Abstract

The human thalamus has been suggested to be involved in executive function, based on animal studies and correlational evidence from functional neuroimaging in humans. Human lesion studies, examining behavioral deficits associated with focal brain injuries, can directly test the necessity of the human thalamus for executive function. The goal of our study was to determine the specific lesion location within the thalamus as well as the potential disruption of specific thalamocortical functional networks, related to executive dysfunction. We assessed executive function in 15 patients with focal thalamic lesions and 34 comparison patients with lesions that spared the thalamus. We found that patients with mediodorsal thalamic lesions exhibited more severe impairment in executive function when compared to both patients with thalamic lesions that spared the mediodorsal nucleus and to comparison patients with lesions outside the thalamus. Furthermore, we employed a lesion network mapping approach to map cortical regions that show strong functional connectivity with the lesioned thalamic subregions in the normative functional connectome. We found that thalamic lesion sites associated with more severe deficits in executive function showed stronger functional connectivity with ACC, dorsomedial PFC, and frontoparietal network, compared to thalamic lesions not associated with executive dysfunction. These are brain regions and functional networks whose dysfunction could contribute to impaired executive functioning. In aggregate, our findings provide new evidence that delineates a thalamocortical network for executive function.

INTRODUCTION

Executive function, also commonly termed cognitive control, describes cognitive processes that support goal-directed behavior. It is closely related to cognitive flexibility, working memory, and inhibition (Miyake, Friedman, Emerson, & Witzki, 2000). Collectively, executive- and control-related functions allow individuals to carry out adaptive and purposeful behaviors that are characteristically human (Lezak, Howieson, Bigler, & Tranel, 2012).

There is a long tradition of mapping the neural substrates of executive function by examining the association between focal brain lesions and neuropsychological outcomes on putative “executive function” tests. A systematic approach of this type of investigation is referred to as lesion symptom mapping. In particular, past studies have highlighted the effects of frontal lobe injuries, suggesting that distinct prefrontal regions support different cognitive processes related to executive function (Tsuchida & Fellows, 2013; Gläscher et al., 2012; Badre, Hoffman, Cooney, & D'Esposito, 2009). Functional neuroimaging studies provide additional evidence characterizing the functional–anatomical relationships of executive function. In contrast to the lesion symptom mapping approach, recent neuroimaging studies have focused on characterizing patterns of functional interactions (also referred to as functional connectivity) between brain regions. Studies suggest that sets of frontal and parietal regions, including the lateral frontal cortex, the dorsomedial frontal cortex, and the posterior parietal cortex, form functional networks whose interactions are involved in cognitive processes related to executive function (Shine et al., 2016; Cole & Schneider, 2007; Dosenbach et al., 2007). The identified cortical networks, including the frontoparietal network (Vincent, Kahn, Snyder, Raichle, & Buckner, 2008), the dorsal attention network (Fox, Corbetta, Snyder, Vincent, & Raichle, 2006), and the cingulo-opercular network (Seeley et al., 2007), are named in accordance with their putative functions and anatomical positions.

Beneath the cerebral cortex, findings from anatomical and functional connectivity studies suggest that the thalamus plays an important role in executive function. In particular, the mediodorsal nucleus has reciprocal and nonreciprocal anatomical connectivity with PFC (Giguere & Goldman-Rakic, 1988; Selemon & Goldman-Rakic, 1988; Goldman-Rakic & Porrino, 1985), and several thalamic nuclei exhibit strong functional connectivity with cortical networks implicated in executive function (Hwang, Bertolero, Liu, & D'Esposito, 2017). These observations, although compelling, are nevertheless correlational and circumstantial evidence. A recent study using deep brain stimulation to manipulate thalamic activity allowed a more causal inference of the role of the mediodorsal nucleus in working memory performance (Peräkylä et al., 2017). Quantifying behavioral change after thalamic injuries also provides more direct evidence for establishing the contribution of the thalamus to executive function. However, it is challenging to map the thalamic loci for executive function using the traditional lesion symptom mapping approach. One challenge is that the thalamus is not a homogenous region; it consists of multiple nuclei, each with a partially distinct anatomical and functional connectivity profiles. Thalamic lesions are rarely restricted to one specific thalamic nucleus, and deficits caused by thalamic lesions often exhibit a wide range of individual variability (Schmahmann, 2003). For example, in one study examining executive function deficits in 19 patients with medial thalamic lesions, eight patients exhibited impairment in executive function, whereas 11 patients did not have such an impairment (Liebermann, Ploner, Kraft, Kopp, & Ostendorf, 2013).

Few studies have attempted to integrate these lines of research in linking executive dysfunction to specific thalamic regions using the lesion approach, while also considering the potential impact of thalamic lesions on a broader functional network that receive anatomical projections from thalamic nuclei. A lesion in a thalamic subregion may disrupt executive function both through direct effects within the thalamus and by disruption of a broader functional network the thalamic region is embedded in. When considering brain functions that arise from interactive network processes, behavioral deficits may result from lesion-induced dysfunction in remote but connected regions—a phenomenon known as “diaschisis” (Carrera & Tononi, 2014; Von Monakow, 1911). Lesion network mapping is a method that attempts to identify the distributed brain regions that are likely affected by a focal brain lesion (Boes et al., 2015). This method uses the site of the brain lesion as a seed ROI for a functional connectivity analysis. This allows identification of brain regions that are functionally connected with the location of the lesion, as inferred using normative functional connectivity data from healthy individuals. Focal lesions that induce similar behavioral deficits through disruption of a broader network should show overlapping spatial patterns in their lesion network maps. Thus, an approach that combines lesion symptom mapping with lesion network mapping is well suited to identify which thalamic regions are critical for executive function, while also inferring the broader brain networks affected by the focal thalamic lesion.

In this study, we tested two main hypotheses: (1) that lesions of the mediodorsal thalamus would be associated with impaired executive function and (2) that thalamic lesions associated with impaired executive function would exhibit stronger functional connectivity with the frontoparietal functional network, when compared to thalamic lesions associated with intact executive function. To test the first hypothesis, we evaluated patients with focal thalamic lesions and comparison patients with lesions that spared the thalamus and PFC. We predicted that patients with lesions to the mediodorsal thalamus would exhibit significantly impaired executive function relative to other participants, as assessed with two classic standardized neuropsychological tests of executive function, the Trail Making Test (TMT) and the Wisconsin Card Sorting Test (WCST). To test the second hypothesis, we compared the lesion network maps of lesions associated with impaired versus preserved executive function, predicting greater connectivity with the frontoparietal network in the impaired group.

METHODS

Participants

We studied 49 neurological patients (24 female participants; mean age at testing = 54.88 years, SD = 11.06 years, range = 39–77 years). All patients were drawn from the Iowa Neurological Patient Registry. Before enrollment, all patients were screened to exclude individuals with learning disabilities, psychiatric disorders, substance abuse, premorbid personality disorders, or other neurological conditions not related to their focal lesion. All patients had focal lesions caused by ischemic or hemorrhagic stroke, and all neuropsychological data were collected in the chronic phase of recovery (at least 3 months after lesion onset; chronicity data reported in Tables 1 and 2). Imaging and neuropsychological data were acquired contemporaneously, separated by at most a few months (M = 4.6 months, SD = 4.5 months). Only patients who completed at least one of the executive function neuropsychology tests of interests (see below) were included.

Table 1. 
Demographic and Test Data for Thalamic Patients
Thalamic PatientDemographicsLesion Size (ml)
AgeSexHandednessChronicity (Months)LateralityEducation (Years)Total Lesion SizeANMDPuVLVP
P1 58 18 12 371 72 
P2 62 12 320 128 192 
P3 68 36 12 1224 120 104 120 
P4 69 Mix 15 16 1457 128 456 
P5 52 12 548 304 128 
P6 57 132 12 1001 224 
P7 60 18 11 520 96 152 72 
P8 52 14 2166 64 752 264 
P9 52 Mix 20 1105 232 144 
P10 32 12 48 16 
P11 68 12 2497 240 608 144 152 
P12 29 15 1830 296 336 
P13 52 14 1387 72 344 336 40 
P14 70 12 2536 192 1232 152 552 
P15 61 12 1940 64 232 18 1344 
Mean 56.13 4F/11M 13R/2Mix 18.53 4R/2B/9L 13.20 1263.33 27.73 194.67 123.87 243.73 84.80 
SD 12.26     32.55   2.34 806.72 42.64 190.99 344.10 329.01 162.85 
  
 TMT Test ScoresWCST Test ScoresIQ Test Scores  
TMT A (z Score)TMT B (z Score)TMT B Minus A (z Score)Perseverative ErrorsCorrect ResponsesTotal ErrorsCategoriesVIQPIQFSIQ  
P1 0.36 0.64 0.28 n/a n/a n/a n/a 87 86 86     
P2 −0.25 −0.39 −0.14 n/a n/a n/a n/a 90 101 94     
P3 −0.1 2.73 2.83 15 95 33 103 122 112     
P4 2.42 11.48 9.06 65 14 104 109 107     
P5 −0.38 3.21 3.59 10 78 20 n/a n/a n/a     
P6 0.08 3.83 3.75 86 12 110 106 108     
P7 1.73 3.25 1.52 49 73 55 n/a n/a n/a     
P8 −0.38 2.1 2.48 73 12 n/a n/a n/a     
P9 0.73 2.1 1.37 63 10 103 113 108     
P10 n/a n/a n/a 16 80 25 88 87 88     
P11 −0.19 0.023 0.213 45 58 70 105 84 96     
P12 0.87 −0.54 −1.41 63 10 95 102 99     
P13 −0.08 0.85 0.93 n/a n/a n/a n/a 119 104 111     
P14 0.1 3.02 2.92 n/a n/a n/a n/a n/a n/a n/a     
P15 1.18 6.31 5.13 n/a n/a n/a n/a n/a n/a n/a     
Mean 0.44 2.76 2.32 16.50 73.40 26.10 5.40 100.40 101.40 100.90     
SD 0.85 3.13 2.63 16.56 11.62 20.85 0.97 10.31 12.44 9.59     
Thalamic PatientDemographicsLesion Size (ml)
AgeSexHandednessChronicity (Months)LateralityEducation (Years)Total Lesion SizeANMDPuVLVP
P1 58 18 12 371 72 
P2 62 12 320 128 192 
P3 68 36 12 1224 120 104 120 
P4 69 Mix 15 16 1457 128 456 
P5 52 12 548 304 128 
P6 57 132 12 1001 224 
P7 60 18 11 520 96 152 72 
P8 52 14 2166 64 752 264 
P9 52 Mix 20 1105 232 144 
P10 32 12 48 16 
P11 68 12 2497 240 608 144 152 
P12 29 15 1830 296 336 
P13 52 14 1387 72 344 336 40 
P14 70 12 2536 192 1232 152 552 
P15 61 12 1940 64 232 18 1344 
Mean 56.13 4F/11M 13R/2Mix 18.53 4R/2B/9L 13.20 1263.33 27.73 194.67 123.87 243.73 84.80 
SD 12.26     32.55   2.34 806.72 42.64 190.99 344.10 329.01 162.85 
  
 TMT Test ScoresWCST Test ScoresIQ Test Scores  
TMT A (z Score)TMT B (z Score)TMT B Minus A (z Score)Perseverative ErrorsCorrect ResponsesTotal ErrorsCategoriesVIQPIQFSIQ  
P1 0.36 0.64 0.28 n/a n/a n/a n/a 87 86 86     
P2 −0.25 −0.39 −0.14 n/a n/a n/a n/a 90 101 94     
P3 −0.1 2.73 2.83 15 95 33 103 122 112     
P4 2.42 11.48 9.06 65 14 104 109 107     
P5 −0.38 3.21 3.59 10 78 20 n/a n/a n/a     
P6 0.08 3.83 3.75 86 12 110 106 108     
P7 1.73 3.25 1.52 49 73 55 n/a n/a n/a     
P8 −0.38 2.1 2.48 73 12 n/a n/a n/a     
P9 0.73 2.1 1.37 63 10 103 113 108     
P10 n/a n/a n/a 16 80 25 88 87 88     
P11 −0.19 0.023 0.213 45 58 70 105 84 96     
P12 0.87 −0.54 −1.41 63 10 95 102 99     
P13 −0.08 0.85 0.93 n/a n/a n/a n/a 119 104 111     
P14 0.1 3.02 2.92 n/a n/a n/a n/a n/a n/a n/a     
P15 1.18 6.31 5.13 n/a n/a n/a n/a n/a n/a n/a     
Mean 0.44 2.76 2.32 16.50 73.40 26.10 5.40 100.40 101.40 100.90     
SD 0.85 3.13 2.63 16.56 11.62 20.85 0.97 10.31 12.44 9.59     

B = bilateral; L = left ; R = right; AN = anterior nucleus; MD = mediodorsal nucleus; Pu = pulvinar; VL = ventrolateral nucleus; VP = ventroposterior nucelus; VIQ = verbal IQ; PIQ = performance IQ; FSIQ = full scale IQ.

Table 2. 
Demographic and Test Data for Comparison
GroupNAge (Years)SexHandednessEducation (Years)VIQPIQFSIQ
All comparison patients 34 54.3 (10.64) 20F/14M 30R/3L/1Mi 13.3 (2.70) 100.4 (10.31) 101.4 (11.44) 100.9 (11.21) 
Comparison patients with lesions in the temporal cortex 56.44 (9.63) 5F/4M 9R 13 (2.45) 101 (8.09) 110.5 (4.73) 106 (7.81) 
Comparison patients with lesions in the parietal cortex 50.22 (8.66) 7F/2M 7R/1L/1Mi 13.67 (2.06) 105.14 (6.59) 112.57 (8.42) 109.42 (6.70) 
Comparison patients with lesions in the occipital cortex 59 (7.05) 5F/3M 7R/1L 14.13 (2.59) 106 (8.72) 117 (12.53) 113.33 (6.5) 
Comparison patients with lesions in the BG 13 54 (12.34) 8F/5M 12R/1L 12.31 (2.72) 92.14 (7.47) 93.33 (12.4) 93.71 (9.83) 
Comparison patients with lesions in the cerebellum 56.33 (9.71) 1f/2M 1R/2L 15.67 (2.52) 110a 99 100 
  
 NChronicity (Months)Lesion LateralityTotal Lesion Size in mlWM Lesion SizeGM Lesion Size  
All comparison patients 34 16.3 (18.28) 14R/20L 1385 (676) 651.76 (571.22) 733.24 (375.06)     
Comparison patients with lesions in the temporal cortex 18.11 (22.87) 2R/7L 1075.44 (418.02) 409.44 (178.49) 655.22 (320.09)     
Comparison patients with lesions in the parietal cortex 26.67 (18.97) 5R/4L 1367.67 (473.33) 791 (710.83) 565.44 (453.93)     
Comparison patients with lesions in the occipital cortex 17.5 (22.42) 2R/6L 1225.13 (722.82) 558.63 (467.35) 642.25 (531.1)     
Comparison patients with lesions in the BG 13 15.25 (17.19) 5R/8L 1236.63 (656.03) 838.23 (656.24) 394.54 (212.97)     
Comparison patients with lesions in the cerebellum 1R/2L 291.67 (137.29) 150.67 (139) 140.33 (213.32)     
  
 NTMT B (z Score)TMT A (z Score)TMT B minus A (z Score)Perseverative ErrorsCorrect ResponsesTotal ErrorsCategories
All comparison patients 34 0.84 (1.75) 0.6 (1.41) 0.24 (1.4) 20 (12.34) 72.04 (9.45) 38.62 (21.47) 4.29 (1.83) 
Comparison patients with lesions in the temporal cortex 0.48 (1.95) 0.49 (2.09) −0.01 (1.4) 17.33 (13.23) 68.83 (7.11) 33 (25.19) 4.33 (2.34) 
Comparison patients with lesions in the parietal cortex 0.6 (1.99) 1 (2.28) −0.4 (0.93) 20.71 (15.54) 66.29 (4.03) 41.14 (26.13) 3.57 (2.51) 
Comparison patients with lesions in the occipital cortex 1.39 (2.1) 0.35 (1.36) 1.04 (1.74) 13.75 (9.81) 64.5 (2.65) 38.25 (30.48) 4 (2.45) 
Comparison patients with lesions in the BG 13 0.7 (1.66) 0.7 (1.04) 0 (1.26) 27 (11.55) 75.57 (11.27) 47.29 (16.41) 3.86 (1.21) 
Comparison patients with lesions in the cerebellum 2.62 (2.75) 0.09 (0.96) 2.53 (1.79) 10a 81 18 
GroupNAge (Years)SexHandednessEducation (Years)VIQPIQFSIQ
All comparison patients 34 54.3 (10.64) 20F/14M 30R/3L/1Mi 13.3 (2.70) 100.4 (10.31) 101.4 (11.44) 100.9 (11.21) 
Comparison patients with lesions in the temporal cortex 56.44 (9.63) 5F/4M 9R 13 (2.45) 101 (8.09) 110.5 (4.73) 106 (7.81) 
Comparison patients with lesions in the parietal cortex 50.22 (8.66) 7F/2M 7R/1L/1Mi 13.67 (2.06) 105.14 (6.59) 112.57 (8.42) 109.42 (6.70) 
Comparison patients with lesions in the occipital cortex 59 (7.05) 5F/3M 7R/1L 14.13 (2.59) 106 (8.72) 117 (12.53) 113.33 (6.5) 
Comparison patients with lesions in the BG 13 54 (12.34) 8F/5M 12R/1L 12.31 (2.72) 92.14 (7.47) 93.33 (12.4) 93.71 (9.83) 
Comparison patients with lesions in the cerebellum 56.33 (9.71) 1f/2M 1R/2L 15.67 (2.52) 110a 99 100 
  
 NChronicity (Months)Lesion LateralityTotal Lesion Size in mlWM Lesion SizeGM Lesion Size  
All comparison patients 34 16.3 (18.28) 14R/20L 1385 (676) 651.76 (571.22) 733.24 (375.06)     
Comparison patients with lesions in the temporal cortex 18.11 (22.87) 2R/7L 1075.44 (418.02) 409.44 (178.49) 655.22 (320.09)     
Comparison patients with lesions in the parietal cortex 26.67 (18.97) 5R/4L 1367.67 (473.33) 791 (710.83) 565.44 (453.93)     
Comparison patients with lesions in the occipital cortex 17.5 (22.42) 2R/6L 1225.13 (722.82) 558.63 (467.35) 642.25 (531.1)     
Comparison patients with lesions in the BG 13 15.25 (17.19) 5R/8L 1236.63 (656.03) 838.23 (656.24) 394.54 (212.97)     
Comparison patients with lesions in the cerebellum 1R/2L 291.67 (137.29) 150.67 (139) 140.33 (213.32)     
  
 NTMT B (z Score)TMT A (z Score)TMT B minus A (z Score)Perseverative ErrorsCorrect ResponsesTotal ErrorsCategories
All comparison patients 34 0.84 (1.75) 0.6 (1.41) 0.24 (1.4) 20 (12.34) 72.04 (9.45) 38.62 (21.47) 4.29 (1.83) 
Comparison patients with lesions in the temporal cortex 0.48 (1.95) 0.49 (2.09) −0.01 (1.4) 17.33 (13.23) 68.83 (7.11) 33 (25.19) 4.33 (2.34) 
Comparison patients with lesions in the parietal cortex 0.6 (1.99) 1 (2.28) −0.4 (0.93) 20.71 (15.54) 66.29 (4.03) 41.14 (26.13) 3.57 (2.51) 
Comparison patients with lesions in the occipital cortex 1.39 (2.1) 0.35 (1.36) 1.04 (1.74) 13.75 (9.81) 64.5 (2.65) 38.25 (30.48) 4 (2.45) 
Comparison patients with lesions in the BG 13 0.7 (1.66) 0.7 (1.04) 0 (1.26) 27 (11.55) 75.57 (11.27) 47.29 (16.41) 3.86 (1.21) 
Comparison patients with lesions in the cerebellum 2.62 (2.75) 0.09 (0.96) 2.53 (1.79) 10a 81 18 
a

Only one cerebellar patient completed the WCST and Wechsler Adult Intelligence Scale. GM = gray matter; WM = white matter.

Participants included 15 patients with lesions restricted to the thalamus and 34 patients with lesions that spared the thalamus, which served as a comparison group. The thalamic group included 11 men and four women (mean age = 56.13 years, SD = 12.26; mean years of education = 13.2, SD = 2.34; Figure 1A and Table 1). The nonthalamic comparison group was included in an attempt to control for nonspecific lesion effects on executive dysfunction (not specific to the thalamus). These individuals were selected from a broader cohort of 852 patients in the Iowa Neurological Patient Registry based on lesion size, as lesions within the thalamus were relatively small compared to typical vascular lesions outside the thalamus (in our patient registry). We tried to minimize any bias that could be introduced by having different lesion sizes between patient groups. Thus, comparison patients had to have a lesion size that was equal or smaller than the largest lesion size we observed in our thalamic patient group (2536 mm3). Individuals with lesions of PFC were also excluded, with PFC defined using the Harvard-Oxford cortical structural atlas (Desikan et al., 2006). This exclusion was based on the known anatomical connectivity with the mediodorsal nucleus and known PFC involvement in executive function tasks. Thirty-four comparison patients met all 3 criteria (20 women; mean age = 54.32 years, SD = 10.64; mean years of education = 13.26, SD = 2.70; Figure 1B and Table 2). Thirty-two comparison patients had lesions that involved both the gray and white matter, and two had lesions only in the white matter. Nine comparison patients had lesions that overlapped with the temporal cortex; nine, with the parietal cortex; eight, with the occipital cortex; 13, with the BG; and three, with the cerebellum (Table 2). Comparison lesions could overlap with more than one brain structure, but no lesions involved the thalamus. There was no statistically significant difference in the lesion size between patient groups (Mann–Whitney U = 190, p = .12). Demographic data for the thalamic and comparison groups, along with IQ scores from the Wechsler Adult Intelligence Scale (Full Scale, Verbal, and Performance), are presented in Tables 1 and 2. All participants gave written informed consent, and the study was approved by the University of Iowa institutional review board.

Figure 1. 

(A) Overlap of lesion sites in patients with thalamic lesions and the major thalamic nuclei. (B) Overlap of lesion sites in comparison patients.

Figure 1. 

(A) Overlap of lesion sites in patients with thalamic lesions and the major thalamic nuclei. (B) Overlap of lesion sites in comparison patients.

Neuropsychological Tests of Executive Function

To evaluate executive function, we analyzed data from two neuropsychology tests: the TMT and the WCST.

The TMT is composed of two parts—Part A and Part B. Part A requires the patient to use a pencil and connect 25 circled numbers in numeric order as fast as possible. The numbers are scattered on a page. Part A is thought to test psychomotor functions and processing speed (Bowie & Harvey, 2006). Stimuli in TMT Part B consist of both numbers and letters, and patients are asked to connect circles between them in an alternating sequence (i.e., 1-A-2-B-3-C), as fast as possible. Part B is more difficult and is considered to be a test of control-related functions that include working memory and cognitive flexibility (Kortte, Horner, & Windham, 2002; Crowe, 1998). The dependent measure of this task is the time patients took to complete the task (for each of Part A and Part B), measured in seconds. All patients' data were normalized to account for age and years of education from population-derived data (Tombaugh, 2004) and converted to z scores by subtracting the normative mean and dividing by the normative standard deviation. Higher TMT Part B z scores are indicative of executive dysfunction. It is important to note that many design variables, including visual properties of the stimuli, the number of stimuli, and the motor responses required (i.e., connecting circles), are similar for TMT Part A and Part B. Thus, impaired performance in Part B but not Part A can provide strong evidence of specificity for associating thalamic lesions with executive dysfunction, while controlling for other factors that may contribute to both tasks, such as psychomotor functions and processing speed.

For WCST, patients were asked to sort 128 cards based on dimensions of the symbol printed on each card (Color, Form, Number). The experimenter provided explicit feedback on patient's sorting choice (“right” or “wrong”), and patients were instructed to use this feedback to adjust their strategy. After 10 correct trials, the sorting dimension changed unbeknownst to the patient, and the task required the patient to infer the correct sorting dimension based on the feedback. The types of errors patients committed can be indicative of impairment of executive function. We focused on perseverative errors for this study, as increased perseverative errors are thought to be indicative of impairment in cognitive flexibility, the ability to detect contingency changes, and using feedback to switch between task sets (Gläscher, Adolphs, & Tranel, 2019; Dehaene & Changeux, 1991). In addition to the raw scores, we also examined percentile scores that are relative to population performance using normative data published in the WCST manual (Heaton, Chelune, Talley, Kay, & Curtiss, 1993).

Anatomical Analysis of Lesion Location

The anatomic location and spatial extent of each lesion were determined by consulting the available T1, T2, or computed tomography data when contraindications to MRI were present. These data were acquired using a variety of sequences over a span of 20 years. For T1 and T2 data, images for most participants were acquired with a 0.9375 × 0.9375 × 1.5 mm3 or 1 × 1 × 1 mm3 resolution; for computed tomography data, data were acquired with a 0.94 × 0.94 mm2 in-plane resolution and different slice thicknesses ranging from 2 to 5 mm. All lesions were manually traced by trained technicians and reviewed in detail by a board-certified neurologist (coauthor A. D. B.), who was blinded to neuropsychological test performance. Lesion masks were then transformed to the Montreal Neurological Institute (MNI) ICBM 152 Nonlinear Asymmetrical template version 2009c space (Fonov, Evans, McKinstry, Almli, & Collins, 2009) using Advanced Normalization Tools' antsRegistration function (Avants, Tustison, & Song, 2009). We used this registration function for spatial normalization because it employs a high-deformation, nonlinear registration procedure, which allows for local deformation to account for differences in size and shape between brain structures. This is advantageous for achieving high registration accuracy for both thalamic and cortical structures. Because lesions negatively affect the accuracy of the transformation to MNI space, we used enantiomorphic normalization, which extracts the voxel intensities from the nondamaged homologue of the lesion volume and inserts it in place of the manually defined lesion mask. As such, normal voxel intensity values are artificially inserted into the lesion space for the purpose of transforming the brain. For bilateral lesions, the lesion mask is converted to a cost-function mask and used to aid in spatial normalization (Nachev, Coulthard, Jäger, Kennard, & Husain, 2008; Brett, Leff, Rorden, & Ashburner, 2001). After transformation, lesion masks went through a second round of manual editing as needed to ensure the anatomical borders of the lesion were accurately represented on the MNI brain, supervised by coauthor A. D. B. (who was blind to the neuropsychological test data at the time). The Morel atlas was used to define the spatial extent of the thalamus and to identify the thalamic nuclei affected by each individual lesion (Krauth et al., 2010; Morel, Magnin, & Jeanmonod, 1997). This atlas was created to map human thalamic nuclei based on cytoarchitecture and myeloarchitecture in stained slices of postmortem tissue collected from five postmortem human brains and further transformed to the MNI space (Krauth et al., 2010). To test our prediction that patients with lesions to the mediodorsal thalamus would exhibit significantly more executive function impairment than patients with lesions that spared the mediodorsal thalamus, we identified patients with and without thalamic lesions that overlapped with the mediodorsal nucleus (Table 2).

Lesion Network Mapping

To identify distributed brain regions that might be functionally disrupted by focal thalamic lesions, we followed the lesion network mapping approach developed by Boes et al. (2015). The validity of this approach was supported by demonstrating that focal subcortical lesions that were heterogeneously distributed were functionally connected with cortical regions thought to be involved in symptom expression across four lesion syndromes (visual hallucinations, auditory hallucinations, aphasia, and poststroke pain; Boes et al., 2015). These findings indicate that lesion network mapping may help to localize behavioral symptoms to affected networks that can be inferred using normative data. To identify the network of brain regions potentially affected by each thalamic lesion for this study, each patient's lesion mask was used as a seed ROI to map regions that exhibit functional connectivity with the thalamic lesion site in a normative functional connectome data set.

The normative functional connectome data set consisted of resting-state fMRI data from 303 participants (mean age = 21.7 years, SD = 2.87, age range =19–27 years, 131 men). Data from these participants were acquired as part of the Brain Genomics Superstruct project (Holmes et al., 2015). We chose this data set specifically because of the high quality of thalamic functional connectivity data, which we have successfully utilized in prior work to map the functional connectivity profile for each thalamic nucleus (Hwang et al., 2017). Previous lesion mapping studies have shown that differences in age between the normative data set and patients with lesions do not significantly impact the lesion-associated networks (Albazron et al., 2019; Boes et al., 2015).

For each normative participant, two 6-min runs of fMRI data were collected using a gradient-echo EPI sequence (repetition time = 3000 msec, echo time = 30 msec, flip angle = 85°, 3 mm3 isotropic voxels with 47 axial slices). Details on MRI data preprocessing were described in our previous article (Hwang et al., 2017). Briefly, brain images were segmented into different tissue classes (white matter, gray matter, and cerebrospinal fluid) using FMRIB Software Library's FAST. These different tissue masks were used to improve cross-modal registration accuracy. Rigid body motion correction was performed, and a boundary-based registration algorithm was used to register the BOLD data with T1-weighted images. The brain-extracted T1 data were then spatially normalized to the MNI-152 space (Fonov et al., 2009) using the same “antsRegistration” function from the Advanced Normalization Tools (Avants et al., 2009) that we used to transform lesion masks, with the difference that no enantiomorphic normalization was performed given that no voxels were damaged. We then performed nuisance regression to further reduce nonneural noise (Behzadi, Restom, Liau, & Liu, 2007). Because the close physical proximity between the thalamus and the ventricles could result in blurring of the fMRI signal, we further regressed out the mean signal from cerebrospinal fluid, white matter, and gray matter that were within five voxels (10 mm) from the thalamus, and no spatial smoothing was performed. After regression, data were bandpass filtered from 0.009 to 0.08 Hz.

To map each patient's lesion-associated network, each lesion mask, after being transformed into MNI space, was used as a seed ROI to extract a mean preprocessed fMRI time series. This mean time series signal was correlated with every voxel across the whole brain, and the resulting maps were averaged across all normative participants to derive a group-averaged seed-based functional connectivity map. This resulting map represents the network of brain regions functionally connected with each thalamic lesion in the normative functional connectome. To examine the overlap of these lesion networks across patients, a voxel-wise one-sample t test was performed using the magnitude of functional connectivity estimates (correlation with the seed signal) as the dependent measure. To compare the lesion-associated networks between impaired and nonimpaired patients, a voxel-wise two-sample t test was performed on the resulting functional connectivity maps. To correct for multiple comparisons, a whole-brain FWE correction was performed using Monte Carlo simulation implemented in the 3dClustSim software from AFNI. To keep the false-positive rate under 5%, we used the updated spatial autocorrelation function to determine the minimum cluster size (Cox, Chen, Glen, Reynolds, & Taylor, 2017). We report corrected results using the cluster-forming threshold of p < .01 and a cluster size threshold of p < .05. The minimum cluster size was 811 spatially contiguous voxels.

To relate these lesion networks with previously identified cortical functional networks, we used a publicly available atlas of canonical cortical networks derived from functional connectivity data (Yeo et al., 2011). We focused on networks previously implicated in executive function, including the cingulo-opercular network (Dosenbach et al., 2007), the frontoparietal network (Cole et al., 2013; Cole & Schneider, 2007), and the dorsal attention network (Corbetta & Shulman, 2002).

Statistical Analyses

To evaluate our first hypothesis that focal thalamic lesions involving the mediodorsal thalamus are associated with impaired executive function, we compared test scores between patients with and without lesions that involved the mediodorsal thalamus using the Mann–Whitney U test, which is suitable for limited sample data sets. To account for the potential confounding effects of different lesion sizes, we performed linear regressions using test scores as the dependent variables and lesion size as the independent variable, and residual scores were used for statistical comparisons. Note that, because not every patient completed all neuropsychological tests, only patients with data from the relevant test were included in each statistical test. Specifically, 14 patients with thalamic lesions completed the TMT, 10 completed the WSCT, nine completed both, and one completed WSCT but not TMT, whereas 31 comparison patients completed the TMT, 21 completed the WCST, and 20 completed both. To correct for multiple comparisons, we performed Bonferroni correction.

To evaluate our second hypothesis that thalamic lesions associated with impaired executive function would exhibit stronger functional connectivity with the frontoparietal functional network, we used two complementary strategies. First, we compared lesion network maps of thalamic lesions associated with impaired versus preserved executive function using a mass univariate t test procedure, evaluating any regions with significant group differences. Second, we compared functional connectivity estimates of each lesion site with each canonical functional network, including the frontoparietal, cingulo-opercular, and dorsal attention networks, between thalamic patients with impaired versus preserved executive function.

RESULTS

Lesions of the Mediodorsal Thalamus Induced More Pronounced Executive Function Impairment

We examined task performances on the TMT Part A and Part B after using published normative data from healthy participants to adjust for the potential confounding effects of age and years of education. We found that, on average, patients with thalamic lesions performed 2 SDs worse than healthy norms on TMT Part B (mean z score = 2.76, SD = 3.13; Table 1). Specifically, 9 of 14 patients had a z score > 2, indicating that their performance on the TMT Part B was slower than 95% of the normative population. For the comparison patients, 8 of 34 patients had z scores of 2 or higher, and the average was within the normative range (mean z score = 0.84, SD = 1.75; Table 2). For the Trail Making Part A, 1 of 14 patients with thalamic lesions performed worse than the population norm (mean z score = 0.44, SD = 0.85; Table 1), and 3 of 45 comparison patients performed worse than the population norm (mean z score = 0.60, SD = 1.41; Table 2).

We tested our first hypothesis, whether patients with lesions that overlapped with the mediodorsal nucleus exhibited poorer performance on tests of executive function. For the 14 patients with thalamic lesions who completed TMT, 11 patients had lesions that overlapped with the mediodorsal nucleus and three patients had lesions that spared the mediodorsal nucleus (Figure 2B). We found that TMT Part B task performance for patients with mediodorsal thalamic lesions was significantly worse than that for comparison patients with lesions outside the thalamus (Mann–Whitney U = 75, corrected p = .0041; Figure 2A) and patients with thalamic lesions not involving the mediodorsal nucleus (Mann–Whitney U = 1, corrected p = .043). There was no difference in TMT Part A performance between patient groups (vs. comparison patients: Mann–Whitney U = 175.5, p = .28; vs. nonmediodorsal thalamic patients: Mann–Whitney U = 17, p = .47; Figure 2A). To better isolate executive functions from visuomotor functions, we subtracted TMT Part A scores from TMT Part B scores. We found that, after subtraction, patients with mediodorsal thalamic lesions had significantly poorer scores when compared to comparison patients with lesions outside the thalamus (Mann–Whitney U = 114, corrected p = .0075; Figure 2C) and patients with nonmediodorsal thalamic lesions (Mann–Whitney U = 1, corrected p = .043; Figure 2C).

Figure 2. 

(A) TMT test scores transformed to z scores that accounted for age and years of education and compared between groups. (B) Lesions that overlapped with the mediodorsal nucleus. MNI coordinates of peak overlapping sites: (6, −14, 2) and (−6, 19, 2). (C) Comparing TMT Part B and Part B minus Part A test scores between patients with and without lesions to the mediodorsal nucleus. The solid dot represents the group mean, and the solid bars depict the 95% bootstrapped confidence intervals. *p < .05.

Figure 2. 

(A) TMT test scores transformed to z scores that accounted for age and years of education and compared between groups. (B) Lesions that overlapped with the mediodorsal nucleus. MNI coordinates of peak overlapping sites: (6, −14, 2) and (−6, 19, 2). (C) Comparing TMT Part B and Part B minus Part A test scores between patients with and without lesions to the mediodorsal nucleus. The solid dot represents the group mean, and the solid bars depict the 95% bootstrapped confidence intervals. *p < .05.

For WCST, we had fewer observations relative to TMT with only 10 patients with thalamic lesions completing the task. Of those 10 patients, nine had lesions that overlapped with the mediodorsal thalamus; thus, we focused on comparing WCST performance with comparison patients. We did not observe a statistically significant difference in the number of perseverative errors between patients with mediodorsal thalamic lesions and comparison patients for either the raw scores (Mann–Whitney U = 168.5, p = .23) or the normative percentile scores (Mann–Whitney U = 170.5, p = .24). Other dependent measures from WCST also did not show a significant between-group difference, including the number of categories achieved (Mann–Whitney U = 175.5.5, p = .28), correct responses (Mann–Whitney U = 197.5, p = .5), or total errors (Mann–Whitney U = 162.5, p = .18). Converting the raw scores to percentile scores did not change the results, nor did accounting for the effects of differential lesion size through linear regression. Because we did not observe a statistically significant difference for WCST, we cannot make the inference on whether thalamic lesions are associated with impaired WCST performance. We however note that 2 of the 10 patients with thalamic lesions involving the mediodorsal nucleus of the thalamus showed more perseverative errors than any of the comparison patients.

Because the medial pulvinar and the anterior nucleus also project to frontal and parietal association cortices and have been implicated in executive function (Hartikainen et al., 2014; Child & Benarroch, 2013; Snow, Allen, Rafal, & Humphreys, 2009), we further compared test scores between patients with and without lesions that overlapped with these two nuclei. We compared test scores from TMT Part B between patients with and without lesions that overlapped with the anterior nucleus (n = 5 and 9, respectively) and did not find a statistically significant difference (Mann–Whitney U = 13.5, p = .12). For patients with and without (n = 3 and 11, respectively) lesions that overlapped with the medial pulvinar, we also did not find any statistically significant difference in TMT Part B test scores (Mann–Whitney U = 11, p = .17).

Impaired Patients' Lesion Networks Overlap with the Frontoparietal Network

To test our second hypothesis, we performed a lesion network mapping analysis to identify the cortical regions potentially affected by those lesions associated with impaired executive function. Nine patients with thalamic lesions who were impaired on TMT Part B (defined as exhibiting a normalized z score greater than 2) showed functional connectivity between the thalamic lesion sites with the dorsomedial PFC, the rostral PFC, the middle and superior frontal gyrus, the inferior frontal cortex, the insula, the temporal pole, the inferior parietal cortex, and the intraparietal sulcus (Figure 3A). To test our hypothesis, we evaluated which functional connectivity patterns were uniquely associated with impaired TMT Part B performance by comparing lesion-associated networks from the nine impaired patients against the five patients with preserved TMT Part B performance (defined as exhibiting a normalized z score smaller than 2), using a two-sample t test. Thalamic lesion sites associated with impaired TMT Part B task performance exhibited stronger functional connectivity with the rostral medial PFC and ACC (Figure 3B) than lesion networks derived from lesions associated with unimpaired TMT Part B performance.

Figure 3. 

(A) Lesion network of patients with thalamic lesions with impaired TMT Part B performance. Cluster corrected at p < .05. (B) Patients with impaired TMT Part B task performance exhibited significantly stronger functional connectivity with the medial PFC and ACC. Peak's MNI coordinate: (−14, 34, 16). Cluster corrected at p < .05.

Figure 3. 

(A) Lesion network of patients with thalamic lesions with impaired TMT Part B performance. Cluster corrected at p < .05. (B) Patients with impaired TMT Part B task performance exhibited significantly stronger functional connectivity with the medial PFC and ACC. Peak's MNI coordinate: (−14, 34, 16). Cluster corrected at p < .05.

In addition to the cortex-wide lesion network mapping analysis, we also compared the strength of functional connectivity with canonical functional networks implicated in executive function, including the frontoparietal, cingulo-opercular, and dorsal attention networks (Power et al., 2011; Yeo et al., 2011). We found that thalamic subregions whose damage was associated with impaired executive function exhibited stronger functional connectivity with the frontoparietal network (Figure 4; Mann–Whitney U = 8, corrected p = .031). We repeated the same analysis for other cortical functional networks and did not find any significant group differences. Altogether, these results suggest that a thalamic lesion may induce a more severe deficit in executive function if the lesioned site had strong functional connectivity with medial prefrontal regions and the frontoparietal functional network.

Figure 4. 

Patients with impaired TMT Part B task performance exhibited stronger functional connectivity with the frontoparietal functional network (top left), but not the cingulo-opercular and dorsal attention networks (left). The network parcellations were obtained from Yeo et al. (2011). Each solid dot represents the group mean, and each solid bar depicts the 95% bootstrapped confidence interval. *p < .05. The bottom right panel depicts the significant cluster from Figure 3, in which patients with impaired TMT Part B task performance exhibited significantly stronger functional connectivity (FC) with the medial prefrontal cortex and ACC. This cluster overlaps with the medial frontal portion of the frontoparietal network.

Figure 4. 

Patients with impaired TMT Part B task performance exhibited stronger functional connectivity with the frontoparietal functional network (top left), but not the cingulo-opercular and dorsal attention networks (left). The network parcellations were obtained from Yeo et al. (2011). Each solid dot represents the group mean, and each solid bar depicts the 95% bootstrapped confidence interval. *p < .05. The bottom right panel depicts the significant cluster from Figure 3, in which patients with impaired TMT Part B task performance exhibited significantly stronger functional connectivity (FC) with the medial prefrontal cortex and ACC. This cluster overlaps with the medial frontal portion of the frontoparietal network.

DISCUSSION

Our findings indicate that lesions involving the mediodorsal nucleus and thalamic subregions that are functionally connected with the dorsomedial PFC, ACC, and frontoparietal network are associated with impaired executive function. The mediodorsal thalamic nucleus has long been acknowledged to be important for executive function (Golden, Graff-Radford, Jones, & Benarroch, 2016; Mitchell, 2015). It has anatomical projections to frontal association areas (Giguere & Goldman-Rakic, 1988), is active during working memory tasks in both animal models and human studies (Peräkylä et al., 2017; Manoach, Greve, Lindgren, & Dale, 2003; Fuster & Alexander, 1971), and could mediate cerebellar and striatal influences on prefrontal function (Albazron et al., 2019; Alexander, DeLong, & Strick, 1986). Although our study is one of several studies that have reported the effects of mediodorsal thalamus lesions on executive function, our inclusion of a lesion comparison group permitted us to make stronger inferences on anatomical specificity. Specifically, past studies compared neuropsychological test performances from thalamic patients with healthy comparisons (or no comparison at all). This is problematic in that it is possible a nonspecific lesion can induce similar impairments as those observed in patients with thalamic lesions or that executive function deficits could be secondary to other common factors that were not explicitly assessed, such as overall intelligence, psychomotor functions, and processing speed. Our findings cannot be attributed to differences in overall intelligence, as the IQ scores were comparable between patient groups. Furthermore, given that we observed impaired performance for TMT Part B but not Part A, our results also cannot be explained by differences in processing speed and psychomotor functions that contribute to both tasks. Thus, by demonstrating task specificity, contrasting thalamic lesions with lesions outside the thalamus (but with comparable lesion size), and evaluating the localization of mediodorsal versus nonmediodorsal lesions, we demonstrate a compelling association between the integrity of the mediodorsal thalamus and performance in a task requiring executive function.

Extending beyond localizing thalamic subregions where damage is associated with impaired executive function, our lesion network mapping investigation further identified ACC and the dorsomedial PFC as brain regions that may interact with the mediodorsal region of the thalamus in tasks involving executive function. This finding is consistent with a previous large-scale lesion symptom mapping study (Gläscher et al., 2012), which demonstrated that ACC and the dorsomedial PFC are necessary for performing the TMT Part B. Both our study and the Gläscher et al. (2012) study utilized TMT Part B as the neuropsychological measure of executive function. The TMT has several advantages, including a well-established population norm and ease of administration in clinical settings. However, it is important to acknowledge that TMT Part B alone is not sufficient to identify the component processes impaired. Executive function is not a unitary function, and TMT Part B likely depends on several component processes including working memory, attention, task-set maintenance, and cognitive flexibility (Kortte et al., 2002; Miyake et al., 2000; Crowe, 1998). Some of these functions have been shown to modulate medial prefrontal activity (Nee, Kastner, & Brown, 2011; Rushworth, Hadland, Gaffan, & Passingham, 2003) and thus may be differentially impacted by thalamic lesions or changes in thalamic connectivity with medial PFC. We are not in a position to comment on specific component processes of executive function disrupted by thalamic lesions—these are questions that should be specifically addressed in future studies.

Resting-state functional connectivity studies have consistently identified several cortical functional networks that include frontal and parietal association areas, including the frontoparietal network (Vincent et al., 2008), the dorsal attention network (Fox et al., 2006), and the cingulo-opercular network (Seeley et al., 2007). However, before our study, we did not know which of these networks may have functional connectivity to sites in the thalamus that, when lesioned, contribute to executive dysfunction. Our results suggest that lesions of the mediodorsal thalamus are associated with impaired executive function, and before injury, these lesioned sites have strong functional connectivity with the frontoparietal network, a network that has been shown to be correlated with TMT Part B performance (Seeley et al., 2007). The mediodorsal thalamus has network properties that are well suited for executive function. For example, mediodorsal thalamus has strong “connector hub” properties (Greene et al., 2020; Hwang et al., 2017), making it well positioned to relay the top–down biasing signals from frontoparietal regions to other functional regions or networks, a mechanism that was previously hypothesized for cognitive control (Cole et al., 2013; Miller & Cohen, 2001). Furthermore, interactions among frontal and parietal association cortices are known to be modulated by executive function (Bowling, Friston, & Hopfinger, 2020). Anatomically, the mediodorsal thalamus is known to have connectivity with both lateral and parietal cortices (Selemon & Goldman-Rakic, 1988), so another potential mechanism of thalamic involvement is that the mediodorsal thalamus mediates the interaction between frontal and parietal regions for executive function.

Thus far, because of the fact that we did not observe a reliable group difference for WCST performance, we restricted our discussion and interpretation on results from the TMT. The two patients who showed impaired WCST performance have mediodorsal thalamic lesions, but there were also patients with mediodorsal thalamic lesions who showed normal WCST performance. Our result appears to contradict a previous study showing that medial thalamic lesions are associated with poorer WCST performance (Liebermann et al., 2013). However, in that study, no comparison patients were included to establish the effects of focal thalamic lesions; thus, the observed impairment could be attributed to nonspecific lesion effects that do not involve the thalamus. Furthermore, similar to our observation that only two patients exhibited impaired WCST performance, not every patient in this prior study was impaired. There are several speculative reasons for our null result. First, we might not have a sufficient sample size; in contrast to the 15 thalamic patients who completed the TMT, only 10 patients completed WCST. A second potential reason is that WCST may depend on a distinct collection of cognitive processes, some of which might engage different cortical and subcortical regions when compared to the TMT. For example, no explicit feedback was given during the TMT tasks, whereas using feedback to adjust cognitive strategies is an essential component of WCST (Gläscher et al., 2019). It is possible that this feedback learning process involves a distinct anatomical substrate. This possibility is supported by a study showing that patients with prefrontal lesions do not reliably show WCST impairment when compared to patients with lesions sparing PFC (Anderson, Damasio, Jones, & Tranel, 1991). A third potential reason is that successful WCST task performance in patients with focal thalamic lesions might recruit thalamic regions outside the mediodorsal thalamus. It is possible that other thalamic nuclei are more important for WCST performance under normal circumstances or other regions are able to compensate in the setting of a mediodorsal lesion. For example, previous neuroimaging studies have found increased activity in the posterior parietal cortex, the TPJ, the BG, and the cerebellum associated with WCST (Lie, Specht, Marshall, & Fink, 2006). One speculation is that these brain regions may contribute to normal WCST performance in our patients. Furthermore, whereas the number of perseverative errors in WCST may be more closely related to feedback adjustment and flexibility in switching task rules, TMT Part B performance may be more closely associated with the ability of using working memory for prospective actions (Sánchez-Cubillo et al., 2009). Specifically, the alternating sequence in TMT Part B requires participants to utilize the memorized content in working memory (i.e., number or letters) to guide their actions. The mediodorsal thalamus has been shown to exhibit sustained delayed activity that encodes prospective actions (Watanabe & Funahashi, 2012), which is particularly relevant for TMT Part B. Thus, the mediodorsal thalamus may be more closely related to working-memory-related functions, and the differential need for working memory between TMT and WCST may explain the differences in performance from patients with medial thalamic lesions. Another potential explanation is that WCST and TMT utilize different outcome measures, accuracy versus RT, and different measures may have different sensitivity. These speculations will need to be tested in future studies.

Our study has several limitations. First, to determine the effects of thalamic lesions on executive function, we included comparison patients with a similar lesion size to ensure that differences in neuropsychological outcomes cannot be attributed more generally to having a focal brain lesion or to differences in lesion size. Given that we did not control for comparison patients' lesion location, our approach cannot determine whether the thalamus or a specific cortical region is more important for executive control, which was not our aim here. This question will have to be answered by a future study that includes patients with focal thalamic lesions, those with focal cortical lesions in a specific region, and comparison patients with lesions outside the focus of study. Second, naturally occurring thalamic lesions typically involve multiple nuclei, and the resolution of imaging and lesion tracing only allows an approximation of the most likely nuclei lesioned. Given that all our patients with anterior nucleus lesions also overlapped with the mediodorsal thalamus, we could not determine whether these two nuclei have differential involvement in executive functions. Furthermore, both the lateral and medial frontal cortices, regions known to be involved in executive function and identified by our lesion network analysis, project to both the mediodorsal and anterior thalamic nuclei (Xiao, Zikopoulos, & Barbas, 2009). As such, we acknowledge the possibility that other nuclei in the region of the mediodorsal nucleus are also contributing to the observed deficits. Related to this topic, the normative functional connectome data that we utilized had a spatial resolution of 3-mm isotropic voxel. The sensitivity and accuracy of the lesion network analysis can be improved by acquiring data with higher spatial resolution. High-resolution fMRI data can be particularly useful for differentiating the anterior nucleus from the mediodorsal nucleus; both are relatively small, spatially close by, and project to different medial frontal regions. Similarly, the accuracy of MRI spatial normalization needs further validation, ideally by comparing MRI registration results with histological analyses to identify specific thalamic nuclei. Future studies should address this open question using improved methods that can more accurately map small thalamic nuclei, such as 7-T MRI. Furthermore, our lesion network mapping approach utilized resting-state fMRI data. It is likely that different executive functions may engage task-dependent coupling between the mediodorsal thalamus and different subregions in the frontal cortex. This task-specific connectivity pattern can only be captured by task fMRI.

A wide range of cognitive impairments have been reported in patients with focal thalamic lesions, including aphasia (Crosson et al., 1986; Graff-Radford, Eslinger, Damasio, & Yamada, 1984), amnesia (Pergola et al., 2016; Graff-Radford, Tranel, Van Hoesen, & Brandt, 1990; von Cramon, Hebel, & Schuri, 1985), attention deficits (de Bourbon-Teles et al., 2014; Snow et al., 2009), and executive dysfunction (Liebermann et al., 2013). Patients with thalamic injuries can sometimes exhibit deficits across multiple domains (Van der Werf et al., 2003). This heterogeneity can likely be attributed to the thalamus's complex connectivity profile. The thalamus is one of the most globally connected structures, and each thalamic nucleus is known to have a partially distinct pattern of connectivity to a system of multiple cortical and subcortical regions. Each system is potentially associated with distinct cognitive functions. Therefore, each thalamic subregion is likely contributing to a specific (or multiple) cognitive function(s), constrained by its connectivity profile (Little et al., 2010). Behavioral deficits may in turn be attributed to the specific brain regions and connectivity among these regions disrupted by the thalamic injury. The principal contribution of our study was to test this prediction, and we were able to identify a specific thalamic subregion in the mediodorsal nucleus with a distinct connectivity pattern to the medial PFC associated with impaired executive function. To conclude, our results indicate that focal thalamic lesions involving the mediodorsal nucleus of the thalamus are associated with impaired executive function, and deficits could result from disrupted connectivity with the dorsomedial PFC, ACC, and the frontoparietal network.

Acknowledgments

K. H. was supported by R01MH122613. D. T. was supported by P50 MH094258 and the Kiwanis Neuroscience Research Foundation. A. D. B. was supported by R01NS114405 and R21MH120441. We thank Fatimah Albazron and Kenneth Manzel for organizing the neuropsychological test data. Portions of this work were conducted on an MRI instrument funded by 1S10OD025025-01.

Reprint requests should be sent to Kai Hwang, Department of Psychological and Brain Sciences, The University of Iowa, G60 PBSB, 340 Iowa Ave., Iowa City, IA 52245, or via e-mail: kai-hwang@uiowa.edu.

REFERENCES

REFERENCES
Albazron
,
F. M.
,
Bruss
,
J.
,
Jones
,
R. M.
,
Yock
,
T. I.
,
Pulsifer
,
M. B.
,
Cohen
,
A. L.
, et al
(
2019
).
Pediatric postoperative cerebellar cognitive affective syndrome follows outflow pathway lesions
.
Neurology
,
93
,
e1561
e1571
.
Alexander
,
G. E.
,
DeLong
,
M. R.
, &
Strick
,
P. L.
(
1986
).
Parallel organization of functionally segregated circuits linking basal ganglia and cortex
.
Annual Review of Neuroscience
,
9
,
357
381
.
Anderson
,
S. W.
,
Damasio
,
H.
,
Jones
,
R. D.
, &
Tranel
,
D.
(
1991
).
Wisconsin card sorting test performance as a measure of frontal lobe damage
.
Journal of Clinical and Experimental Neuropsychology
,
13
,
909
922
.
Avants
,
B. B.
,
Tustison
,
N.
, &
Song
,
G.
(
2009
).
Advanced normalization tools (ANTS)
.
Insight Journal
,
2
,
1
35
.
Badre
,
D.
,
Hoffman
,
J.
,
Cooney
,
J. W.
, &
D'Esposito
,
M.
(
2009
).
Hierarchical cognitive control deficits following damage to the human frontal lobe
.
Nature Neuroscience
,
12
,
515
522
.
Behzadi
,
Y.
,
Restom
,
K.
,
Liau
,
J.
, &
Liu
,
T. T.
(
2007
).
A component based noise correction method (CompCor) for BOLD and perfusion based fMRI
.
Neuroimage
,
37
,
90
101
.
Boes
,
A. D.
,
Prasad
,
S.
,
Liu
,
H.
,
Liu
,
Q.
,
Pascual-Leone
,
A.
,
Caviness
,
V. S.
, Jr.
, et al
(
2015
).
Network localization of neurological symptoms from focal brain lesions
.
Brain
,
138
,
3061
3075
.
Bowie
,
C. R.
, &
Harvey
,
P. D.
(
2006
).
Administration and interpretation of the trail making test
.
Nature Protocols
,
1
,
2277
2281
.
Bowling
,
J. T.
,
Friston
,
K. J.
, &
Hopfinger
,
J. B.
(
2020
).
Top–down versus bottom–up attention differentially modulate frontal–parietal connectivity
.
Human Brain Mapping
,
41
,
928
942
.
Brett
,
M.
,
Leff
,
A. P.
,
Rorden
,
C.
, &
Ashburner
,
J.
(
2001
).
Spatial normalization of brain images with focal lesions using cost function masking
.
Neuroimage
,
14
,
486
500
.
Carrera
,
E.
, &
Tononi
,
G.
(
2014
).
Diaschisis: Past, present, future
.
Brain
,
137
,
2408
2422
.
Child
,
N. D.
, &
Benarroch
,
E. E.
(
2013
).
Anterior nucleus of the thalamus: Functional organization and clinical implications
.
Neurology
,
81
,
1869
1876
.
Cole
,
M. W.
,
Reynolds
,
J. R.
,
Power
,
J. D.
,
Repovs
,
G.
,
Anticevic
,
A.
, &
Braver
,
T. S.
(
2013
).
Multi-task connectivity reveals flexible hubs for adaptive task control
.
Nature Neuroscience
,
16
,
1348
1355
.
Cole
,
M. W.
, &
Schneider
,
W.
(
2007
).
The cognitive control network: Integrated cortical regions with dissociable functions
.
Neuroimage
,
37
,
343
360
.
Corbetta
,
M.
, &
Shulman
,
G. L.
(
2002
).
Control of goal-directed and stimulus-driven attention in the brain
.
Nature Reviews Neuroscience
,
3
,
201
215
.
Cox
,
R. W.
,
Chen
,
G.
,
Glen
,
D. R.
,
Reynolds
,
R. C.
, &
Taylor
,
P. A.
(
2017
).
fMRI clustering in AFNI: False-positive rates redux
.
Brain Connectivity
,
7
,
152
171
.
Crosson
,
B.
,
Parker
,
J. C.
,
Kim
,
A. K.
,
Warren
,
R. L.
,
Kepes
,
J. J.
, &
Tully
,
R.
(
1986
).
A case of thalamic aphasia with postmortem verification
.
Brain and Language
,
29
,
301
314
.
Crowe
,
S. F.
(
1998
).
The differential contribution of mental tracking, cognitive flexibility, visual search, and motor speed to performance on parts A and B of the trail making test
.
Journal of Clinical Psychology
,
54
,
585
591
.
de Bourbon-Teles
,
J.
,
Bentley
,
P.
,
Koshino
,
S.
,
Shah
,
K.
,
Dutta
,
A.
,
Malhotra
,
P.
, et al
(
2014
).
Thalamic control of human attention driven by memory and learning
.
Current Biology
,
24
,
993
999
.
Dehaene
,
S.
, &
Changeux
,
J. P.
(
1991
).
The Wisconsin card sorting test: Theoretical analysis and modeling in a neuronal network
.
Cerebral Cortex
,
1
,
62
79
.
Desikan
,
R. S.
,
Ségonne
,
F.
,
Fischl
,
B.
,
Quinn
,
B. T.
,
Dickerson
,
B. C.
,
Blacker
,
D.
, et al
(
2006
).
An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest
.
Neuroimage
,
31
,
968
980
.
Dosenbach
,
N. U. F.
,
Fair
,
D. A.
,
Miezin
,
F. M.
,
Cohen
,
A. L.
,
Wenger
,
K. K.
,
Dosenbach
,
R. A. T.
, et al
(
2007
).
Distinct brain networks for adaptive and stable task control in humans
.
Proceedings of the National Academy of Sciences, U.S.A.
,
104
,
11073
11078
.
Fonov
,
V. S.
,
Evans
,
A. C.
,
McKinstry
,
R. C.
,
Almli
,
C. R.
, &
Collins
,
D. L.
(
2009
).
Unbiased nonlinear average age-appropriate brain templates from birth to adulthood
.
Neuroimage
,
47(Suppl. 1)
,
S102
.
Fox
,
M. D.
,
Corbetta
,
M.
,
Snyder
,
A. Z.
,
Vincent
,
J. L.
, &
Raichle
,
M. E.
(
2006
).
Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems
.
Proceedings of the National Academy of Sciences, U.S.A.
,
103
,
10046
10051
.
Fuster
,
J. M.
, &
Alexander
,
G. E.
(
1971
).
Neuron activity related to short-term memory
.
Science
,
173
,
652
654
.
Giguere
,
M.
, &
Goldman-Rakic
,
P. S.
(
1988
).
Mediodorsal nucleus: Areal, laminar, and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys
.
Journal of Comparative Neurology
,
277
,
195
213
.
Gläscher
,
J.
,
Adolphs
,
R.
,
Damasio
,
H.
,
Bechara
,
A.
,
Rudrauf
,
D.
,
Calamia
,
M.
, et al
(
2012
).
Lesion mapping of cognitive control and value-based decision making in the prefrontal cortex
.
Proceedings of the National Academy of Sciences, U.S.A.
,
109
,
14681
14686
.
Gläscher
,
J.
,
Adolphs
,
R.
, &
Tranel
,
D.
(
2019
).
Model-based lesion mapping of cognitive control using the Wisconsin card sorting test
.
Nature Communications
,
10
,
20
.
Golden
,
E. C.
,
Graff-Radford
,
J.
,
Jones
,
D. T.
, &
Benarroch
,
E. E.
(
2016
).
Mediodorsal nucleus and its multiple cognitive functions
.
Neurology
,
87
,
2161
2168
.
Goldman-Rakic
,
P. S.
, &
Porrino
,
L. J.
(
1985
).
The primate mediodorsal (MD) nucleus and its projection to the frontal lobe
.
Journal of Comparative Neurology
,
242
,
535
560
.
Graff-Radford
,
N. R.
,
Eslinger
,
P. J.
,
Damasio
,
A. R.
, &
Yamada
,
T.
(
1984
).
Nonhemorrhagic infarction of the thalamus: Behavioral, anatomic, and physiologic correlates
.
Neurology
,
34
,
14
23
.
Graff-Radford
,
N. R.
,
Tranel
,
D.
,
Van Hoesen
,
G. W.
, &
Brandt
,
J. P.
(
1990
).
Diencephalic amnesia
.
Brain
,
113
,
1
25
.
Greene
,
D. J.
,
Marek
,
S.
,
Gordon
,
E. M.
,
Siegel
,
J. S.
,
Gratton
,
C.
,
Laumann
,
T. O.
, et al
(
2020
).
Integrative and network-specific connectivity of the basal ganglia and thalamus defined in individuals
.
Neuron
,
105
,
742
758
.
Hartikainen
,
K. M.
,
Sun
,
L.
,
Polvivaara
,
M.
,
Brause
,
M.
,
Lehtimäki
,
K.
,
Haapasalo
,
J.
, et al
(
2014
).
Immediate effects of deep brain stimulation of anterior thalamic nuclei on executive functions and emotion–attention interaction in humans
.
Journal of Clinical and Experimental Neuropsychology
,
36
,
540
550
.
Heaton
,
R. K.
,
Chelune
,
G. J.
,
Talley
,
J. L.
,
Kay
,
G. G.
, &
Curtiss
,
G.
(
1993
).
Wisconsin card sorting test, revised and expanded
.
Odessa, FL
:
Psychological Assessment Resources
.
Holmes
,
A. J.
,
Hollinshead
,
M. O.
,
O'Keefe
,
T. M.
,
Petrov
,
V. I.
,
Fariello
,
G. R.
,
Wald
,
L. L.
, et al
(
2015
).
Brain genomics superstruct project initial data release with structural, functional, and behavioral measures
.
Scientific Data
,
2
,
150031
.
Hwang
,
K.
,
Bertolero
,
M. A.
,
Liu
,
W. B.
, &
D'Esposito
,
M.
(
2017
).
The human thalamus is an integrative hub for functional brain networks
.
Journal of Neuroscience
,
37
,
5594
5607
.
Kortte
,
K. B.
,
Horner
,
M. D.
, &
Windham
,
W. K.
(
2002
).
The trail making test, part B: Cognitive flexibility or ability to maintain set?
Applied Neuropsychology
,
9
,
106
109
.
Krauth
,
A.
,
Blanc
,
R.
,
Poveda
,
A.
,
Jeanmonod
,
D.
,
Morel
,
A.
, &
Székely
,
G.
(
2010
).
A mean three-dimensional atlas of the human thalamus: Generation from multiple histological data
.
Neuroimage
,
49
,
2053
2062
.
Lezak
,
M. D.
,
Howieson
,
D. B.
,
Bigler
,
E. D.
, &
Tranel
,
D.
(
2012
).
Neuropsychological assessment
(
Vol. 5
, 5th ed., p.
1161
).
New York
:
Oxford University Press
.
Lie
,
C.-H.
,
Specht
,
K.
,
Marshall
,
J. C.
, &
Fink
,
G. R.
(
2006
).
Using fMRI to decompose the neural processes underlying the Wisconsin card sorting test
.
Neuroimage
,
30
,
1038
1049
.
Liebermann
,
D.
,
Ploner
,
C. J.
,
Kraft
,
A.
,
Kopp
,
U. A.
, &
Ostendorf
,
F.
(
2013
).
A dysexecutive syndrome of the medial thalamus
.
Cortex
,
49
,
40
49
.
Little
,
D. M.
,
Kraus
,
M. F.
,
Joseph
,
J.
,
Geary
,
E. K.
,
Susmaras
,
T.
,
Zhou
,
X. J.
, et al
(
2010
).
Thalamic integrity underlies executive dysfunction in traumatic brain injury
.
Neurology
,
74
,
558
564
.
Manoach
,
D. S.
,
Greve
,
D. N.
,
Lindgren
,
K. A.
, &
Dale
,
A. M.
(
2003
).
Identifying regional activity associated with temporally separated components of working memory using event-related functional MRI
.
Neuroimage
,
20
,
1670
1684
.
Miller
,
E. K.
, &
Cohen
,
J. D.
(
2001
).
An integrative theory of prefrontal cortex function
.
Annual Review of Neuroscience
,
24
,
167
202
.
Mitchell
,
A. S.
(
2015
).
The mediodorsal thalamus as a higher order thalamic relay nucleus important for learning and decision-making
.
Neuroscience & Biobehavioral Reviews
,
54
,
76
88
.
Miyake
,
A.
,
Friedman
,
N. P.
,
Emerson
,
M. J.
, &
Witzki
,
A. H.
(
2000
).
The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis
.
Cognitive Psychology
,
41
,
49
100
.
Morel
,
A.
,
Magnin
,
M.
, &
Jeanmonod
,
D.
(
1997
).
Multiarchitectonic and stereotactic atlas of the human thalamus
.
Journal of Comparative Neurology
,
387
,
588
630
.
Nachev
,
P.
,
Coulthard
,
E.
,
Jäger
,
H. R.
,
Kennard
,
C.
, &
Husain
,
M.
(
2008
).
Enantiomorphic normalization of focally lesioned brains
.
Neuroimage
,
39
,
1215
1226
.
Nee
,
D. E.
,
Kastner
,
S.
, &
Brown
,
J. W.
(
2011
).
Functional heterogeneity of conflict, error, task-switching, and unexpectedness effects within medial prefrontal cortex
.
Neuroimage
,
54
,
528
540
.
Peräkylä
,
J.
,
Sun
,
L.
,
Lehtimäki
,
K.
,
Peltola
,
J.
,
Öhman
,
J.
,
Möttönen
,
T.
, et al
(
2017
).
Causal evidence from humans for the role of mediodorsal nucleus of the thalamus in working memory
.
Journal of Cognitive Neuroscience
,
29
,
2090
2102
.
Pergola
,
G.
,
Danet
,
L.
,
Barbeau
,
E. J.
,
Eustache
,
P.
,
Planton
,
M.
,
Raposo
,
N.
, et al
(
2016
).
Review of thalamic amnesia after infarct: The role of the mammillothalamic tract and mediodorsal nucleus
.
Neurology
,
86
,
1928
.
Power
,
J. D.
,
Cohen
,
A. L.
,
Nelson
,
S. M.
,
Wig
,
G. S.
,
Barnes
,
K. A.
,
Church
,
J. A.
, et al
(
2011
).
Functional network organization of the human brain
.
Neuron
,
72
,
665
678
.
Rushworth
,
M. F. S.
,
Hadland
,
K. A.
,
Gaffan
,
D.
, &
Passingham
,
R. E.
(
2003
).
The effect of cingulate cortex lesions on task switching and working memory
.
Journal of Cognitive Neuroscience
,
15
,
338
353
.
Sánchez-Cubillo
,
I.
,
Periáñez
,
J. A.
,
Adrover-Roig
,
D.
,
Rodríguez-Sánchez
,
J. M.
,
Ríos-Lago
,
M.
,
Tirapu
,
J.
, et al
(
2009
).
Construct validity of the trail making test: Role of task-switching, working memory, inhibition/interference control, and visuomotor abilities
.
Journal of the International Neuropsychological Society
,
15
,
438
450
.
Schmahmann
,
J. D.
(
2003
).
Vascular syndromes of the thalamus
.
Stroke
,
34
,
2264
2278
.
Seeley
,
W. W.
,
Menon
,
V.
,
Schatzberg
,
A. F.
,
Keller
,
J.
,
Glover
,
G. H.
,
Kenna
,
H.
, et al
(
2007
).
Dissociable intrinsic connectivity networks for salience processing and executive control
.
Journal of Neuroscience
,
27
,
2349
2356
.
Selemon
,
L. D.
, &
Goldman-Rakic
,
P. S.
(
1988
).
Common cortical and subcortical targets of the dorsolateral prefrontal and posterior parietal cortices in the rhesus monkey: Evidence for a distributed neural network subserving spatially guided behavior
.
Journal of Neuroscience
,
8
,
4049
4068
.
Shine
,
J. M.
,
Bissett
,
P. G.
,
Bell
,
P. T.
,
Koyejo
,
O.
,
Balsters
,
J. H.
,
Gorgolewski
,
K. J.
, et al
(
2016
).
The dynamics of functional brain networks: Integrated network states during cognitive task performance
.
Neuron
,
92
,
544
554
.
Snow
,
J. C.
,
Allen
,
H. A.
,
Rafal
,
R. D.
, &
Humphreys
,
G. W.
(
2009
).
Impaired attentional selection following lesions to human pulvinar: evidence for homology between human and monkey
.
Proceedings of the National Academy of Sciences, U.S.A.
,
106
,
4054
4059
.
Tombaugh
,
T. N.
(
2004
).
Trail making test A and B: Normative data stratified by age and education
.
Archives of Clinical Neuropsychology
,
19
,
203
214
.
Tsuchida
,
A.
, &
Fellows
,
L. K.
(
2013
).
Are core component processes of executive function dissociable within the frontal lobes? Evidence from humans with focal prefrontal damage
.
Cortex
,
49
,
1790
1800
.
Van der Werf
,
Y. D.
,
Scheltens
,
P.
,
Lindeboom
,
J.
,
Witter
,
M. P.
,
Uylings
,
H. B. M.
, &
Jolles
,
J.
(
2003
).
Deficits of memory, executive functioning and attention following infarction in the thalamus; a study of 22 cases with localised lesions
.
Neuropsychologia
,
41
,
1330
1344
.
Vincent
,
J. L.
,
Kahn
,
I.
,
Snyder
,
A. Z.
,
Raichle
,
M. E.
, &
Buckner
,
R. L.
(
2008
).
Evidence for a frontoparietal control system revealed by intrinsic functional connectivity
.
Journal of Neurophysiology
,
100
,
3328
3342
.
von Cramon
,
D. Y.
,
Hebel
,
N.
, &
Schuri
,
U.
(
1985
).
A contribution to the anatomical basis of thalamic amnesia
.
Brain
,
108
,
993
1008
.
Von Monakow
,
C.
(
1911
).
Localization of brain functions
.
Journal für Psychologie und Neurologie
,
17
,
185
200
.
Watanabe
,
Y.
, &
Funahashi
,
S.
(
2012
).
Thalamic mediodorsal nucleus and working memory
.
Neuroscience & Biobehavioral Reviews
,
36
,
134
142
.
Xiao
,
D.
,
Zikopoulos
,
B.
, &
Barbas
,
H.
(
2009
).
Laminar and modular organization of prefrontal projections to multiple thalamic nuclei
.
Neuroscience
,
161
,
1067
1081
.
Yeo
,
B. T. T.
,
Krienen
,
F. M.
,
Sepulcre
,
J.
,
Sabuncu
,
M. R.
,
Lashkari
,
D.
,
Hollinshead
,
M.
, et al
(
2011
).
The organization of the human cerebral cortex estimated by intrinsic functional connectivity
.
Journal of Neurophysiology
,
106
,
1125
1165
.