Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Aug 19.
Published in final edited form as: Psychopharmacology (Berl). 2020 Apr 19;237(6):1873–1883. doi: 10.1007/s00213-020-05505-z

Intrinsic connections between thalamic sub-regions and the lateral prefrontal cortex are differentially impacted by acute methylphenidate

Adam X Gorka 1, Tiffany R Lago 1, Nicholas Balderston 1,2, Salvatore Torrisi 1, Bari Fuchs 3, Christian Grillon 1, Monique Ernst 1
PMCID: PMC7437544  NIHMSID: NIHMS1617803  PMID: 32307560

Abstract

Background

The thalamus is a major target of dopaminergic projections and is densely connected with the prefrontal cortex. A better understanding of how dopamine changes thalamo-cortical communication may shed light on how dopamine supports cognitive function. Methylphenidate has been shown to facilitate cognitive processing and reduce connectivity between the thalamus and lateral prefrontal cortex.

Aims

The thalamus is a heterogeneous structure, and the present study sought to clarify how the intrinsic connections of thalamic sub-regions are differentially impacted by acute dopamine transporter blockade.

Methods

Sixty healthy volunteers were orally administered either 20 mg of methylphenidate (N = 29) or placebo (N = 31) in a double-blind, randomized, between-subject design. Multi-echo fMRI was used to assess intrinsic functional connectivity of sub-regions of the thalamus during a resting state scan. An N-back working-memory paradigm provided a measure of cognitive performance.

Results

Acute methylphenidate significantly reduced connectivity of the lateral prefrontal cortex with the motor and somatosensory sub-regions of the thalamus and reduced connectivity with the parietal and visual sub-regions at a trend level. Connectivity with the premotor, prefrontal, and temporal sub-regions was not impacted. The intrinsic connectivity between the thalamus and the lateral prefrontal cortex was not associated with working-memory performance.

Conclusions

Methylphenidate decreases functional connections between the lateral prefrontal cortex and thalamus broadly, while sparing intrinsic connectivity with thalamic sub-regions involved with working-memory and language related processes. Collectively, our results suggest that the dopamine transporter regulates functional connections between the prefrontal cortex and non-cognitive areas of the thalamus.

Keywords: Acute methylphenidate, Thalamus, Prefrontal cortex, Intrinsic functional connectivity, Resting state fMRI, Dopamine

Introduction

The thalamus is a heterogeneous midline structure that impacts cortical activity by serving as a relay for sensory and motor signaling and by facilitating cortico-cortical communication via its association nuclei (Sherman 2011). Reciprocal connections between the thalamus and prefrontal cortex support cognitive operations, including the maintenance of information during multi-tasking and the suppression of distracting information (Pergola et al. 2018; Zikopoulos and Barbas 2006). Pharmacological augmentation of dopamine signaling improves cognitive performance (Spencer et al. 2015), which may in part be mediated by the effect of dopamine on the thalamus. The thalamus receives dopaminergic projections from the midbrain, lateral parabrachial nucleus of the pons, hypothalamus, and periaqueductal gray matter and exhibits some of the densest dopaminergic innervation of the brain, outside of the striatum (García-Cabezas et al. 2007; Sánchez-González et al. 2005). However, there are significant regional differences in the termination patterns of the dopaminergic afferents within the thalamus suggesting that specific connections between the thalamus and prefrontal cortex are differentially affected by dopamine signaling. A better understanding of how dopamine function influences thalamo-cortical communication may shed light on the role of dopamine in supporting high-level cognitive function.

The lateral prefrontal cortex is critically involved with working memory (Owen et al. 1998) and language processing (Koelsch et al. 2009). Researchers have suggested that the maintenance and rehearsal of speech-based information (i.e., the phonological loop) is an important aspect of working memory (Baddeley 1992). Additionally, evidence suggests that the mediodorsal (MD) and ventral anterior (VA) nuclei of the thalamus support cognitive processes (Mitchell 2015) and play a critical role in the cognitive and motor aspects of speech (Barbas et al. 2013). Importantly, dopamine signaling modulates the function of the prefrontal cortex and is important for working memory processes (Ranganath and Jacob 2016). However, it is presently unclear how communication between the thalamus and prefrontal cortex is impacted by dopamine.

Methylphenidate increases dopaminergic signaling through the blockade of dopamine transporters (Volkow et al. 1998). Acute methylphenidate reduces the short-range functional connectivity of the thalamus (Konova et al. 2015), and research suggests that short-range connectivity reflects intra-regional functional connections (Tomasi and Volkow 2012). Additionally, acute methylphenidate decreases intrinsic functional connectivity (iFC) between the thalamus and the lateral prefrontal cortex (Farr et al. 2014; Mueller et al. 2014). However, researchers have not consistently replicated these findings (Konova et al. 2013; Ramaekers et al. 2013). Acute methylphenidate may differentially decrease iFC between the prefrontal cortex and sub-regions of the thalamus based on regional differences in the expression of the dopamine transporter. This functional heterogeneity may explain inconsistent findings within the literature. It is possible that decreased iFC between the lateral prefrontal cortex and the thalamus reflects either direct connections which support the maintenance of information within working memory or indirect connections which support the suppression of distracting information.

Evidence suggests that direct projections between the thalamus and prefrontal cortex can support working memory processes. The MD and VA nuclei of the thalamus both exhibit anatomical connections with the prefrontal cortex (Collins et al. 2018; Xiaob and Barbas 2004), and lesions of the MD (Kupferschmidt and Gordon 2018) and VA (Tanaka 2007) nuclei impair working memory. Additionally, areas of the VA nucleus are activated in human participants during working memory paradigms (de Bourbon-Teles et al. 2014). Activation of the MD nucleus amplifies local prefrontal network connectivity (Schmitt et al. 2017) while MD inhibition disrupts synchrony with the prefrontal cortex and impairs working memory processes (Parnaudeau et al. 2013). It is possible that acute methylphenidate may function to reduce iFC between the lateral prefrontal cortex and regions of the thalamus consistent with the MD and VA nuclei (i.e., the prefrontal sub-region of the thalamus), which support the maintenance of information within working memory.

Alternatively, evidence suggests that indirect projections between the thalamus and prefrontal cortex may support the suppression of distracting information within working memory. The inhibitory thalamic reticular nucleus (TRN) is a thin and widely diffuse structure surrounding the thalamus. Prefrontal projections to the TRN are thought to regulate attentional resources and to suppress distracting information (Zikopoulos and Barbas 2006). Indeed, prefrontal projections can inhibit thalamic sensory nuclei, via the TRN, during cross-modal divided-attention tasks (Wimmer et al. 2015). Researchers have suggested that dopamine impacts neural gain and results in neurons being more sensitive to signals and less sensitive to noise (Hauser et al. 2016). Therefore, an alternative possibility is that acute methylphenidate functions to reduce iFC between the lateral prefrontal cortex and non-prefrontal regions of the thalamus, which may reflect the suppression of distracting information within working memory.

Previous research has utilized thalamic regions of interest (Behrens et al. 2003) to demonstrate that acute methylphenidate changes the intrinsic connectivity of individual thalamic sub-regions (Demiral et al. 2018). However, to date, no research has directly compared prefrontal iFC across sub-regions of the thalamus to determine which areas are most impacted by the influence of acute methylphenidate. We set out to expand upon these findings by determining which areas of the thalamus exhibit reduced iFC with the lateral prefrontal cortex during the administration of 20 mg of acute methylphenidate. Previous research from our lab has demonstrated that a dose of 20 mg can impact working memory processes (Ernst et al. 2016). A reduction in iFC between the lateral prefrontal cortex and the prefrontal sub-region of the thalamus may reflect the maintenance of information within working memory via direct projections between the thalamus and prefrontal cortex. Alternatively, reduction in iFC between the lateral prefrontal cortex and the thalamus broadly, with the exception of the prefrontal sub-region, may reflect the suppression of distracting information via indirect projections from the prefrontal cortex to the inhibitory TRN nucleus. We hypothesize that methylphenidate will reduce functional connectivity between the lateral prefrontal cortex and thalamus and that this effect will be specific to non-prefrontal regions of the thalamus, reflecting the suppression of sources of distracting information. The suppression of distracting information would be consistent with increased neural gain which results from increased dopaminergic signaling (Hauser et al. 2016). Additionally, we hypothesize that reduced connectivity between the lateral prefrontal cortex and the thalamus will be associated with improved working-memory performance during an N-back working memory paradigm.

Methods

Participants

Seventy volunteers were recruited via flyers, print advertisements, and Internet listservs. All participants provided informed consent approved by the National Institute of Mental Health (NIMH) Combined Neuroscience Institutional Review Board. Participants were free from the following exclusion criteria: (a) past or current Axis I psychiatric disorder as assessed through a clinician administered SCID-I/NP (First et al. 2002), (b) current substance abuse or past substance dependence as assessed through a clinician administered SCID-I/NP, (c) current of past diagnosis of Attention Deficit Hyperactivity Disorder as assessed by a clinician (see Supplementary Materials for details), (d) first-degree relative with a psychotic disorder, (e) medical condition conflicting with safety or design of the study, (f) brain abnormality on MRI as assessed by a radiologist, (g) positive toxicology screen, or (h) MRI contraindication. Ten participants were excluded from resting state analyses: 2 participants had missing resting state sequences because of time constraints; 2 participants were excluded for head motion (see “BOLD fMRI preprocessing and analysis”); 3 participants were excluded because of scanner artifact caused by hair extensions, a clinical finding, and transcranial magnetic stimulation earlier that day from a different research study; and 3 participants were excluded due to non-convergence of multi-echo independent component analysis during EPI preprocessing. The final sample for fMRI analysis consisted of 60 participants (M = 27.4 ± 6.44 SD years; 32 women). Demographic information for each drug group is presented in Table 1. History of substance use, alcohol use, and attention-deficit/hyperactivity disorder symptoms are presented in Supplemental Table 1. Finally, brain-behavior relationships were assessed for 53 participants who had overlapping N-back working-memory accuracy values (placebo: N = 27, M = 26.4 ± 6.47 SD years, 14 women; methylphenidate: N = 26; M = 28.0 ± 6.84 SD years; 15 women).

Table 1.

Sample sizes and demographic information for participants in fMRI analyses as a function of drug group

Placebo Methylphenidate
Number of participants N = 31 N = 29
Gender 16 female/15 male 16 female/13 male
Age 27.03 ± (6.35) 27.86 ± (6.63)
Weight (kg) 75.53 ± (15.88) 75.74 ± (16.51)
Open in a new tab

Experimental procedure

Participants were asked to abstain from eating 1 h before administration of the drug. Additionally, participants were instructed not to consume nicotine or alcohol during the 24 h before their study visit. Following informed consent, participants were orally administered either placebo or 20 mg of methylphenidate in a between-subject manner. Both participants and research personnel were blinded to the drug condition. Following the collection of anatomical sequences, resting-state EPI data was collected ~ 80 min after drug administration. As peak blood concentrations for oral methylphenidate occur 1.5–2 h after ingestion (Swanson and Volkow 2003), participants performed two runs of an experimental N-back working-memory paradigm beginning ~ 90 min after drug administration (for details, see “N-back working-memory paradigm”). The influence of methylphenidate on BOLD responses during the N-back working-memory paradigm is outside the scope of this paper and will be reported elsewhere. However, here, working-memory accuracy is utilized as a dependent variable to assess the impact of both methylphenidate and thalamic iFC with the lateral prefrontal cortex on cognitive performance.

Resting state scan

Participants were instructed to remain still and keep their eyes open while looking at a central fixation cross on a blank screen. The resting state scan lasted for 8 min.

N-back working-memory paradigm

An experimental N-back paradigm was collected immediately following the resting state scan. This working memory study was designed to determine whether acute methylphenidate reduced the deleterious effect of threat of electrical shock on working-memory capacity. Therefore, the WM task was performed in two conditions, a safe, and a threat (threat of unpredictable shocks) condition. Here, accuracy during the N-back paradigm is used as a between-subject measure of cognitive performance. Participants performed a task wherein they viewed a series of consecutively presented letters. Participants were instructed to press a button to indicate whether the current letter matched or did not match a previous target letter. Trials were presented in blocks of 18, and each trial consisted of letters presented for 0.5 s followed by a fixed 2.0 s inter-trial interval. Working-memory load was manipulated by designating the target as the letter that immediately preceded the current letter (1-back) or by designating the target as the letter that occurred three trials beforehand (3-back). Accuracy during the 3-back condition is only meaningful after 3 trials have already occurred, and consequently, accuracy was measured for the last 15 of 18 trials across 1-back and 3-back conditions. Additionally, measures of accuracy did not include trials in which participants received an aversive electrical shock. Reaction times were measured on trials during which participants responded accurately, not including trials in which participants received an electrical shock or trials in which reaction times which were 3 standard deviations above or below the mean reaction time. Thirty-three percent of trials matched the target in both the 1-back and 3-back conditions. Participants performed 2 runs of the experimental paradigm, and each run contained 4 blocks of alternating “safe” and “threat” conditions. During safe blocks, participants were told that they would not receive any aversive electrical shocks. During threat blocks, participants were told that they might receive an aversive electrical shock at any time (for details, see Supplemental Materials). A total of six electrical stimuli were delivered during the experiment (three per run). Block order was pseudo-randomized so that participants were never asked to sequentially complete the same level of N-back more than twice. Of the 60 participants included in our resting state analyses, only 55 participants successfully completed the experimental N-back task following the resting state scan. Additionally, 2 participants were excluded for accuracy scores during the 3-back/safe condition three standard deviations below the mean so not to bias our stepwise regression analysis. Consequently, 53 participants had overlapping resting state and N-back task data for the assessment of brain-behavior relationships.

BOLD fMRI data acquisition

MRI data were acquired using a 3T Siemens MAGNETOM Skyra (Erlangen, Germany) fMRI system and a 32-channel head coil. Blood oxygen level-dependent (BOLD) signal was acquired using T2*-weighted multi-echo EPI across 32 ascending interleaved axial slices using the following parameters: TR = 2000 ms, TEs = 12 ms, 24.48 ms, 36.96 ms, flip angle = 70°, acquisition matrix = 64 × 64, 3 × 3 × 3 mm voxels, 240 volumes. Slices were collected with an anterior-to-posterior phase encoding direction. A T1-weighted multi-echo MPRAGE was also collected using the following parameters: TR = 2530 ms, TEs = 1.69 ms, 3.55 ms, 5.41 ms, 7.27 ms, flip angle = 7°, acquisition matrix = 256 × 256 mm, 1 × 1 × 1 mm voxels.

BOLD fMRI preprocessing and analysis

Most preprocessing and analyses were performed using AFNI (Cox 1996). DICOM images from the T1-weighted multi-echo MPRAGE were imported using the DIMON command, and the resulting images corresponding to each echo time were averaged using 3dcalc. FreeSurfer (Fischl et al. 2002) was used to generate region specific segmentation volume images (i.e., aparc + aseg) which were subsequently binarized and multiplied with the averaged T1 image from FreeSurfer to skull strip the image. The first four functional volumes were discarded to allow for steady state equilibrium. The EPI preprocessing pipeline was generated using the afni_proc.py script. Functional volumes were slice time corrected, re-aligned to the minimum outlier volume, co-registered to the corresponding T1 weighted anatomical image, and non-linearly warped to the “MNIa_caez_colin27_T1_18” template with 3dQwarp. The TEDANA python script, part of the Multi Echo Independent Component Analysis package, was also used to denoise the time series by using the T2* decay of BOLD signal across echoes (Kundu et al. 2017). Within the TEDANA script, the kdaw value was set at 5 to facilitate convergence. Functional volumes maintained their 3 mm isotropic voxel resolution and were smoothed using a 6-mm FWHM kernel.

After noise cleaning with the TEDANA method, first-level general linear models were used to statistically remove variance associated with head motion. The six head motion parameters and their first order derivatives were added as regressors into the model, along with individual regressors corresponding to volumes in which 10% of the voxels were outliers. Previous research has demonstrated that even small amounts of head motion can effect measures of iFC, and we statistically censored (i.e., scrubbed) individual volumes where the Euclidean norm motion derivative was greater than 0.2 mm based on the recommendations of Power et al. (2014). Participants whose head motion exceeded 0.2 mm on more than 15% of functional volumes were excluded from analysis.

The 3dNetCorr program (Taylor and Saad 2013) was used to calculate correlations between the time series from each thalamic region of interest. These correlations coefficients were then Fisher transformed to Z statistics for analysis. The average Z statistic from those resulting correlations was calculated for each participant, which served as a measure of intra-thalamic connectivity. Measures of iFC with each thalamic region of interest were calculated via Pearson correlations between the time series from each thalamic region of interest and all voxels. Pearson correlation maps were then Fisher transformed for group-level analysis.

Regions of interest

Regions of interest (ROI) for thalamic sub-regions were created to assess seed-based iFC using the Anatomy toolbox (Behrens et al. 2003). Specifically, all thalamic sub-regions probability maps were thresholded at 50%, resampled to 3 mm isotropic resolution, and binarized. Subsequently, all voxels within any thalamic sub-region ROI that overlapped with voxels from any other thalamic sub-region were removed (see Supplemental Figure 1 and Supplemental Table 2 for more details).

Statistical and analytic strategy

The impact of acute methylphenidate on measures of intra-thalamic connectivity was assessed using between-subject ANCOVA models with drug as a between-subject variable and gender, age, and weight as nuisance covariates.

Correction for multiple comparisons (cluster level alpha = 0.05) across the whole brain gray matter, based on cluster extent, was calculated using a cluster-forming threshold of (P < 0.0005, k = 13, NN = 1) via updated versions of 3dFWHMx and 3dClustSim. Importantly, these updates incorporate a mixed autocorrelation function (acf) to better model non-Gaussian noise structure (Cox et al. 2016; Eklund et al. 2016). Analysis of covariance (ANCOVA) models were used to determine whether methylphenidate affected measures of intra-thalamic connectivity (i.e., the average Z statistic) while controlling for effects of gender, age, and weight. The AFNI program 3dMVM (Chen et al. 2014) was used to model interactions between drug (2 levels: placebo vs. methylphenidate) and seed region (7 levels: motor vs. parietal vs. premotor vs. prefrontal vs. somatosensory vs. temporal vs. visual thalamus ROIs) while controlling for effects of gender, age, and weight. Parameter estimates of iFC were extracted from resulting clusters in the prefrontal cortex, and the identified interactions were probed using ANCOVA models with drug as a between-subject variable, and gender, age, and weight as nuisance covariates. The inclusion of these nuisance covariates did not impact the pattern of results (see Supplemental Materials for details). Post-hoc ANCOVA analyses were performed using SYSTAT v13.1 using an alpha of P < 0.007 to control for multiple comparisons using Bonferroni correction. Bar graphs and strip charts were created using R Studio Version 1.2.1335. Main effects of methylphenidate on the iFC of each thalamic sub-region and exploratory analyses investigating interactions between drug and thalamic sub-region at more liberal statistical thresholds are presented in Supplementary Materials for completeness.

The effect of methylphenidate on accuracy of the N-back working-memory performance as a function of threat condition and cognitive load was assessed using repeated-measure ANCOVA models with drug (2 levels: placebo vs. methylphenidate) as a between-subject factor and cognitive load (2 levels: 1-back vs. 3-back) and threat (2 levels: safe vs. threat) as within-subject factors, while controlling for gender, age, and weight. However, to reduce statistical complexity, brain-behavior relationships were assessed using accuracy during high cognitive load (i.e., 3-back) in the safe condition as a dependent variable reflecting cognitive performance. Additionally, because prior work has suggested that methylphenidate impacts response speed (Linssen et al. 2012), reaction times during high cognitive load (i.e., 3-back) in the safe condition were also included as a dependent variable when assessing brain-behavior relationships. In order to determine which neural variables explain the most variance in cognitive performance during the experimental N-back paradigm, we performed forward stepwise regression analyses, whereby the model begins with no independent variables and predictors are subsequently added if they significantly improve model fit. Measures of intra-thalamic connectivity (i.e., the average Z statistic) and parameter estimates characterizing iFC between each thalamic ROI and the cluster identified in the lateral prefrontal cortex via mass-univariate analysis were entered as independent variables and N-back accuracy during high cognitive load in the safe condition served as the dependent variable. Stepwise regression models controlled for gender, age, and weight.

Results

IFC

Intra-thalamus connectivity

Z statistics corresponding to the association among the time series from each sub-region of the thalamus with each other are reported in Table 2 for participants who received either methylphenidate or placebo. ANCOVA analysis demonstrated that drug (F(1,55) = 9.47, P < 0.005, ηp2 = 0.15) significantly affected measures of intra-thalamic connectivity (i.e., the average Z statistic) while controlling for age, gender, and weight of the participant (Fig. 1). Intra-thalamic connectivity during methylphenidate (M = 0.4480 ± SD = 0.11) was significantly reduced compared to placebo (M = 0.5482 ± SD = 0.13).

Table 2.

Means and standard deviations of the Fisher transformed correlation coefficients (i.e., Z statistics) corresponding to the association between the time series from each thalamic region of interest

Motor Parietal Premotor Prefrontal Temporal Somato Visual
Placebo (N =31)
 Motor -
 Parietal .64 (.14) -
 Premotor .80 (.16) .60 (.16) -
 Prefrontal .55 (.18) .74 (.20) .84 (.18) -
 Temporal .41 (.16) .66 (.22) .57 (.17) 1.02 (.23) -
 Somatosensory .69 (.16) .80 (.20) .53 (.17) .46 (.16) .35 (.15) -
 Visual .23 (.13) .53 (.22) .26 (.15) .33 (.15) .31 (.17) .20 (.14) -
Methylphenidate (N = 29)
 Motor -
 Parietal .50 (.13) -
 Premotor .71 (.10) .46 (.14) -
 Prefrontal .44 (.15) .59 (.18) .70 (.17) -
 Temporal .34 (.15) .58 (.17) .48 (.16) .89 (.17) -
 Somatosensory .58 (.11) .68 (.12) .42 (.12) .36 (.13) .30 (.15) -
 Visual .16 (.12) .46 (.10) .17 (.11) .21 (.11) .22 (.11) .17 (.11) -
Open in a new tab
Fig. 1.

Fig. 1

Open in a new tab

Bar graph and strip chart illustrating average intra-thalamic intrinsic functional connectivity (iFC) as a function of drug condition. Z statistics from associations between each sub-region of the thalamus were averaged for each participant to create measures of intra-thalamic iFC. Error bars = 1SE

IFC of thalamus sub-regions

Repeated measure ANCOVA analysis demonstrated that drug interacted with thalamic sub-region to predict iFC between the thalamus and a cluster in the lateral prefrontal cortex within Brodmann area 6 (x = − 50, y = 7, z = 30, F(6,330) = 6.32, 25 voxels) that extended into rostral areas of Brodmann area 8 (Fig. 2). Post-hoc ANCOVA analyses demonstrated that methylphenidate reduced iFC between the lateral prefrontal cortex and regions of the motor (F(1,55) = 8.95, P = 0.004, ηp2 = 0.14) and somatosensory (F(1,55) = 11.35, P = 0.001, ηp2 = 0.17) sub-regions of the thalamus. There was a trend toward reduced iFC between this cluster and regions of the parietal (F(1,55) = 7.64, P = 0.008, ηp2 = 0.12) and visual (F(1,55) = 5.13, P = 0.027, ηp2 = 0.09) sub-regions of the thalamus, but these effects did not survive correction for multiple comparisons. Methylphenidate did not impact iFC between this cluster and the premotor (F(1,55) = 1.27, P = 0.266, ηp2 = 0.02), prefrontal (F(1,55) = 0.07, P = 0.799, ηp2 = 0.001), or temporal (F(1,55) = 0.40, P = 0.530, ηp2 = 0.007) sub-regions of the thalamus (Fig. 3). The only other cluster exhibiting an interaction between drug and thalamic sub-region which survived correction for multiple comparisons was located within the temporal sub-region of the thalamus (x = 8, y = − 15, z = 12, F(6,330) = 6.46, 14 voxels).

Fig. 2.

Fig. 2

Open in a new tab

Sagittal (left) and axial (right) illustrations of a statistical parametric map representing an interaction between drug and sub-region of the thalamus within left lateralized prefrontal cortex, consistent with Broca’s region

Fig. 3.

Fig. 3

Open in a new tab

Bar graphs and strip chart illustrating significantly reduced intrinsic functional connectivity (iFC) during methylphenidate compared to placebo within motor, parietal, somatosensory, and visual sub-regions, but not premotor, prefrontal, or temporal areas. †P < 0.05, *P < 0.007

Working memory

Impact of methylphenidate on working memory

Repeated measure ANCOVA analysis demonstrated there was a significant effect of working memory load (F(1,48) = 5.70, P = 0.021, ηp2 = 0.11) such that higher working memory was associated with lower accuracy during the N-back paradigm when controlling for gender, age, and weight (Table 3). Main effects of the threat condition, drug, and all interactions failed to reach significance (all Ps > 0.15).

Table 3.

Accuracy (% correct trials), number of correct responses (safe condition: out of 60 trials; threat condition: out of 54 trials), and reaction times (in milliseconds) for N-back performance for participants who were administered placebo (N = 27) and methylphenidate (N = 26). Standard deviations are in parentheses

Condition Placebo Methylphenidate
Safe
1-back		 Safe
3-back		 Threat
1-back		 Threat
3-back		 Safe
1-back		 Safe
3-back		 Threat
1-back		 Threat
3-back
Accuracy 93.3 ± (6.7) 73.5 ± (9.2) 93.7 ± (7.9) 73.4 ± (9.2) 95.1 ± (7.7) 72.6 ± (9.9) 93.2 ± (5.6) 75.4 ± (10.5)
No. of correct trials 55.98 ± (4.0) 44.1 ± (5.5) 50.60 ± (4.3) 39.64 ± (5.0) 57.06 ± (4.6) 43.56 ± (5.9) 50.38 ± (3.0) 40.72 ± (5.7)
Reaction times 741 ± (112) 923 ± (260) 731 ± (116) 930 ± (225) 707 ± (176) 895 ± (285) 725 ± (182) 912 ± (303)
Open in a new tab

Repeated measure ANCOVA analyses predicting reactions times failed to find significant effects of drug, working memory load, threat condition, or any interaction when controlling for gender, age, and weight (all Ps > 0.15). However, significant effects of working memory load were observed when nuisance covariates were removed from the model (See Supplemental Materials for details).

Impact of thalamic iFC on working memory

The selection procedure of our stepwise regression model predicting accuracy during high memory loads in the safe condition during the N-back failed to include measures of intra-thalamic connectivity or iFC between the lateral prefrontal cortex and any sub-region of the thalamus. Additionally, the selection procedure of our stepwise regression model predicting reaction times during high memory loads in the safe condition during the N-back failed to include measures of intra-thalamic connectivity or iFC between the lateral prefrontal cortex and any sub-region of the thalamus.

Discussion

We set out to determine which sub-regions of the thalamus exhibit reduced iFC with the lateral prefrontal cortex during acute methylphenidate and whether reduced iFC between the thalamus and lateral prefrontal cortex is related to working memory performance in healthy human volunteers. Acute methylphenidate significantly reduced iFC between the lateral prefrontal cortex and motor and somatosensory sub-regions of the thalamus, and there was a trend toward reduced iFC with the parietal and visual sub-regions which did not formally survive correction for multiple comparisons. In contrast, methylphenidate did not impact the premotor, prefrontal, and temporal sub-regions of the thalamus. These results may reflect inhibition of non-cognitive areas of the thalamus via indirect (i.e., the thalamic reticular nucleus) connections between the thalamus and prefrontal cortex (Wimmer et al. 2015; Zikopoulos and Barbas 2006). This may reflect the suppression of distracting sources of information. Additionally, we replicated previous work demonstrating that methylphenidate administration reduces intra-regional functional connectivity within the thalamus (Konova et al. 2015). Lastly, although working memory load significantly impacted accuracy during the N-back paradigm, neither measures of intra-thalamic connectivity nor iFC between the thalamus and the lateral prefrontal cortex were associated with individual differences in working-memory performance.

Owing to the thin and distributed morphology of the inhibitory thalamic reticular nucleus (TRN), we were unable to directly measure functional activity within this structure using conventional fMRI. However, previous research has demonstrated that acute methylphenidate impacts iFC of individual thalamic sub-regions (Demiral et al. 2018). Therefore, the present results may reflect the role of the TRN in regulating functional connections between the prefrontal cortex and thalamus. Importantly, both the prefrontal and temporal thalamic sub-regions include areas of the MD nucleus, while both the prefrontal and premotor sub-regions include areas of the VA nucleus (Behrens et al. 2003). This may explain why the premotor, prefrontal, and temporal sub-regions of the thalamus are similarly spared from the action of acute methylphenidate. Afferents from the prefrontal cortex project primarily to the areas of the TRN that innervate the MD and VA nuclei and send less prominent projections to regions of the TRN that innervate the thalamic sensory nuclei (Zikopoulos and Barbas 2006). Prefrontal projections can increase activity of the areas of TRN that inhibit thalamic sensory nuclei during cross-modal divided-attention tasks (Wimmer et al. 2015). Furthermore, prefrontal projections to the TRN have been proposed to regulate attentional resources and suppress distracting information (Zikopoulos and Barbas 2006). The present results may reflect increased TRN inhibition of non-cognitive areas of the thalamus, possibly due to D1 receptor type activation within the TRN (Huang et al. 1992), as a consequence of increased dopaminergic signaling following acute methylphenidate.

Importantly, neurons within the lateral prefrontal cortex also express the dopamine transporter (García-Cabezas et al. 2007; Sánchez-González et al. 2005), and previous research has suggested that the lateral prefrontal cortex is involved with both the maintenance of information within working memory (Sakai et al. 2002) and the suppression of distracting information (Jacob and Nieder 2014). As such, it is unclear from our data whether dopamine signaling within the thalamus or the prefrontal cortex is driving the pattern of results observed here. In addition to working memory processes, sub-regions of the thalamus and lateral prefrontal cortex are also involved in language processing. Previous research has suggested that the MD and VA nuclei play a critical role in the cognitive and motor aspects of speech (Barbas et al. 2013). Additionally, Broca’s area and adjacent cortical regions, such as BA 47 and the ventral part of BA 6, process language related information (Hagoort 2014). Researchers have termed this larger, functionally defined cortical area as Broca’s region (Hagoort 2014), and the lateral prefrontal cluster identified in our analyses is located within an area consistent with this region. The present results suggest that acute methylphenidate suppresses input from non-language related thalamic nuclei to frontal language centers, which may be responsible for maintaining semantic information within working-memory.

Previous research has demonstrated that pharmacological augmentation of dopamine signaling improves cognitive performance (Spencer et al. 2015). Although our results failed to demonstrate an impact of acute methylphenidate on cognitive performance, there was a significant effect of memory load on accuracy scores. This suggests that our N-back paradigm successfully engaged working memory processes. However, between-subject differences in working-memory capacity were not associated with iFC between the thalamus and the lateral prefrontal cortex or measures of intra-thalamic connectivity. The experimental N-back paradigm was designed to determine whether acute methylphenidate reduced the deleterious effect of threat of electrical shock on working-memory performance. As such, between-subject differences in accuracy during this paradigm may not be suitable for the assessment of associations between iFC and behavior. Additionally, previous research has demonstrated that lesions of the MD thalamus have the most impact on working-memory performance during cognitive interference paradigms (Mitchell 2015). Future research should investigate whether iFC between the thalamus and prefrontal cortex is related to performance during cognitive interference paradigms, consistent with the importance of this neural circuit in the suppression of distracting information.

Previous research has demonstrated that acute methylphenidate affects iFC between the thalamus and the neural targets important for sensory and motor processing, including the visual network, sensorimotor network, and regions of the striatum and cerebellum (Demiral et al. 2018; Farr et al. 2014; Konova et al. 2013; Mueller et al. 2014). Although the ventral lateral motor nucleus and the lateral posterior somatosensory nucleus of the thalamus densely express the dopamine transporter (García-Cabezas et al. 2007; Sánchez-González et al. 2005), results failed to show evidence that acute methylphenidate differentially impacts thalamic connectivity with regions of the motor and somatosensory cortices. Future research will be required to address the specificity of the observed patterns and to further delineate how acute dopamine transporter blockade influences functional connections between thalamic sub-regions and the cortical mantle broadly.

This study is not without limitations. First, although it was possible to detect significant differences in the impact of methylphenidate on the iFC of different thalamic sub-regions, it is possible that our voxel size and smoothing process prevented our ability to observe differences between smaller thalamic sub-regions (i.e., the visual, motor, and somatosensory regions). Future research employing ultra-high field MRI and anatomical atlases which delineate specific thalamic nuclei (Iglesias et al. 2018) will be needed to confirm and expand upon the findings reported here. Second, our measures of thalamic-prefrontal iFC were collected at rest independently of our experimental paradigm. Measures of thalamic-prefrontal iFC at rest were not associated with working memory performance. Future studies should employ larger samples sizes and cognitive interference paradigms to determine whether iFC collected during rest is related to individual differences in cognitive capacity. Additionally, future work should assess whether the working-memory processes impact connectivity between the thalamus and prefrontal cortex and whether circuit function during cognitive processes is associated with performance. Third, despite previous research demonstrating that acute methylphenidate impacts cognitive performance (Spencer et al. 2015), acute methylphenidate was not associated with accuracy or reaction time during our experimental N-back paradigm. We selected a dose of 20 mg based on previous finding from our lab (Ernst et al. 2016), and it is possible that a larger dose would result in larger effects on working memory. Lastly, methylphenidate antagonizes both the dopamine and norepinephrine transporters (Arnsten et al. 2007; Ding et al. 1997), and previous research has demonstrated that acute methylphenidate occupies norepinephrine transporter binding sites within the human thalamus, in vivo (Hannestad et al. 2010). Therefore, although an emphasis on dopamine transporter antagonism is a more parsimonious explanation for the pattern of results reported here, we cannot conclude with certainty that changes in iFC during acute methylphenidate are the result of dopaminergic signaling specifically.

Conclusion

Collectively, results suggest that acute methylphenidate decreases iFC between the lateral prefrontal cortex and non-cognitive sub-regions of the thalamus. Specifically, methylphenidate significantly reduced iFC between the lateral prefrontal cortex and motor and somatosensory sub-regions of the thalamus and reduced iFC with the parietal and visual sub-regions at a trend level. In contrast, methylphenidate did not impact the premotor, prefrontal, and temporal sub-regions of the thalamus. These patterns may reflect the role of dopamine signaling within the TRN in weakening functional connectivity between the prefrontal cortex and the thalamus, broadly, while sparing connectivity with sub-regions that support cognitive processes. Connections between the thalamus and prefrontal cortex have been proposed to mediate the suppression of distracting information during multi-tasking. The present findings suggest that acute methylphenidate reduces functional connections between the prefrontal cortex and non-cognitive areas of the thalamus.

Supplementary Material

Suppl

Acknowledgments

This work utilized the computational resources of the NIH HPC Biowulf cluster (http://hpc.nih.gov).

Funding information This work was supported by the Intramural Research Program of the National Institutes of Mental Health, project no. ZIAMH002798 (clinical protocol 02-M-0321, NCT00047853 ) to CG.

Footnotes

Conflict of interest The authors declare that there is no conflict of interest.

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00213–020-05505-z) contains supplementary material, which is available to authorized users.

References

  1. Arnsten AF, Scahill L, Findling RL (2007) Alpha-2 adrenergic receptor agonists for the treatment of attention-deficit/hyperactivity disorder: emerging concepts from new data. J Child Adolesc Psychopharmacol 17(4):393–406 [DOI] [PubMed] [Google Scholar]
  2. Baddeley A (1992) Working memory. Science 255(5044):556–559. 10.1126/science.1736359 [DOI] [PubMed] [Google Scholar]
  3. Barbas H, García-Cabezas MÁ, Zikopoulos B (2013) Frontal-thalamic circuits associated with language. Brain Lang 126(1):49–61 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Behrens TE, Johansen-Berg H, Woolrich MW, Smith SM, Wheeler-Kingshott CAM, Boulby PA, Barker GJ, Sillery EL, Sheehan K, Ciccarelli O (2003) Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nat Neurosci 6(7):750–757 [DOI] [PubMed] [Google Scholar]
  5. Chen G, Adleman NE, Saad ZS, Leibenluft E, Cox RW (2014) Applications of multivariate modeling to neuroimaging group analysis: a comprehensive alternative to univariate general linear model. Neuroimage 99:571–588 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Collins DP, Anastasiades PG, Marlin JJ, Carter AG (2018) Reciprocal circuits linking the prefrontal cortex with dorsal and ventral thalamic nuclei. Neuron 98(2):366–379 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cox RW (1996) AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res 29(3): 162–173 [DOI] [PubMed] [Google Scholar]
  8. Cox RW, Reynolds RC, Taylor PA (2016) AFNI and clustering: false positive rates redux. BioRxiv 065862. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. de Bourbon-Teles J, Bentley P, Koshino S, Shah K, Dutta A, Malhotra P, Egner T, Husain M, Soto D (2014) Thalamic control of human attention driven by memory and learning. Curr Biol 24(9):993–999 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Demiral ŞB, Tomasi D, Wiers CE, Manza P, Shokri-Kojori E, Studentsova Y, Wang G-J, Volkow ND (2018) Methylphenidate’s effects on thalamic metabolism and functional connectivity in cannabis abusers and healthy controls. Neuropsychopharmacology 1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ding Y-S, Fowler JS, Volkow ND, Dewey SL, Wang G-J, Logan J, Gatley SJ, Pappas N (1997) Chiral drugs: comparison of the pharmacokinetics of [11C] d-threo and L-threo-methylphenidate in the human and baboon brain. Psychopharmacology 131(1):71–78 [DOI] [PubMed] [Google Scholar]
  12. Eklund A, Nichols TE, Knutsson H (2016) Cluster failure: why fMRI inferences for spatial extent have inflated false-positive rates. Proc Natl Acad Sci 113(28):7900–7905 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ernst M, Lago T, Davis A, Grillon C (2016) The effects of methylphenidate and propranolol on the interplay between induced-anxiety and working memory. Psychopharmacology 233(19):3565–3574. 10.1007/s00213-016-4390-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Farr OM, Zhang S, Hu S, Matuskey D, Abdelghany O, Malison RT, Li CR (2014) The effects of methylphenidate on resting-state striatal, thalamic and global functional connectivity in healthy adults. Int J Neuropsychopharmacol 17(8):1177–1191 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. First MB, Spitzer RL, Gibbon M, Williams JB (2002) Structured clinical interview for DSM-IV-TR axis I disorders, research version, patient edn. SCID-I/P, New York [Google Scholar]
  16. Fischl B, Salat DH, Busa E, Albert M, Dieterich M, Haselgrove C, Van Der Kouwe A, Killiany R, Kennedy D, Klaveness S (2002) Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain. Neuron 33(3):341–355 [DOI] [PubMed] [Google Scholar]
  17. García-Cabezas MÁ, Rico B, Sánchez-González MÁ, Cavada C (2007) Distribution of the dopamine innervation in the macaque and human thalamus. Neuroimage 34(3):965–984 [DOI] [PubMed] [Google Scholar]
  18. Hagoort P (2014) Nodes and networks in the neural architecture for language: Broca’s region and beyond. Curr Opin Neurobiol 28:136–141 [DOI] [PubMed] [Google Scholar]
  19. Hannestad J, Gallezot J-D, Planeta-Wilson B, Lin S-F, Williams WA, van Dyck CH, Malison RT, Carson RE, Ding Y-S (2010) Clinically relevant doses of methylphenidate significantly occupy norepinephrine transporters in humans in vivo. Biol Psychiatry 68(9):854–860 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hauser TU, Fiore VG, Moutoussis M, Dolan RJ (2016) Computational psychiatry of ADHD: neural gain impairments across Marrian levels of analysis. Trends Neurosci 39(2):63–73 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Huang Q, Zhou D, Chase K, Gusella JF, Aronin N, DiFiglia M (1992) Immunohistochemical localization of the D1 dopamine receptor in rat brain reveals its axonal transport, pre- and postsynaptic localization, and prevalence in the basal ganglia, limbic system, and thalamic reticular nucleus. Proc Natl Acad Sci 89(24):11988–11992 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Iglesias JE, Insausti R, Lerma-Usabiaga G, Bocchetta M, Van Leemput K, Greve DN, van der Kouwe A, Fischl B, Caballero-Gaudes C, Paz-Alonso PM (2018) A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology. NeuroImage 183: 314–326. 10.1016/j.neuroimage.2018.08.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jacob SN, Nieder A (2014) Complementary roles for primate frontal and parietal cortex in guarding working memory from distractor stimuli. Neuron 83(1):226–237. 10.1016/j.neuron.2014.05.009 [DOI] [PubMed] [Google Scholar]
  24. Koelsch S, Schulze K, Sammler D, Fritz T, Müller K, Gruber O (2009) Functional architecture of verbal and tonal working memory: an FMRI study. Hum Brain Mapp 30(3):859–873 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Konova AB, Moeller SJ, Tomasi D, Volkow ND, Goldstein RZ (2013) Effects of methylphenidate on resting-state functional connectivity of the mesocorticolimbic dopamine pathways in cocaine addiction. JAMA Psychiatry 70(8):857–868 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Konova AB, Moeller SJ, Tomasi D, Goldstein RZ (2015) Effects of chronic and acute stimulants on brain functional connectivity hubs. Brain Res 1628:147–156 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kundu P, Voon V, Balchandani P, Lombardo MV, Poser BA, Bandettini PA (2017) Multi-echo fMRI: a review of applications in fMRI denoising and analysis of BOLD signals. Neuroimage 154:59–80 [DOI] [PubMed] [Google Scholar]
  28. Kupferschmidt DA, Gordon JA (2018) The dynamics of disordered dialogue: prefrontal, hippocampal and thalamic miscommunication underlying working memory deficits in schizophrenia. Brain Neurosci Adv 2:2398212818771821 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Linssen AMW, Vuurman E, Sambeth A, Riedel WJ (2012) Methylphenidate produces selective enhancement of declarative memory consolidation in healthy volunteers. Psychopharmacology 221(4):611–619 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mitchell AS (2015) The mediodorsal thalamus as a higher order thalamic relay nucleus important for learning and decision-making. Neurosci Biobehav Rev 54:76–88. 10.1016/j.neubiorev.2015.03.001 [DOI] [PubMed] [Google Scholar]
  31. Mueller S, Costa A, Keeser D, Pogarell O, Berman A, Coates U, Reiser MF, Riedel M, Möller H-J, Ettinger U (2014) The effects of methylphenidate on whole brain intrinsic functional connectivity. Hum Brain Mapp 35(11):5379–5388 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Owen AM, Stern CE, Look RB, Tracey I, Rosen BR, Petrides M (1998) Functional organization of spatial and nonspatial working memory processing within the human lateral frontal cortex. Proc Natl Acad Sci 95(13):7721–7726. 10.1073/pnas.95.13.7721 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Parnaudeau S, O’Neill P-K, Bolkan SS, Ward RD, Abbas AI, Roth BL, Balsam PD, Gordon JA, Kellendonk C (2013) Inhibition of mediodorsal thalamus disrupts thalamofrontal connectivity and cognition. Neuron 77(6):1151–1162. 10.1016/j.neuron.2013.01.038 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Pergola G, Danet L, Pitel A-L, Carlesimo GA, Segobin S, Pariente J, Suchan B, Mitchell AS, Barbeau EJ (2018) The regulatory role of the human mediodorsal thalamus. Trends Cogn Sci [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE (2014) Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage 84:320–341 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ramaekers JG, Evers EA, Theunissen EL, Kuypers KPC, Goulas A, Stiers P (2013) Methylphenidate reduces functional connectivity of nucleus accumbens in brain reward circuit. Psychopharmacology 229(2):219–226 [DOI] [PubMed] [Google Scholar]
  37. Ranganath A, Jacob SN (2016) Doping the mind: dopaminergic modulation of prefrontal cortical cognition. Neuroscientist 22(6):593–603. 10.1177/1073858415602850 [DOI] [PubMed] [Google Scholar]
  38. Sakai K, Rowe JB, Passingham RE (2002) Active maintenance in prefrontal area 46 creates distractor-resistant memory. Nat Neurosci 5(5):479–484 [DOI] [PubMed] [Google Scholar]
  39. Sánchez-González MÁ, García-Cabezas MÁ, Rico B, Cavada C (2005) The primate thalamus is a key target for brain dopamine. J Neurosci 25(26):6076–6083 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schmitt LI, Wimmer RD, Nakajima M, Happ M, Mofakham S, Halassa MM (2017) Thalamic amplification of cortical connectivity sustains attentional control. Nature 545(7653):219–223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sherman SM (2011) Functioning of circuits connecting thalamus and cortex. Compr Physiol 7(2):713–739 [DOI] [PubMed] [Google Scholar]
  42. Spencer RC, Devilbiss DM, Berridge CW (2015) The cognition-enhancing effects of psychostimulants involve direct action in the prefrontal cortex. Biol Psychiatry 77(11):940–950 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Swanson JM, Volkow ND (2003) Serum and brain concentrations of methylphenidate: implications for use and abuse. Neurosci Biobehav Rev 27(7):615–621 [DOI] [PubMed] [Google Scholar]
  44. Tanaka M (2007) Cognitive signals in the primate motor thalamus predict saccade timing. J Neurosci 27(44):12109–12118. 10.1523/JNEUROSCI.1873-07.2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Taylor PA, Saad ZS (2013) FATCAT: (an efficient) functional and tractographic connectivity analysis toolbox. Brain Connect 3(5): 523–535 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Tomasi D, Volkow ND (2012) Abnormal functional connectivity in children with attention-deficit/hyperactivity disorder. Biol Psychiatry 71(5):443–450 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Volkow ND, Wang G-J, Fowler JS, Gatley SJ, Logan J, Ding Y-S, Hitzemann R, Pappas N (1998) Dopamine transporter occupancies in the human brain induced by therapeutic doses of oral methylphenidate. Am J Psychiatr 155(10):1325–1331 [DOI] [PubMed] [Google Scholar]
  48. Wimmer RD, Schmitt LI, Davidson TJ, Nakajima M, Deisseroth K, Halassa MM (2015) Thalamic control of sensory selection in divided attention. Nature 526(7575):705–709 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Xiaob D, Barbas H (2004) Circuits through prefrontal cortex, basal ganglia, and ventral anterior nucleus map pathways beyond motor control. Thalamus Relat Syst 2(4):325–343 [Google Scholar]
  50. Zikopoulos B, Barbas H (2006) Prefrontal projections to the thalamic reticular nucleus form a unique circuit for attentional mechanisms. J Neurosci 26(28):7348–7361 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Suppl
NIHMS1617803-supplement-Suppl.docx (989.5KB, docx)

RESOURCES