connectivity modalities. Binarized network metrics after thresholding by edge weight (connectivity strength).
Tables
Connectome Features Probabilistic methodology sensitive to weak connections yielding a fully-populated, un-thresholded connectome Cortex parcellated into the standardized, relatively dense, and functionally relevant HCP-MMP1.0 atlas Large normative sample size (N = 1065) Enables comparison with other measures in the WU-Minn HCP and other cohorts Idx. Parcel Orig. Network Idx. Parcel Orig. Network Idx. Parcel Orig. Network 1 V1 1 Cingulo-opercular 61 46 84 Cingulo-opercular 121 IP1 145 Frontoparietal 2 ProS 121 Visual 62 9-46d 86 Cingulo-opercular 122 PFm 149 Frontoparietal 3 DVT 142 Visual 63 43 99 Cingulo-opercular 123 p10p 170 Frontoparietal 4 MST 2 Visual 64 PFcm 105 Cingulo-opercular 124 p47r 171 Frontoparietal 5 V6 3 Visual 65 PoI2 106 Cingulo-opercular 125 A1 24 Auditory 6 V2 4 Visual 66 FOP4 108 Cingulo-opercular 126 52 103 Auditory 7 V3 5 Visual 67 MI 109 Cingulo-opercular 127 RI 104 Auditory 8 V4 6 Visual 68 FOP1 113 Cingulo-opercular 128 TA2 107 Auditory 9 V8 7 Visual 69 FOP3 114 Cingulo-opercular 129 PBelt 124 Auditory 10 V3A 13 Visual 70 PFop 147 Cingulo-opercular 130 MBelt 173 Auditory 11 V7 16 Visual 71 PF 148 Cingulo-opercular 131 LBelt 174 Auditory 12 IPS1 17 Visual 72 PoI1 167 Cingulo-opercular 132 A4 175 Auditory 13 FFC 18 Visual 73 FOP5 169 Cingulo-opercular 133 7m 30 Default mode 14 V3B 19 Visual 74 PI 178 Cingulo-opercular 134 POS1 31 Default mode 15 LO1 20 Visual 75 a32pr 179 Cingulo-opercular 135 23d 32 Default mode 16 LO2 21 Visual 76 p24 180 Cingulo-opercular 136 v23ab 33 Default mode 17 PIT 22 Visual 77 PEF 11 Dorsal attention 137 d23ab 34 Default mode 18 MT 23 Visual 78 7PL 46 Dorsal attention 138 31pv 35 Default mode 19 LIPv 48 Visual 79 MIP 50 Dorsal attention 139 a24 61 Default mode 20 VIP 49 Visual 80 LIPd 95 Dorsal attention 140 d32 62 Default mode 21 PH 138 Visual 81 6a 96 Dorsal attention 141 p32 64 Default mode 22 V6A 152 Visual 82 PFt 116 Dorsal attention 142 10r 65 Default mode 23 VMV1 153 Visual 83 AIP 117 Dorsal attention 143 47m 66 Default mode 24 VMV3 154 Visual 84 PHA3 127 Dorsal attention 144 8Av 67 Default mode 25 V4t 156 Visual 85 TE2p 136 Dorsal attention 145 8Ad 68 Default mode 26 FST 157 Visual 86 PHT 137 Dorsal attention 146 9m 69 Default mode 27 V3CD 158 Visual 87 PGp 143 Dorsal attention 147 8BL 70 Default mode 28 LO3 159 Visual 88 IP0 146 Dorsal attention 148 9p 71 Default mode 29 VMV2 160 Visual 89 55b 12 Language 149 10d 72 Default mode 30 VVC 163 Visual 90 PSL 25 Language 150 47l 76 Default mode 31 4 8 Somatomotor 91 SFL 26 Language 151 9a 87 Default mode 32 3b 9 Somatomotor 92 STV 28 Language 152 10v 88 Default mode 33 5m 36 Somatomotor 93 44 74 Language 153 10pp 90 Default mode 34 5L 39 Somatomotor 94 45 75 Language 154 OFC 93 Default mode 35 24dd 40 Somatomotor 95 IFJa 79 Language 155 47s 94 Default mode 36 24dv 41 Somatomotor 96 IFSp 81 Language 156 EC 118 Default mode 37 7AL 42 Somatomotor 97 STGa 123 Language 157 PreS 119 Default mode 38 7PC 47 Somatomotor 98 A5 125 Language 158 H 120 Default mode 39 1 51 Somatomotor 99 STSda 128 Language 159 PHA1 126 Default mode 40 2 52 Somatomotor 100 STSdp 129 Language 160 STSvp 130 Default mode 41 3a 53 Somatomotor 101 TPOJ1 139 Language 161 TGd 131 Default mode 42 6d 54 Somatomotor 102 TGv 172 Language 162 TE1a 132 Default mode 43 6mp 55 Somatomotor 103 RSC 14 Frontoparietal 163 TE2a 134 Default mode 44 6v 56 Somatomotor 104 POS2 15 Frontoparietal 164 PGi 150 Default mode 45 OP4 100 Somatomotor 105 7Pm 29 Frontoparietal 165 PGs 151 Default mode 46 OP1 101 Somatomotor 106 8BM 63 Frontoparietal 166 PHA2 155 Default mode 47 OP2-3 102 Somatomotor 107 8C 73 Frontoparietal 167 31pd 161 Default mode 48 FOP2 115 Somatomotor 108 a47r 77 Frontoparietal 168 31a 162 Default mode 49 Ig 168 Somatomotor 109 IFJp 80 Frontoparietal 169 25 164 Default mode 50 FEF 10 Cingulo-opercular 110 IFSa 82 Frontoparietal 170 s32 165 Default mode 51 5mv 37 Cingulo-opercular 111 p9-46v 83 Frontoparietal 171 STSva 176 Default mode 52 23c 38 Cingulo-opercular 112 a9-46v 85 Frontoparietal 172 TE1m 177 Default mode 53 SCEF 43 Cingulo-opercular 113 a10p 89 Frontoparietal 173 PCV 27 Multimodal 54 6ma 44 Cingulo-opercular 114 11l 91 Frontoparietal 174 TPOJ2 140 Multimodal 55 7Am 45 Cingulo-opercular 115 13l 92 Frontoparietal 175 TPOJ3 141 Multimodal 56 p24pr 57 Cingulo-opercular 116 i6-8 97 Frontoparietal 176 PeEc 122 Multimodal 57 33pr 58 Cingulo-opercular 117 s6-8 98 Frontoparietal 177 TF 135 Multimodal 58 a24pr 59 Cingulo-opercular 118 AVI 111 Frontoparietal 178 Pir 110 Orbito-affective 59 p32pr 60 Cingulo-opercular 119 TE1p 133 Frontoparietal 179 AAIC 112 Orbito-affective 60 6r 78 Cingulo-opercular 120 IP2 144 Frontoparietal 180 pOFC 166 Orbito-affective The Idx indices refer to the parcel order in Figure 1A. The Orig. indices refer to the original parcel order presented in Glasser et al. (2016). All indices refer to the left hemisphere, adding 180 yields the homologous right hemisphere indices.
Location Data structure Test or analysis N Uncertainty [CI95%] Extended Data Fig. 1-1D Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 64,620
64,620
64,620
64,620λ = 23.8 [23.5, 24.0]
λ = 22.8 [22.7, 22.9]
λ = 22.2 [22.1, 22.2]
λ = 23.4 [23.3, 23.6]Fig. 2A Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 16,110 λ = 23.1 [22.8, 23.3] Fig. 2B Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 16,110 λ = 23.9 [23.7, 24.2] Fig. 2C Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 32,400 λ = 32.8 [32.5, 33.0] Fig. 2D Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 64,620 λ = 23.4 [23.3, 23.6] Extended Data Fig. 2-2B Gaussian predictor
Gaussian responseLinear correlation 1065 r = −0.14 [−0.20, −0.08] Fig. 3D Gaussian predictor
Exponential responseNonlinear regression (iterative optimization) 12,924 λ = 27.8 [27.4, 28.2] Fig. 3F Gaussian predictor
Gaussian responseLinear correlation 1065 r = 0.70 [0.67, 0.73] Fig. 4C Gaussian predictor
Gaussian responseLinear correlation 80 r = 0.35 [0.14, 0.53] Fig. 8A Gaussian predictor
Gaussian responseLinear correlation 16,110
16,110
32,400r = −0.10 [−0.12, −0.09]
r = −0.12 [−0.13, −0.10]
r = −0.11 [−0.12, −0.10]Fig. 8B Gaussian predictor
Gaussian responseLinear correlation 351
351
66
91
231
231
28
780
780
10r = −0.17 [−0.27, −0.06]
r = −0.13 [−0.23, −0.02]
r = −0.41 [−0.60, −0.19]
r = −0.26 [−0.44, −0.06]
r = −0.30 [−0.42, −0.18]
r = −0.30 [−0.40, −0.17]
r = −0.56 [−0.77, −0.24]
r = −0.12 [−0.19, −0.05]
r = −0.17 [−0.24, −0.10]
r = −0.74 [−0.93, −0.20]Fig. 9B Gaussian predictor
Gaussian responseLinear correlation 19,667
19,667
64,620r = 0.43 [0.42, 0.44]
r = 0.23 [0.21, 0.24]
r = 0.06 [0.05, 0.07]Extended Data Fig. 9-1 Gaussian predictor
Gaussian responseLinear correlation 8483
8483
16,110
8370
8370
16,110r = 0.42 [0.40, 0.44]
r = 0.22 [0.20, 0.24]
r = 0.06 [0.05, 0.07]
r = 0.40 [0.38, 0.42]
r = 0.22 [0.20, 0.24]
r = 0.11 [0.10, 0.13]Where multiple uncertainties are listed for a figure panel, they correspond to the statistics read left-to-right, top-to-bottom in that panel. For Figure 8B, only uncertainties for significant correlations are listed. Uncertainties for Figures 6-8, 10 are not shown. Extended Data Figure 6-1 contains bootstrapped 95% confidence intervals for the 180 means shown in Figure 6, n = 179. Figure 7 shows bootstrapped 95% confidence intervals in gray; the values of these intervals for all distance bins are available in the figure source data at https://doi.org/10.5281/zenodo.4060485. For Figure 10, means across shuffled matrices are only necessary to account for arbitrary ordering among tied edge weights, and the bootstrapped 95% confidence intervals for these means are vanishingly small. The values of these intervals at all network densities are also included in the figure source data. For nonlinear regressions, confidence intervals are estimated using R−1, the inverse R factor from QR decomposition of the Jacobian, the degrees of freedom for error, and the root mean squared error. For linear correlations, the confidence intervals are based on an asymptotic normal distribution of 0.5*log((1+r)/(1–r)), with an approximate variance equal to 1/(N – 3). For descriptive statistics, e.g., means, empirical 95% confidence intervals are estimated by bootstrapping with 2000 iterations.
Extended Data Figure 1-1
Comparison of normalization methods. Shown are the (A) connectivity matrices, (B) distributions of pairwise connectivity, (C) the pre-log distribution of Fpt (D) relationships between connectivity and fiber tract length for four normalization methods. Download Figure 1-1, TIF file.
Extended Data Figure 2-1
Alternative models for fitting connectivity strength as a function of fiber tract length. Each gray marker shows the average pair-wise Fpt between two parcels and fiber tract length between them, as also shown in Figure 2D. The colored traces show maximum likelihood estimates for several listed functional forms. The AIC, AICc, and aBIC columns contain the Akaike, corrected Akaike, and Bayesian information criteria, respectively. While the Gaussian fits explain slightly more variance and have a slightly lower AIC than the exponential fit, the exponential has fewer parameters and is consistent with histological non-human primate evidence (Markov et al., 2013; Donahue et al., 2016; Theodoni et al., 2020). Download Figure 2-1, TIF file.
Extended Data Figure 2-2
Effect of motion during the dMRI scan. A, Time course of displacement relative to initial position for one subject (996782). The six runs of the HCP dMRI protocol can be seen. B, Exponential fall-off coefficient λ is only modestly affected by motion, r = 0.140, p = 4.6E-6. Each marker represents a subject. Download Figure 2-2, TIF file.
Extended Data Figure 6-1
Differential connectivity between contralateral homologous parcels versus the mean of all other contralateral parcels. Confidence intervals are Bonferroni-corrected for multiple comparisons. Download Figure 6-1, DOCX file.
Extended Data Figure 8-1
Myelination difference connectivity matrix. This provides an estimate for the difference in hierarchical level between cortical parcels. Values have been fractionally scaled. Note that the color scale has been reversed when compared to Figure 1, as |Δmyelination| is inversely proportional to connectivity. Download Figure 8-1, TIF file.
Extended Data Figure 8-2
Pearson correlations between the Fpt from each left hemisphere parcel to all others and the target parcels’ myelination indices; p values are Bonferroni-corrected for multiple comparisons. Download Figure 8-2, DOCX file.
Extended Data Figure 8-3
Pearson correlations between the Fpt from each right hemisphere parcel to all others and the target parcels’ myelination indices; p values are Bonferroni-corrected for multiple comparisons. Download Figure 8-3, DOCX file.
Extended Data Figure 9-1
Within-hemisphere comparison of probabilistic dMRI tractography, CCEP, and rs-fMRI connectivity. For the left and right hemisphere, the distribution of pairwise non-zero connection strengths and correlations among the three modalities are shown. The least-squares linear fit is shown in red. All within-hemisphere findings are concordant with the overall findings, shown in Figure 9. Download Figure 9-1, TIF file.
In this issue
