Figures and data in An afferent white matter pathway from the pulvinar to the amygdala facilitates fear recognition

Figures
Tables
Additional files

8 figures, 13 tables and 1 additional file

Figures

Figure 1

Download asset Open asset

Global and local probabilistic tractography reconstructions of the subcortical route to the amygdala.

Fibres reconstructed between the superior colliculus and the pulvinar are shown in (A) and between the pulvinar and amygdala in (B). Both (A) and the top row of (B) show 3D-renders of the major ROIs (amygdala in pink, pulvinar in orange, superior colliculus in green), as well as the left anterior thalamic radiation (orange) and the right corticospinal tract (pink). The reconstructed fibres for all participants are rendered by semi-transparent white streamlines. Streamlines of a single example participant are shown below in boxes. The bottom half of (B) shows a different, top-down perspective of the pulvinar to amygdala pathway, illustrating the right inferior frontal occipital fasciculus (light blue) and the left inferior longitudinal fasciculus (dark blue). (C) shows the global fibre counts (left graph) and average fibre density (right graph) for global and local tractography, respectively (SC-PUL in yellow, PUL-AMG in purple). Group average is indicated by solid white line, with dotted lines indicating the 95% confidence interval. (D) shows the average number of fibres terminating in each subregion of the pulvinar and amygdala, for both global and local tractography. 3D renders of the subregions (colour-coded to match the graphs) are shown on the left. Graphs of the number of fibres terminating in pulvinar subregions are shown for tracts connecting the superior colliculus and pulvinar (first and second graphs) and the pulvinar and amygdala (third and fourth graphs), while fibres terminating in amygdala subregions are shown for tracts connecting the pulvinar and amygdala (fifth and sixth graphs). Error bars represent 95% confidence intervals.

https://doi.org/10.7554/eLife.40766.003

Figure 2

Download asset Open asset

Relationship between behavioural performance and tractography.

(A) Histograms of scores (out of eight) for correctly recognising the emotional expression of angry, fearful, and sad faces from the Penn Recognition Test. Normal distribution function is plotted. (B) The relationships between recognition of angry (purple), fearful (red), and sad (blue) faces (x-axes) and the predicted local fibre densities for the left (top row) and right (bottom row) PUL-AMG connection (y-axes), as produced by a multivariate ANOVA. Fearful face recognition accuracy was significantly related to local fibre density for the left and right PUL-AMG connections (β and p values shown for parameter estimates from multivariate ANOVA, *p < .05).

https://doi.org/10.7554/eLife.40766.004

Figure 3

Download asset Open asset

Whole-brain fMRI activation for faces and shapes.

Face activation is shown by hot colours and shape activation is shown by cool colours. MNI coordinates are shown at the bottom. Labels indicate significant (p < 0.05, whole brain FWE) activation in the superior colliculus (SC), pulvinar (PUL), amygdala (AMG), inferior occipital gyrus (IOG), and fusiform gyrus (FG). Results are overlayed on an averaged MNI152 T1 template.

https://doi.org/10.7554/eLife.40766.005

Figure 4

Download asset Open asset

Dynamic causal modelling model space and estimated probabilities.

(A) Diagram of the model space constructed for the fMRI activation to viewing faces. The top row shows various types of model designs in the Cortical family, and the bottom row shows model designs in the Dual family (which includes a subcortical route). Each model variation included different combinations of forwards and backwards connections, indicated by dashed arrows. The model numbers are shown above each model variation. Every model contained input to the inferior occipital gyrus but some models also contained input to the superior colliculus (green) only, pulvinar (orange) only, or both (black). (B) The expected (top row) probabilities and exceedance probabilities (bottom row) for each of the 102 models. The colour of the bars indicates the different type of input for that family (i.e., SC, PUL, or both), according to A). Individual model probabilities are shown on the left (including a diagram of the winning model) and the probabilities for the families are shown on the right.

https://doi.org/10.7554/eLife.40766.006

Figure 5

Download asset Open asset

The strength of individual connections from the winning dynamic causal model.

Parameter estimates of the modulatory connection strength of the winning model. Dot plots of up to 234 participants are shown with horizontal lines for the mean and vertical lines for one standard deviation. Density distribution is represented by each violin plot. The long horizontal dotted line across the graph represents the prior (set to 1), where * indicates the connection was significantly greater than 1 (p<0.05, corrected for multiple comparisons). A diagram of the winning model (left and right hemispheres are shown) is at the top right. Greyed-out connections indicate those that were not significant.

https://doi.org/10.7554/eLife.40766.007

Figure 6

Download asset Open asset

The relationship between structural and effective connectivity for the subcortical route.

(**A and B**) show 3D renders of the ROIs used in the dynamic causal modelling stage. A = anterior, p = posterior, R = right, L = left. (A) 3D-rendered tracks generated by global tractography overlaid for all 622 participants (pink/orange for PUL-AMG, green/yellow for SC-PUL). (B) Direction of information flow according to the winning dynamic causal model (exceedance probability = 98.01%) illustrated using an axial view of the network. Dotted lines indicate subcortical input/connections. (C) Positive correlation between global track count for the left and right PUL-AMG connections and the modulatory strength of the same connection in the winning effective connectivity model (with 95% CI shown).

https://doi.org/10.7554/eLife.40766.008

Appendix 1—figure 1

Download asset Open asset

Participants with above or below threshold fMRI signal in all VOIs for inclusion into DCM stage.

Threshold was set at p<0.05 uncorrected, using the [Faces – Shapes] contrast. Each column is a participant (N = 622) and each row is a different region (SC = superior colliculus, PUL = pulvinar, AMG = amygdala, FG = fusiform gyrus, IOG = inferior occipital gyrus) with two sessions each (i.e. each fMRI run) per hemisphere (L = left, R = right). Red indicates below-threshold signal and green indicates above-threshold signal. The bottom row (‘ALL’) indicates whether the participants were included or excluded from further DCM analysis, based on whether they had any below-threshold VOIs.

https://doi.org/10.7554/eLife.40766.024

Author response image 1

Download asset Open asset

Tables

Appendix 1—table 1

Global tractography fibre counts.

2 (pathway: SC-PUL, PUL-AMG)×2 (hemisphere: left, right) design. Four outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.011

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	* Pathway	(3,4)	433.286	2.842 × 10⁻⁷³*	0.413
	* Hemisphere	(3,4)	7.583	0.006 *	0.012
	* Pathway × Hemisphere	(1,616)	16.025	7.000 × 10⁻⁵*	0.025
	Left vs.Right SC-PUL					617	1.013	.311
	* Left vs. Right PUL-AMG					617	−9.785	4.070 × 10⁻²¹*
	Pathway × Head Motion	(1,616)	0.313	.576	5.090 × 10^-4 *
	Hemisphere × Head Motion	(1,616)	0.055	.813	9.100 × 10^-5 *
	Pathway × Hemisphere × Head Motion	(1,616)	0.332	.710	2.250 × 10^-4 *
No Outliers Removed	* Pathway	(1,620)	32.416	2.194 × 10⁻⁸*	.049
	* Hemisphere	(1,620)	9.175	.003 *	.015
	* Pathway × Hemisphere	(1,620)	5.807	.016 *	.009
	Left vs. Right SC-PUL					621	0.654	.514
	* Left vs. Right PUL-AMG					621	−8.904	5.906 × 10⁻¹⁸*
	Pathway × Head Motion	(1,620)	0.030	.863	4.800 × 10⁻⁵
	Hemisphere × Head Motion	(1,620)	1.155	.283	.002
	Pathway × Hemisphere × Head Motion	(1,620)	0.332	.565	.001

*p <.05

Appendix 1—table 2

Local tractography fibre counts.

2 (pathway: SC-PUL, PUL-AMG)×2 (hemisphere: left, right) design. Thirteen outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.012

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	* Pathway	(1,607)	69.586	4.930 × 10⁻¹⁶*	0.103
	Hemisphere	(1,607)	1.387	0.239	0.002
	* Pathway × Hemisphere	(1,607)	162.48	3.828 × 10⁻³³*	0.211
	* Left vs. Right SC-PUL					608	10.749	8.596 × 10⁻²⁵*
	* Left vs. Right PUL-AMG					608	−18.205	1.960 × 10⁻⁵⁹*
	Pathway × Head Motion	(1,607)	1.815	.178	.003
	Hemisphere × Head Motion	(1,607)	1.39	.239	.002
	Pathway × Hemisphere × Head Motion	(1,607)	1.425	.233	.002
No Outliers Removed	* Pathway	(1,620)	59.749	4.365 × 10⁻¹⁴*	.088
	Hemisphere	(1,620)	0.757	.385	.001
	* Pathway ×Hemisphere	(1,620)	116.962	4.316 × 10⁻²⁵*	.159
	* Left vs. Right SC-PUL					621	9.378	1.243 × 10⁻¹⁹*
	* Left vs. Right PUL-AMG					621	−17.515	4.068 × 10⁻⁵⁶*
	Pathway × Head Motion	(1,620)	1.222	.269	.002
	Hemisphere × Head Motion	(1,620)	1.511	.219	.002
	Pathway × Hemisphere × Head Motion	(1,620)	0.617	.433	.001

*p <.05

Appendix 1—table 3

Global tractography fibre counts (terminations in pulvinar subregions) for SC-PUL.

5 (cluster: inferior, medial, anterior, superior, lateral)×2 (hemisphere: left, right) design. Ten outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.013

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	Hemisphere	(1,610)	3.817	.051	.090
	* Cluster	(3,2060)	214.015	1.584 × 10⁻¹³³*	.260
	*Hemisphere × Cluster	(3,2014)	23.624	2.870 × 10⁻¹⁶*	.037
	* Left: Inferior vs. Anterior					611	4.820	1.815 × 10⁻⁶*
	* Left: Anterior vs. Medial					611	2.754	0.006 *
	* Left: Medial vs. Lateral					611	11.635	2.114 × 10⁻²⁸*
	* Left: Lateral vs. Superior					611	15.938	4.456 × 10⁻⁴⁸*
	* Right:Anterior vs. Inferior					611	4.228	2.715 × 10⁻⁵*
	* Right: Inferior vs. Lateral					611	13.982	9.574 × 10⁻³⁹*
	Right: Lateral vs. Medial					611	2.045	0.041
	* Right: Medial vs. Superior					611	21.108	1.076 × 10⁻⁷⁴*
	Hemisphere × Head Motion	(1,610)	0.620	0.431	.001
	Cluster × Head Motion	(3,2060)	0.478	0.720	.001
	Hemisphere × Cluster × Head Motion	(3,2014)	0.306	0.840	.001
No Outliers Removed	* Hemisphere	(1,620)	5.749	.017 *	.009
	* Cluster	(1,913)	52.174	2.998 × 10⁻¹⁷*	.078
	*Hemisphere × Cluster	(3,1743)	16.591	4.129 × 10⁻¹⁰*	.026
	* Left: Inferior vs. Anterior					621	4.191	3.200 × 10^{-5 *}
	Left: Anterior vs. Medial					621	2.314	.021
	* Left: Medial vs. Lateral					621	6.385	3.361 × 10^{-10 *}
	* Left: Lateral vs. Superior					621	14.235	5.357 × 10^{-40 *}
	* Right: Anterior vs. Inferior					621	3.695	2.390 × 10^{-5 *}
	* Right: Inferior vs. Lateral					621	7.619	9.590 × 10^{-14 *}
	Right:Lateral vs. Medial					621	−0.187	.852
	* Right: Medial vs. Superior					621	12.957	3.715 × 10^{-34 *}
	Hemisphere × Head Motion	(1,620)	1.555	.213	.003
	Cluster × Head Motion	(1,913)	0.159	.786	2.570 × 10⁻⁴
	Hemisphere × Cluster × Head Motion	(3,1743)	0.706	.539	.001

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected)

Appendix 1—table 4

Global tractography fibre counts (terminations in pulvinar subregions) for PUL-AMG.

5 (cluster: inferior, medial, anterior, superior, lateral)×2 (hemisphere: left, right) design. Sixty-eight outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.014

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	* Hemisphere	(1,552)	22.147	3.197 × 10⁻⁶*	.039
	* Cluster	(2,1139)	354.261	2.662 × 10⁻¹²³*	.391
	*Hemisphere × Cluster	(2,1135)	15.334	1.946 × 10⁻⁷*	.027
	* Left: Inferior vs. Medial					533	24.849	4.248 × 10⁻⁹²*
	* Left: Medial vs. Lateral					533	4.209	2.998 × 10⁻⁵*
	Left: Lateral vs. Anterior					533	2.458	.014
	* Left: Anterior vs. Superior					533	4.665	3.868 × 10⁻⁶*
	* Right:Inferior vs. Lateral					533	25.742	1.184 × 10⁻⁹⁶*
	* Right: Lateral vs. Superior					533	7.609	1.197 × 10⁻¹³*
	Right: Superior vs. Medial					533	2.044	.041
	Right: Medial vs. Anterior					533	4.133	4.141 × 10⁻⁵*
	Hemisphere × Head Motion	(1,552)	1.067	.302	.002
	Cluster × Head Motion	(2,1139)	1.885	.151	.003
	Hemisphere ×Cluster × Head Motion	(2,1135)	0.601	.553	.001
No Outliers Removed	* Hemisphere	(1,620)	33.464	1.152 × 10^{-8 *}	.051
	* Cluster	(2,1153)	320.406	3.145 × 10^{-105 *}	.341
	*Hemisphere × Cluster	(2,1390)	14.935	9.810 × 10^{-8 *}	.024
	* Left:Inferior vs. Anterior					621	23.865	8.002 × 10^{-90 *}
	Left: Anterior vs. Medial					621	3.099	.002
	* Left: Medial vs. Lateral					621	3.021	.003
	* Left:Lateral vs. Superior					621	3.835	1.380 × 10^{-4 *}
	* Right: Anterior vs. Inferior					621	25.424	2.827 × 10^{-98 *}
	* Right: Inferior vs. Lateral					621	7.909	1.190 × 10^{-14 *}
	Right: Lateral vs. Medial					621	0.657	.511
	* Right: Medial vs. Superior					621	3.925	9.600 × 10^{-5 *}
	Hemisphere × Head Motion	(1,620)	0.278	.598	4.490 × 10⁻⁴
	Cluster × Head Motion	(2,1153)	1.134	.319	.002
	Hemisphere × Cluster × Head Motion	(2,1390)	0.512	.620	.001

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected).

Appendix 1—table 5

Global tractography fibre counts (terminations in amygdala subregions) for PUL-AMG.

3 (cluster: centromedial, basolateral, superficial)×2 (hemisphere: left, right) design. Sixteen outliers removed.

https://doi.org/10.7554/eLife.40766.015

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	Hemisphere	(1,604)	1.419	.234	.002
	* Cluster	(2,1161)	80.779	1.909 × 10^{-32 *}	.118
	* Hemisphere × Cluster	(2,1171)	28.278	2.060 × 10^{-12 *}	.045
	* Left: Basolateral vs. Centromedial					605	8.741	2.279 × 10^{-17 *}
	* Left: Centromedial vs. Superficial					605	5.354	1.222 × 10^{-7 *}
	* Right: Basolateral vs. Superficial					605	11.515	7.018 × 10^{-28 *}
	* Right: Superficial vs. Centromedial					605	10.700	1.376 × 10^{-24 *}
	Hemisphere × Head Motion	(1,604)	0.586	.444	.695 × 10⁻⁴
	Cluster × Head Motion	(2,1161)	0.892	.407	.001
	Hemisphere × Cluster × Head Motion	(2,1171)	0.261	.764	4.315 × 10⁻⁴
No Outliers Removed	Hemisphere	(1,620)	2.295	.130	.004
	* Cluster	(2,1212)	79.253	2.017 × 10^{-32 *}	.113
	* Hemisphere × Cluster	(2,1201)	27.754	3.363 × 10^{-12 *}	.043
	* Left: Basolateral vs. Centromedial				621	8.806	1.284 × 10^{-17 *}
	* Left: Centromedial vs. Superficial				621	5.119	4.098 × 10^{-7 *}
	* Right: Basolateral vs. Superficial				621	11.753	6.132 × 10^{-29 *}
	* Right: Superficial vs. Centromedial				621	10.027	4.961 × 10^{-22 *}
	Hemisphere × Head Motion	(1,620)	0.014	.907	2.200 × 10⁻⁴
	Cluster × Head Motion	(2,1212)	0.515	.593	.001
	Hemisphere × Cluster × Head Motion	(2,1201)	0.164	.842	2.650 × 10⁻⁴

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected)

Appendix 1—table 6

Local tractography fibre counts (terminations in pulvinar subregions) for SC-PUL.

5 (cluster: inferior, medial, anterior, superior, lateral)×2 (hemisphere: left, right) design. Sixty-seven outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.016

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	* Hemisphere	(1,553)	23.577	1.564 × 10^{-6 *}	.041
	* Cluster	(2,1010)	873.922	6.337 × 10^{-209 *}	.612
	*Hemisphere × Cluster	(2,1176)	64.370	1.002 × 10^{-28 *}	.104
	* Left: Anterior vs. Inferior					554	22.373	1.709 × 10^{-79 *}
	* Left: Inferior vs. Medial					554	9.742	8.353 × 10^{-21 *}
	* Left:Medial vs. Lateral					554	30.710	1.089 × 10^{-121 *}
	Left:Lateral vs. Superior*					554	16.514	4.103 × 10^{-50 *}
	* Right:Anterior vs. Inferior					554	39.241	4.436 × 10^{-162 *}
	* Right: Inferior vs. Medial					554	9.157	1.022 × 10^{-18 *}
	* Right: Medial vs. Lateral					554	25.287	2.203 × 10^{-94 *}
	* Right:Lateral vs. Superior					554	15.213	6.282 × 10^{-44 *}
	Hemisphere × Head Motion	(1,553)	0.358	0.550	.001
	Cluster × Head Motion	(2,1010)	0.495	0.593	.001
	Hemisphere × Cluster × Head Motion	(2,1176)	0.604	0.557	.001
No Outliers Removed	* Hemisphere	(1,620)	23.600	2.000 × 10^{-6 *}	.037
	* Cluster	(2,1244)	831.764	1.817 × 10^{-230 *}	.573
	*Hemisphere × Cluster	(2,1354)	69.262	1.107 × 10^{-31 *}	.100
	* Left: Anterior vs. Inferior					621	22.745	9.069 × 10^{-84 *}
	* Left:Inferior vs. Medial					621	8.436	2.312 × 10^{-16 *}
	* Left:Medial vs. Lateral					621	29.562	1.448 × 10^{-120 *}
	Left:Lateral vs. Superior*					621	13.832	3.972 × 10^{-39 *}
	* Right: Anterior vs. Inferior					621	39.100	1.326 × 10^{-169 *}
	* Right:Inferior vs. Medial					621	7.436	3.463 × 10^{-13 *}
	* Right:Medial vs. Lateral					621	23.544	4.380 × 10^{-88 *}
	* Right:Lateral vs. Superior					621	12.964	3.461 × 10^{-34 *}
	Hemisphere × Head Motion	(1,620)	0.161	.688	2.60 × 10⁻⁴
	Cluster × Head Motion	(2,1244)	0.830	.437	.001
	Hemisphere × Cluster × Head Motion	(2,1354)	1.023	.365	.002

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected)

Appendix 1—table 7

Local tractography fibre counts (terminations in pulvinar subregions) for AMG-PUL.

5 (cluster: inferior, medial, anterior, superior, lateral)×2 (hemisphere: left, right) design. Seventy-eight outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.017

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	* Hemisphere	(1,542)	108.301	3.019 × 10^{-23 *}	.167
	* Cluster	(1,544)	2159.827	9.415 × 10^{-192 *}	.799
	*Hemisphere × Cluster	(1,544)	105.772	7.093 × 10^{-23 *}	.163
	* Left: Inferior vs. Lateral					543	57.694	8.996 × 10^{-234 *}
	* Left: Lateral vs. Medial					543	10.633	4.080 × 10^{-24 *}
	* Left: Medial vs. Anterior					543	12.744	9.913 × 10^{-33 *}
	* Left: Anterior vs. Superior					543	5.897	6.503 × 10^{-9 *}
	* Right: Inferior vs. Lateral					543	59.079	1.332 × 10^{-238 *}
	* Right: Lateral vs. Superior					543	27.277	7.886 × 10^{-104 *}
	* Right: Superior vs. Anterior					543	14.006	2.839 × 10^{-38 *}
	Right: Anterior vs. Medial					543	0.815	.416
	Hemisphere × Head Motion	(1,542)	4.749	.030 *	0.009
	Cluster × Head Motion	(1,544)	19.569	1.146 × 10^{-5 *}	0.035
	Hemisphere × Cluster × Head Motion	(1,544)	4.689	.031 *	0.009
No Outliers Removed	* Hemisphere	(1,620)	110.587	6.523 × 10^{-24 *}	.151
	* Cluster	(1,622)	2139.649	4.887 × 10^{-204 *}	.775
	*Hemisphere × Cluster	(1,624)	107.010	2.219 × 10^{-23 *}	.147
	* Left: Inferior vs. Lateral					621	57.120	2.174 × 10^{-249 *}
	* Left: Lateral vs. Medial					621	10.169	1.427 × 10^{-22 *}
	* Left:Medial vs. Anterior					621	10.609	2.812 × 10^{-24 *}
	Left: Anterior vs. Superior*					621	5.153	3.446 × 10^{-7 *}
	* Right: Inferior vs. Lateral					621	61.529	1.963 × 10^{-266 *}
	* Right:Lateral vs. Superior					621	24.241	7.327 × 10^{-92 *}
	Right: Superior vs. Anterior					621	10.891	2.122 × 10^{-25 *}
	Right: Anterior vs. Medial					621	0.227	.821
	Hemisphere × Head Motion	(1,620)	4.775	.029 *	.147
	Cluster × Head Motion	(1,622)	19.739	1.000 × 10^{-5 *}	.031
	Hemisphere × Cluster × Head Motion	(1,624)	4.568	.033	.007

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected)

Appendix 1—table 8

Local tractography fibre counts (terminations in amygdala subregions) for PUL-AMG.

3 (cluster: centromedial, basolateral, superficial)×2 (hemisphere: left, right) design. Thirty-seven outliers removed out of 622.

https://doi.org/10.7554/eLife.40766.018

Outliers removed	ANOVA main effects					t-tests for Simple Effects
	Effect	Df	F	P	η_p²	Df	T	P
	Hemisphere	(1,620)	511.278	5.154 × 10^{-83 *}	.452
	* Cluster	(2,989)	308.079	7.593 × 10^{-88 *}	.332
	*Hemisphere × Cluster	(2,1201)	27.754	3.363 × 10^{-12 *}	.043
	* Left: Basolateral vs. Centremedial					621	10.374	2.323 × 10^{-23 *}
	* Left: Centromedial vs. Superficial					621	19.573	8.248 × 10^{-67 *}
	* Right: Centromedial vs. Basolateral					621	25.136	1.037 × 10^{-96 *}
	* Right: Basolateral vs. Superficial					621	27.037	5.308 × 10^{-107 *}
	Hemisphere × Head Motion	(1,620)	296.624	.015 *	.009
	Cluster × Head Motion	(2,989)	0.050	4.774 × 10⁻⁸⁴	.324
	Hemisphere × Cluster × Head Motion	(2,1201)	0.917	.380	.001
No Outliers Removed	Hemisphere	(1,620)	511.278	5.154 × 10^{-83 *}	.452
	* Cluster	(2,989)	308.079	7.593 × 10^{-88 *}	.332
	*Hemisphere × Cluster	(2,1201)	27.754	3.363 × 10^{-12 *}	.043
	* Left: Basolateral vs. Centremedial					621	10.374	2.323 × 10^{-23 *}
	* Left: Centromedial vs. Superficial					621	19.573	8.248 × 10^{-67 *}
	* Right: Centromedial vs. Basolateral					621	25.136	1.037 × 10^{-96 *}
	* Right: Basolateral vs. Superficial					621	27.037	5.308 × 10^{-107 *}
	Hemisphere × Head Motion	(1,620)	296.624	.015 *	.009
	Cluster × Head Motion	(2,989)	0.050	4.774 × 10⁻⁸⁴	.324
	Hemisphere × Cluster × Head Motion	(2,1201)	0.917	.380	.001

*p <.05 (for t-tests, only * if <0.05 Bonferroni corrected)

Appendix 1—table 9

Paired t-tests between local streamline count and null distribution.

https://doi.org/10.7554/eLife.40766.019

Comparison	N	M	SEM	95%	T	Df	P
Left SC- PUL	618	1384.876	16.704	[1352.073, 1417.68]	82.907	617	<1 × 10^{-243 *}
Right SC- PUL	618	−499.555	16.430	[−531.821, –467.289]	−30.404	617	8.832 × 10^{-125 *}
Left PUL- AMG	618	351.669	8.535	[334.908, 368.431]	41.202	617	2.722 × 10^{-179 *}
Right PUL-AMG	620	577.535	10.385	[557.142, 597.929]	55.614	619	6.091 × 10^{-243 *}

*p <.05

Appendix 1—table 10

Global tractography multivariate GLM.

Fibre counts for left and right SC-PUL and PUL-AMG entered as dependent variables. Recognition scores (out of eight) for fearful, sad, and angry expressions entered as covariates of interest. Head motion entered as covariate of no interest. Bootstrapping at 1000 iterations. Four outliers removed with residuals more than three standard deviations from the mean.

https://doi.org/10.7554/eLife.40766.020

Outliers removed	Multivariate tests
	Effect		Wilk’s Λ	Df	F	P	η_p²
	Fearful		.995	(4, 609)	0.715	.582	.005
	Sad		.994	(4, 609)	0.915	.455	.006
	Angry		.998	(4, 609)	0.294	.882	.002
	Head Motion		.999	(4, 609)	0.223	.926	.001
	Parameter estimates
	Dependent variable	Parameter	B	T	P	η_p²	95%
	Left SC-PUL	Fearful	−0.350	−1.583	.114	.004	[−0.784, 0.084]
		Sad	−0.312	−1.520	.129	.004	[−0.715, 0.091]
		Angry	−0.157	−0.687	.493	.001	[−0.607, 0.292]
		Head Motion	−0.359	−0.677	.498	.001	[−1.401, 0.682]
	Right SC-PUL	Fearful	−0.208	−0.933	.351	.001	[−0.645, 0.23]
		Sad	−0.042	−0.202	.840	6.700 × 10⁻⁵	[−0.448, 0.364]
		Angry	−0.046	−0.197	.844	6.400 × 10⁻⁵	[−0.499, 0.408]
		Head Motion	−0.268	−0.501	.616	4.110 × 10⁻⁴	[−1.317, 0.781]
	Left PUL-AMG	Fearful	−0.006	−0.051	.959	4.000 × 10⁻⁶	[−0.225, 0.213]
		Sad	−0.010	−0.099	.921	1.600 × 10⁻⁵	[−0.213, 0.193]
		Angry	−0.024	−0.205	.837	6.900 × 10⁻⁵	[−0.251, 0.203]
		Head Motion	0.103	0.386	.700	2.430 × 10⁻⁴	[−0.422, 0.628]
	Left PUL-AMG	Fearful	−0.071	−0.611	.541	.001	[−0.297, 0.156]
		Sad	0.101	0.941	.347	.001	[−0.11, 0.311]
		Angry	−0.101	−0.846	.398	.001	[−0.336, 0.134]
		Head Motion	−0.103	−0.373	.709	2.280 × 10⁻⁴	[−0.647, 0.44]
No Outliers Removed	Multivariate Tests
	Effect		Wilk’s Λ	df	F	p	η_p²
	Fearful		.998	(4,613)	0.349	.845	.002
	Sad		.994	(4,613)	0.957	.430	.006
	Angry		.996	(4,613)	0.569	.685	.004
	Head Motion		.998	(4,613)	0.287	.886	.002
	Parameter Estimates
	Dependent Variable	Parameter	B	t	p	η_p²	95% CI
	Left SC-PUL	Fearful	−0.794	−0.809	.419	.001	[−2.723, 1.134]
		Sad	0.053	0.058	.954	5.000 × 10⁻⁷	[−1.736, 1.842]
		Angry	0.057	0.056	.955	5.000 × 10⁻⁷	[−1.938, 2.052]
		Head Motion	−0.147	−0.062	.950	6.000 × 10⁻⁷	[−4.763, 4.47]
	Right SC-PUL	Fearful	−0.612	−0.605	.546	.001	[−2.597, 1.374]
		Sad	0.482	0.514	.607	4.290 × 10⁻⁴	[−1.36, 2.325]
		Angry	0.314	0.3	.764	1.460 × 10⁻⁴	[−1.74, 2.369]
		Head Motion	−0.703	−0.29	.772	1.370 × 10⁻⁴	[−5.458, 4.052]
	Left PUL-AMG	Fearful	−0.072	−0.433	.665	3.050 × 10⁻⁴	[−0.4, 0.255]
		Sad	0.018	0.117	.907	2.200 × 10⁻⁵	[−0.286, 0.322]
		Angry	−0.109	−0.631	.529	.001	[−0.448, 0.23]
		Head Motion	0.191	0.479	.632	3.720 × 10⁻⁴	[−0.593, 0.976]
	Left PUL-AMG	Fearful	−0.126	−0.871	.384	.001	[−0.41, 0.158]
		Sad	0.12	0.894	.372	.001	[−0.144, 0.384]
		Angry	−0.135	−0.905	.366	.001	[−0.429, 0.158]
		Head Motion	−0.012	−0.033	.973	2.000 × 10⁻⁶	[−0.692, 0.669]

Appendix 1—table 11

Local tractography multivariate GLM.

Average apparent fibre density for left and right SC-PUL and PUL-AMG entered as dependent variables. Recognition scores (out of eight) for fearful, sad, and angry expressions entered as covariates of interest. Head motion entered as covariate of no interest. Bootstrapping at 1000 iterations. Fifteen outliers removed with residuals more than three standard deviations from the mean.

https://doi.org/10.7554/eLife.40766.021

Outliers removed	Multivariate tests
	Effect		Wilk’s Λ	Df	F	P	η_p²
	* Fearful		.984	(4,598)	2.501	.042 *	.016
	Sad		.991	(4,598)	1.363	.245	.009
	Angry		.993	(4,598)	1.111	.350	.007
	* Head Motion		.977	(4,598)	3.451	.008 *	.023
	Parameter estimates
	Dependent variable	Parameter	B	T	P	η_p²	95%
	Left SC-PUL	Fearful	0	−0.003	0.998	1.093 × 10⁻⁸	[−0.163, 0.162]
		Sad	0.084	1.099	0.272	0.002	[−0.066, 0.235]
		Angry	−0.078	−0.901	0.368	0.001	[−0.249, 0.093]
		Head Motion	−0.135	−0.681	0.496	0.001	[−0.526, 0.255]
	Right SC-PUL	Fearful	−0.046	−0.645	0.519	0.001	[−0.185, 0.094]
		Sad	−0.008	−0.123	0.902	2.500 × 10⁻⁵	[−0.137, 0.121]
		Angry	0.043	0.574	0.566	0.001	[−0.104, 0.189]
		Head Motion	−0.146	−0.859	0.39	0.001	[−0.48, 0.188]
	Left PUL-AMG	*fearful	0.14	2.887	0.004 *	0.014	[0.045, 0.235]
		Sad	−0.084	−1.855	0.064	0.006	[−0.172, 0.005]
		Angry	−0.052	−1.018	0.309	0.002	[−0.152, 0.048]
		Head Motion	−0.21	−1.802	0.072	0.005	[−0.438, 0.019]
	Right PUL AMG	*fearful	0.143	2.523	0.012 *	0.010	[0.032, 0.255]
		Sad	−0.042	−0.787	0.431	0.001	[−0.145, 0.062]
		Angry	0.041	0.686	0.493	0.001	[−0.076, 0.158]
		* Head Motion	−0.484	−3.55	4.150 × 10⁻⁴ *	0.021	[−0.752,–0.216]
No Outliers Removed	Multivariate Tests
	Effect		Wilk’s Λ	df	F	p	η_p²
	* Fearful		.983	(4,613)	2.669	.031 *	.017
	Sad		.993	(4,613)	1.080	.366	.007
	Angry		.997	(4,613)	0.499	.736	.003
	* Head Motion		.982	(4,613)	2.792	.026 *	.018
	Parameter Estimates
	Dependent Variable	Parameter	B	t	p	η_p²	95% CI
	Left SC-PUL	Fearful	−0.006	−0.051	.960	4.000 × 10⁻⁶	[−0.231, 0.22]
		Sad	0.121	1.138	.255	.002	[−0.088, 0.33]
		Angry	−0.080	−0.674	.500	.001	[−0.313, 0.153]
		Head Motion	−0.063	−0.229	.819	8.500 × 10⁻⁵	[−0.602, 0.477]
	Right SC-PUL	Fearful	−0.097	−1.037	.300	.002	[−0.28, 0.086]
		Sad	0.017	0.197	.844	6.300 × 10⁻⁵	[−0.153, 0.187]
		Angry	0.023	0.233	.816	8.800 × 10⁻⁵	[−0.167, 0.212]
		Head Motion	−0.165	−0.738	.461	.001	[−0.604, 0.274]
	Left PUL-AMG	* Fearful	0.137	2.650	.008 *	.011	[0.036, 0.239]
		Sad	−0.075	−1.563	.119	.004	[−0.17, 0.019]
		Angry	−0.049	−0.904	.366	.001	[−0.154, 0.057]
		Head Motion	−0.193	−1.554	.121	.004	[−0.437, 0.051]
	Left PUL-AMG	* Fearful	0.137	2.334	.020 *	.009	[0.022, 0.253]
		Sad	−0.053	−0.976	.330	.002	[−0.16, 0.054]
		Angry	0.001	0.019	.985	6.126 × 10⁻⁷	[−0.118, 0.121]
		* Head Motion	−0.465	−3.303	.001 *	.017	[−0.741,–0.189]

Appendix 1—table 12

Partial correlations between fibre density of each pathway and the corresponding DCM parameter estimate.

Scatterplots show the fibre density (x-axis) residuals (after regressing against head motion) and the DCM parameter estimate (y-axis) residuals (after regressing against head motion). *p<0.05, Bonferroni-corrected

https://doi.org/10.7554/eLife.40766.022

Appendix 1—table 13

Significance of DCM parameter estimates.

https://doi.org/10.7554/eLife.40766.023

Outliers removed	Parameter	M	SD	Df	T	P	95%
	* Left SC- PUL	1.063	0.115	229	8.285	<0.001	[1.048, 1.078]
	* Right SC-PUL	1.059	0.124	233	7.295	<0.001	[1.043, 1.075]
	* Left PUL-AMG	1.834	1.000	233	12.753	<0.001	[1.705, 1.962]
	* Right PUL-AMG	1.731	0.765	230	14.514	<0.001	[1.632, 1.83]
	* Left IOG-FG	1.685	0.715	228	14.485	<0.001	[1.592, 1.778]
	* Right IOG-FG	2.415	2.140	233	10.111	<0.001	[2.139, 2.69]
	* Left FG-AMG	1.213	0.258	233	12.614	<0.001	[1.18, 1.246]
	* Right FG-AMG	1.190	0.217	230	13.308	<0.001	[1.162, 1.218]
	Left FG- IOG	0.988	0.211	230	−0.842	0.401	[0.961, 1.016]
	Right FG- IOG	1.037	0.270	231	2.095	0.037	[1.002, 1.072]
	* Left AMG-FG	1.813	1.330	231	9.306	<0.001	[1.641, 1.985]
	* Right AMG-FG	1.878	1.101	228	12.078	<0.001	[1.735, 2.022]
No Outliers Removed	Parameter	M	SD	df	t	p	95% CI
	* Left SC- PUL	1.066	0.157	236	6.481	<0.001	[1.046, 1.086]
	* Right SC-PUL	1.068	0.190	236	5.510	<0.001	[1.044, 1.093]
	* Left PUL-AMG	1.916	1.235	236	11.415	<0.001	[1.758, 2.074]
	* Right PUL-AMG	1.821	0.944	236	13.382	<0.001	[1.7, 1.942]
	* Left IOG-FG	1.864	1.214	236	10.957	<0.001	[1.709, 2.02]
	* Right IOG-FG	2.772	4.249	236	6.422	<0.001	[2.229, 3.316]
	* Left FG- AMG	1.251	0.446	236	8.661	<0.001	[1.194, 1.308]
	* Right FG-AMG	1.218	0.278	236	12.081	<0.001	[1.182, 1.253]
	Left FG- IOG	1.028	0.329	236	1.311	0.191	[0.986, 1.07]
	Right FG- IOG	1.069	0.348	236	3.071	0.002	[1.025, 1.114]
	* Left AMG-FG	2.057	2.166	236	7.517	<0.001	[1.78, 2.335]
	* Right AMG-FG	2.092	1.596	236	10.536	<0.001	[1.888, 2.296]

*p <.001

Additional files

Transparent reporting form: https://doi.org/10.7554/eLife.40766.009
Download elife-40766-transrepform-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)

Article PDF

Open citations (links to open the citations from this article in various online reference manager services)

Mendeley

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

Jessica McFadyen
Jason B Mattingley
Marta I Garrido

(2019)

An afferent white matter pathway from the pulvinar to the amygdala facilitates fear recognition

eLife 8:e40766.

https://doi.org/10.7554/eLife.40766

Share this article

Cite this article

Global and local probabilistic tractography reconstructions of the subcortical route to the amygdala.

Relationship between behavioural performance and tractography.

Whole-brain fMRI activation for faces and shapes.

Dynamic causal modelling model space and estimated probabilities.

The strength of individual connections from the winning dynamic causal model.

The relationship between structural and effective connectivity for the subcortical route.

Participants with above or below threshold fMRI signal in all VOIs for inclusion into DCM stage.

Global tractography fibre counts.

Local tractography fibre counts.

Global tractography fibre counts (terminations in pulvinar subregions) for SC-PUL.

Global tractography fibre counts (terminations in pulvinar subregions) for PUL-AMG.

Global tractography fibre counts (terminations in amygdala subregions) for PUL-AMG.

Local tractography fibre counts (terminations in pulvinar subregions) for SC-PUL.

Local tractography fibre counts (terminations in pulvinar subregions) for AMG-PUL.

Local tractography fibre counts (terminations in amygdala subregions) for PUL-AMG.

Paired t-tests between local streamline count and null distribution.

Global tractography multivariate GLM.

Local tractography multivariate GLM.

Partial correlations between fibre density of each pathway and the corresponding DCM parameter estimate.

Significance of DCM parameter estimates.

Transparent reporting form

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)