Connectivity between the superior colliculus and the amygdala in humans and macaque monkeys: virtual dissection with probabilistic DTI tractography

Abstract

It has been suggested that some cortically blind patients can process the emotional valence of visual stimuli via a fast, subcortical pathway from the superior colliculus (SC) that reaches the amygdala via the pulvinar. We provide in vivo evidence for connectivity between the SC and the amygdala via the pulvinar in both humans and rhesus macaques. Probabilistic diffusion tensor imaging tractography revealed a streamlined path that passes dorsolaterally through the pulvinar before arcing rostrally to traverse above the temporal horn of the lateral ventricle and connect to the lateral amygdala. To obviate artifactual connectivity with crossing fibers of the stria terminalis, the stria was also dissected. The putative streamline between the SC and amygdala traverses above the temporal horn dorsal to the stria terminalis and is positioned medial to it in humans and lateral to it in monkeys. The topography of the streamline was examined in relation to lesion anatomy in five patients who had previously participated in behavioral experiments studying the processing of emotionally valenced visual stimuli. The pulvinar lesion interrupted the streamline in two patients who had exhibited contralesional processing deficits and spared the streamline in three patients who had no deficit. Although not definitive, this evidence supports the existence of a subcortical pathway linking the SC with the amygdala in primates. It also provides a necessary bridge between behavioral data obtained in future studies of neurological patients, and any forthcoming evidence from more invasive techniques, such as anatomical tracing studies and electrophysiological investigations only possible in nonhuman species.

Keywords: DTI tractography; amygdala; pulvinar; superior colliculus; threat.

Fig. 1.

Masks used for the virtual dissection of possible connectivity between the superior colliculus (SC) and the amygdala, and for the dissection of the stria terminalis in the first group of human participants. Left, axial sections; middle, coronal sections; right, sagittal sections. A: SC masks (shown in yellow). B: pulvinar masks (shown in blue). C: amygdala masks for first and second groups (shown in green). D: bed nucleus of the stria terminalis masks (shown in magenta). E: stria terminalis waypoint mask subjacent to the frontal horn of the lateral ventricle between the head of the caudate nucleus and the thalamus.

Fig. 2.

Probabilistic tractography between the SC and the amygdala via the pulvinar in one representative participant from the first group. A: three-dimensional (3D) reconstructions of the tract (shown in red) linking the SC (shown in yellow) and the amygdala (shown in green) via the pulvinar (shown in blue). Tracts have been thresholded by 10% (see methods for details). The individual structures and tracts in the 3D reconstructions have been dilated by 2–4 mm for visualization purposes only. Coronal (B) and sagittal (C) sections show the location of the tract relative to the amygdala (green), pulvinar (blue), and SC (yellow). The probabilistic data are presented unthresholded (and actual size) as a percentage of the total number of traces linking the SC and amygdala (that pass through the pulvinar).

Fig. 3.

3D reconstructions of the SC-pulvinar-amygdala tract (shown in red) in 3 additional representative participants from the first (A) and second (B) groups. Tracts have been thresholded by 10% (see methods for details). The individual structures and tracts in the 3D reconstructions have been dilated by 2–6 mm for visualization purposes only. C: composite tract of all 12 participants from the second group (higher resolution). The composite tract has been thresholded such that only voxels that were identified in at least 75% of participants are shown.

Fig. 4.

Masks used for the dissection of possible connectivity between the SC and the amygdala in the monkey brain. Left, axial sections; middle, coronal sections; right, sagittal sections. A: SC masks (shown in yellow). B: pulvinar masks (shown in blue). C: amygdala masks (shown in green).

Fig. 5.

Probabilistic tractography between the SC and the amygdala via the pulvinar in the monkey brain. Top: 3D reconstructions of the tract (shown in red) linking the SC (shown in yellow) and the amygdala (shown in green) via the pulvinar (shown in blue). Coronal (middle) and sagittal (bottom) sections show the location of the tract relative to the amygdala (green), pulvinar (blue), and SC (yellow). The probabilistic data are presented unthresholded as a percentage of the total number of traces linking the SC and amygdala (that pass through the pulvinar).

Fig. 6.

3D reconstructions of the SC-pulvinar-amygdala tract (shown in red) in the remaining 6 monkey subjects (in whom the tract was identified). Tracts have been thresholded by 10% (see methods for details). Unthresholded data are shown as transparency.

Fig. 7.

Overlap of binarized streamlines between the SC and the amygdala for all 12 participants (the second group) aligned to the Montreal Neurological Institute T1-weighted standard atlas in the coronal plane (A) and 8 monkeys (B). The color scale shows the proportion of participants whose streamline runs through each voxel.

Fig. 8.

Location of the putative tract linking the SC and the amygdala relative to the stria terminalis in the human brain. A: 3D reconstructions of the SC-pulvinar-amygdala tract (shown in red) and the stria terminalis (shown in purple). Tracts have been thresholded by 10% (see methods for details). Unthresholded data are shown as transparency. The individual structures and tracts have been expanded in size for visualization purposes only (SC: 4 mm; amygdala and pulvinar: 2 mm; tract: 2 mm). B: coronal sections showing the location of the tract relative to the stria terminalis. The probabilistic data for both the SC-pulvinar-amygdala tract and the stria terminalis are presented unthresholded as a percentage of the total number of traces linking the starting and termination masks.

Fig. 9.

Location of the putative tract linking the SC and the amygdala relative to the stria terminalis in the monkey brain. A: 3D reconstruction of the SC-pulvinar-amygdala tract (shown in red) and the stria terminalis (shown in purple). Tracts have been thresholded by 10% (see methods for details). B: coronal sections showing the location of the tract relative to the stria terminalis. The probabilistic data for both the SC-pulvinar-amygdala tract and the stria terminalis are presented unthresholded as a percentage of the total number of traces linking the starting and termination masks.

Fig. 10.

Probabilistic tractography between the SC and the amygdala via the pulvinar in patient DG. A: 3D reconstructions of the tract (shown in red) in relation to the site of the lesion (shown in green). Coronal (B) and sagittal (C) sections show the location of the tract relative to the lesion (green). Data are presented on a subtraction of the T2-weighted image [SO image generated by DTIfit from the FSL FDT Diffusion Toolbox] from the mean diffusivity image (MD image generated by DTIfit from the FSL FDT Diffusion Toolbox). This image highlights the lesion as a hypointense, cystic cavity in the pulvinar (masked with translucent green), surrounded by tissue with hemosiderin iron staining (high signal) from the previous hemorrhage. The probabilistic data are presented unthresholded (and actual size) as a percentage of the total number of traces linking the SC and amygdala (that pass through the pulvinar).

Fig. 11.

Coronal, axial, and sagittal slices for two patients who exhibited a contralesional behavioral deficit in the emotional recognition tasks [Ward et al. 2005 (patient SM; A) and 2007 (patient CJ; B)].

Fig. 12.

Coronal, axial, and sagittal slices for two patients [TN (A) and GJ (B)] who, like patient DG, exhibited preserved abilities following pulvinar damage (Ward et al. 2007).