Neuropsychological evidence of multi-domain network hubs in the human thalamus

  1. Kai Hwang  Is a corresponding author
  2. James M Shine
  3. Joel Bruss
  4. Daniel Tranel
  5. Aaron Boes
  1. Department of Psychological and Brain Sciences, The University of Iowa & The University of Iowa College of Medicine, United States
  2. Cognitive Control Collaborative, The University of Iowa & The University of Iowa College of Medicine, United States
  3. Iowa Neuroscience Institute, The University of Iowa & The University of Iowa College of Medicine, United States
  4. Department of Psychiatry, The University of Iowa & The University of Iowa College of Medicine, United States
  5. Brain and Mind Center, The University of Sydney, Australia
  6. Department of Neurology, The University of Iowa & The University of Iowa College of Medicine, United States
  7. Department of Pediatrics, The University of Iowa & The University of Iowa College of Medicine, United States
8 figures, 1 table and 2 additional files

Figures

Thalamic patients performed significantly worse on multiple neuropsychological tests compared to comparison patients.

(A) Overlap of lesions in patients with thalamic lesions and comparison patients with cortical lesions. (B) Neuropsychological test scores from patients with thalamic lesions and comparison patients. All test scores transformed to z-score using published population norm. We inverted z-scores from TMT Part B to facilitate comparison, for all tests negative z-scores indicate more severe impairment. For each plot, the solid dot depicts the mean, and the bar depicts the 95% bootstrapped confidence interval. * Indicates corrected p < 0.05. TMT: Trail Making Test; COWA: Controlled Oral Word Association Test; RAVLT: Rey Auditory-Verbal Learning Test.

Figure 2 with 1 supplement
Lesions associated with impaired performance on tasks across multiple cognitive domains.

(A) Left panel: overlap of lesion masks from subjects with impaired test performance on each task. Impaired task performance defined as z < −1.645 (95 percentile in z distribution). Right panel: overlap of lesion sites associated with impairment on each individual task (summing each individual task’s lesion map from the left panel). (B) Table showing each thalamus patient’s task performance on 10 different neuropsychology tests. For all tasks, negative z-scores indicate more pronounced impairment. Both Trail Making Test (TMT) Part A and Part B scores were inverted to match the directionality of other tests. (C) Left panel: classifying thalamic patients into groups that exhibit impairment in one versus multiple tasks across cognitive domains. Right panel: Lesion sites in patients with or without impairment across multiple tasks.

Figure 2—figure supplement 1
Lesion sites of all thalamic patients.
Figure 3 with 1 supplement
Comparing the degree of average behavioral impairment (x axis) and multi-domain behavioral impairment (y axis) between thalamus patients and comparison patients.

More negative average impairment score represents more severe behavioral impairment. Higher multi-domain impairment score indicate more cognitive domains were affected. Individual dots represent individual patients. The solid line indicates a fitted regression line for the comparison patient group. The shaded area represents the 95% confidence interval.

Figure 3—figure supplement 1
Lesion sites 145 comparison patients.
Neurosynth meta-analyses.

Top four rows: brain regions associated with putative cognitive processes assessed by the Trail Making Test Part B (TMT Part B), Boston Naming Test (BNT), Controlled Oral Word Association Test (COWA), Rey Auditory-Verbal Learning Test (RAVLT) delayed recall, and RAVLT delayed recognition tests. Color bar represents the strength of association in z-score. Bottom row: overlap between the top four maps. Maps for ‘naming’ and ‘fluency’ overlapped in the left frontal cortex, maps for ‘recall’ and ‘recognition’ overlapped in temporal cortices. All four maps overlapped with six voxels in the left middle frontal gyrus.

Lesions to thalamic regions with strong hub properties are associated with behavioral impairment across cognitive domains.

(A) Thalamic lesion sites associated with multi-domain impairment are located in the anterior-medio-dorsal thalamus. (B) Right panel: left medial and anterior thalamus is associated with prominent hub property (measured by participation coefficient [PC]). (C) Kernel density plot of voxel-wise PC values from thalamic lesion sites associated with multi-domain and single-domain impairment. Voxel-wise PC values were significantly higher for multi-domain lesion sites. The y-axis was scaled so area under the curve is summated to 1. (D) In a group of 235 subjects, PC values were significantly higher in multi-domain versus single-domain lesion sites in the thalamus. Each data dot indicates PC value from one normative subject.

Comparison lesions with strong hub properties are associated with behavioral impairment across cognitive domains.

(A) Comparison lesion sites associated with multi-domain impairment are located in the lateral frontal and posterior parietal regions. (B) Frontal and parietal regions exhibit strong connector hub properties (measured by participation coefficient [PC]). (C) Kernel density plot of voxel-wise PC values from comparison lesion sites associated with multi-domain and single-domain impairment. Voxel-wise PC values were significantly higher for multi-domain lesion sites. The y-axis was scaled so area under the curve is summated to one. (D) Comparing the distribution of PC values in comparison and thalamic lesion sites. (E) Comparing multiple-domain impairment score between thalamic and expanded comparison patients with multi-domain impairment.

The multi-domain lesion site does not show strong functional specificity.

(A) Voxel-wise kernel density plot of thalamocortical functional connectivity weight for each cortical network. Voxels in the multi-domain lesion site exhibit stronger functional connectivity with the cingulo-opercular (CO), limbic (Lm), frontoparietal (FP), and default mode (DMN) networks. Weaker connectivity with visual (V), somatomotor (SM), and dorsal attention (DA) networks. (B) Voxel-wise kernel density plot of thalamocortical functional connectivity weight with cortical regions identified via Neurosynth meta-analyses. For both (A) and (B), higher weight ratio indicates functional specificity, suggesting that voxel is selectively interacting with a specific cortical system. (C) The multi-domain lesion site overlaps with higher-order thalamic nuclei, including the anterior (AN), ventromedial (VM), mediodorsal (MD), intra-laminar (IL), ventro-anterior (VA), and ventrolateral (VL) nuclei.

The multi-domain lesion site contains higher concentration of matrix cells.

(A) The anterior-medial thalamus had a relative higher expression of CALB1 (lower panel), which overlapped with the multi-domain lesion sites (upper panel). Color intensity in the lower panel denotes the relative expression of CALB1 and PVALB in different thalamic voxels (Jones, 2009; Jones, 2001). (B) Voxel-wise kernel density plot of normalized CALB1 expression level for the multi-domain versus the single-domain lesion sites. (C) Voxel-wise kernel density plot of normalized PVALB expression level for the multi-domain versus the single-domain lesion sites.

Tables

Table 1
Group comparisons on neuropsychological test results.
Thalamic patientsComparison patients
Mean z-scoreSDMean z-scoreSDRandomized permutation p
Trail Making Test Part A0.381.080.061.480.10
Trail Making Test Part B–2.152.940.31.55<0.001
Boston Naming Task–0.290.730.30.640.0033
Controlled Oral Word Association Test–0.41.30.381.160.055
Rey Auditory-Verbal Learning
Immediate Learning–0.680.71–0.60.850.76
Total Learning–0.791.37–0.220.920.11
Delayed Recall–0.731.290.450.940.0021
Delayed Recognition–1.362.70.120.860.0049
Rey Complex Figure Test
Delayed Recall–0.160.860.0940.940.41
Copy Test–0.511.170.0640.560.098

Additional files

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)

  1. Kai Hwang
  2. James M Shine
  3. Joel Bruss
  4. Daniel Tranel
  5. Aaron Boes
(2021)
Neuropsychological evidence of multi-domain network hubs in the human thalamus
eLife 10:e69480.
https://doi.org/10.7554/eLife.69480