Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2017 Feb 16;38(5):2580–2588. doi: 10.1002/hbm.23542

Altered neural mechanisms of cognitive control in patients with primary progressive multiple sclerosis: An effective connectivity study

Ekaterina Dobryakova 1,2,✉, Maria Assunta Rocca 1,3, Paola Valsasina 1, John DeLuca 2,4, Massimo Filippi 1,3
PMCID: PMC6866882  PMID: 28205364

Abstract

Primary progressive multiple sclerosis (PPMS) leads to physical and cognitive disability. Specifically, cognitive deficits in PPMS have been explained by both grey matter atrophy and white matter lesions. However, existing research still lacks in the understanding of how the brain of a patient with PPMS functions under cognitive control demands. Thus, the aim of the current study was to examine information integration in patients with PPMS using a search‐based effective connectivity method. Fourteen patients with PPMS and 22 age‐ and gender‐matched healthy controls (HC) performed the Stroop task, a cognitively demanding interference task that taxes neural resources required for cognitive control and response inhibition. Results showed that compared to HC, PPMS patients exhibited poor behavioral performance and alterations in information flow, manifested in the form of the loss of top‐down connections, reversal of connections, and hyperconnectivity. Significant correlations were observed between connection strengths and behavioral measures. The connection between the posterior parietal cortex (PCC) and left posterior parietal lobule, which was present in both groups, showed a negative correlation with performance accuracy on incongruent trials. The connection between the left dorsolateral prefrontal cortex and PCC showed a positive correlation with performance accuracy on incongruent trials. However, the adaptive nature of this connection was not significant on a behavioral level as the PPMS group performed significantly worse compared to the HC group during the Stroop task. Thus, the current study provides important evidence about effective connectivity patterns that can be characterized as maladaptive cerebral re‐organization in the PPMS brain. Hum Brain Mapp 38:2580–2588, 2017. © 2017 Wiley Periodicals, Inc.

Keywords: effective connectivity, multiple sclerosis, primary progressive, Stroop, executive control, interference, fMRI, Bayes networks

INTRODUCTION

Multiple sclerosis (MS) is a neurological disorder that results in axonal loss due to demyelination in the white matter (WM) [Miller and Leary, 2007]. The damage caused by disease processes leads not only to physical, but also to cognitive disability. As is the case with other MS phenotypes, patients with primary progressive (PP) MS disease course have been shown to have significant cognitive problems in several domains [Miller and Leary, 2007]. Cognitive deficits in primary progressive multiple sclerosis (PPMS) have been partially explained by grey matter atrophy and focal WM lesions [Camp et al., 1999; Riccitelli et al., 2011; Tur et al., 2011]. In addition, previous functional neuroimaging studies in PPMS have shown differences in neural network functioning in association with motor [Ceccarelli et al., 2010; Ciccarelli et al., 2006; Filippi et al., 2002] and cognitive tasks [Rocca et al., 2010; Smith et al., 2002]. Notably, Filippi et al. [2002] and Ceccarelli et al. [2010] reported altered recruitment of the sensorimotor network in patients with PPMS during simple motor task performance compared to healthy individuals. Working memory and processing speed have also been shown to lead to increased activation of the fronto‐parietal network in individuals with PPMS [Rocca et al., 2010].

Further exploration of neural correlates of cognitive deficits in MS has focused on examining functional connectivity between networks at rest (i.e., resting‐state functional connectivity) [Dogonowski et al., 2013a, 2013b; Lowe et al., 2008; Rocca et al., 2010, 2012]. However, a potentially, more direct representation of brain function and information integration between regions can be provided by examining effective connectivity during a specific task. Effective connectivity allows one to examine the influence of one brain region on another in the network that is engaged in a specific task performance [Friston, 1994]. We recently showed that the effective connectivity patterns during a cognitive control task differ between MS phenotypes [Dobryakova et al., 2016]. However, our previous study was limited to patients with relapse‐onset disease, including relapsing‐remitting, secondary progressive and benign MS. We expect that exploring differences in information integration through effective connectivity in PPMS subjects will provide important pieces of information about reorganization of cognitive control function, given the unique clinical profile of PPMS [Miller and Leary, 2007].

Thus, in the current study, we used a recently developed effective connectivity approach, based on Bayes causal networks [Ramsey et al., 2010, 2011], to examine information integration during the Stroop task in PPMS. The effective connectivity analysis allowed us to investigate the causal relationship between brain regions while subjects performed this cognitively demanding task. It was hypothesized that individuals with PPMS would exhibit abnormal connectivity of prefrontal regions as the Stroop task heavily relies on executive processes.

METHODS

Participants

Fourteen patients with PPMS [Montalban et al., 2009] and 22 age‐ and gender‐matched healthy controls (HCs) consented to participate in the study (Table 1). The inclusion criteria were: (1) right‐handedness; (2) absence of clinical involvement of the right upper limb (for patients); (3) native Italian speaking; (4) ability to discriminate colors accurately; (5) normal or corrected‐to‐normal vision; (6) no concomitant therapy with antidepressants and psychoactive drugs; and (7) no history of major medical or psychiatric disorders. The institutional review board approved the study and a written informed consent was obtained from all study participants.

Table 1.

Main demographic, clinical, structural magnetic resonance imaging (MRI) and behavioral characteristics of healthy control (HC) and primary progressive multiple sclerosis (PPMS) groups

HC PPMS P a
Age (SD) [years] 46.3 (11.8) 50.5 (11.3) 0.3
Gender (M/F) 13/9 12/2 0.3+
Median EDSS (range) — 5.5 (4.0–7.0) —
Disease Duration (SD) [years] — 13.1 (8.9) —
T2 LV (SD) [ml] — 12.7 (12.3) —
NBV (SD) [ml] 1564 (97) 1443 (74) 0.001
Correct responses (congruent) [%] 96.42 90.62 0.05
Correct responses (incongruent) [%] 95.34 83.74 0.003
RT (congruent) [s] 0.86 1.02 0.001
RT (incongruent) [s] 0.94 1.15 0.0001
Open in a new tab
a

Two‐sample t test for non‐paired data; +Mann‐Whitney U test.

Abbreviations: SD = standard deviation; EDSS = Expanded Disability Status Scale; T2 LV = T2 lesion volume; NBV = normalized brain volume; RT = reaction time.

MRI Acquisition

A 3.0 Tesla scanner (Philips, Best, the Netherlands) was used to obtain the following sequences: (a) dual echo turbo spin echo (TSE) (repetition time [TR]/echo time [TE] = 2,599/16–80 ms; echo train length = 5; flip angle = 150°, 44 contiguous, 3‐mm thick, axial slices, matrix size = 256 × 256, field of view [FOV] = 240 mm2; acquisition time = 7 min 21 s); (b) single‐shot echo planar imaging (EPI) for functional MRI (fMRI) acquisition (TR/TE = 2,000/30 ms, flip angle = 85°, FOV = 240 mm2, matrix = 128 × 128; slice thickness = 4 mm; 250 sets of 30 contiguous axial slices; total acquisition time= 8 min 20 s); and (c) 3D T1‐weighted fast field echo (FFE) for quantification of brain atrophy (TR = 25 ms, TE = 4.6 ms, flip angle = 30°, FOV = 230 mm2, matrix = 256 × 256, slice thickness = 1 mm, 220 contiguous axial slices, in‐plane resolution = 0.89 × 0.89 mm2; acquisition time = 6 min 33 s).

fMRI Paradigm

A modified version of the canonical Stroop task previously validated in HC and MS patients [Rocca et al., 2009] was used during acquisition of fMRI data. The task was programmed with the Presentation software (http://www.neuro-bs.com, Version 0.70). Briefly, the Stroop task included three different trials: color‐congruent trials (C) (e.g., “red” printed in red), neutral trials (N) (e.g., “dog” printed in red) and incongruent trials (I) (e.g., “red” printed in green), separated by a fixation cross. A detailed description of the task design can be found elsewhere [Rocca et al., 2009]. Behavioral data (i.e., reaction times [RT] and percentage of correct responses) during the performance of the Stroop task were recorded from each study subject.

fMRI Analysis

fMRI data were analyzed using SPM8 software. Prior to statistical analysis, all images were realigned to the mean scan, spatially normalized into the standard MNI space, and smoothed with a 10‐mm, 3D‐Gaussian filter. Subjects were included in the statistical analysis if they had a maximum translation/rotation lower than 3.0 mm in all directions.

Event‐related correct responses for each condition were modeled using the general linear model (GLM) and the theory of Gaussian fields [Friston et al., 1995]. A first‐level design matrix, including motion parameters as regressors, was built and specific effects were tested by applying appropriate linear contrasts on t statistical parametric maps (SPMt). At this stage, hemodynamic changes during the congruent and incongruent conditions, as well as those related to the Stroop interference effect (incongruent vs. neutral), were defined. To determine commonly activated brain regions across the whole sample of subjects, a second level analysis was performed on the interference contrast (one‐way ANOVA model, effects of interest, P < 0.001, family‐wise error corrected for multiple comparisons). Based on the second level results and on previous literature [Dedovic et al., 2009; Ridderinkhof et al., 2004; Roberts and Hall, 2008], 14 regions of interest (ROI) were selected for effective connectivity analysis (listed in Table 2).

Table 2.

List of regions of interest (ROIs) selected as nodes for the effective connectivity analysis

ROIs MNI space coordinate x MNI space coordinate y MNI space coordinate z
R VMPFC 13 55 0
L VMPFC −13 55 0
dACC 2 3 50
R CN 11 4 5
L CN −11 4 5
L INS 48 8 −1
R INS −48 8 −1
PCC 1 −41 23
R PPL 31 −53 61
L PPL −31 −53 61
R IFG 31 25 −24
L IFG −31 25 −24
R DLPFC 26 38 42
L DLPFC −26 38 42
Open in a new tab

The selection was done based on the brain circuit activated during the Stroop interference task in the HC group (full factorial model, P < 0.001, family‐wise error corrected for multiple comparisons).

Abbreviations: L = left; R = right; DLPFC = dorsolateral prefrontal cortex; VMPFC = ventromedial prefrontal cortex; PPL = posterior parietal lobule; dACC = dorsolateral anterior cingulate; CN = caudate nucleus; PCC = posterior cingulate cortex; INS = insula; IFG = inferior frontal gyrus.

Effective Connectivity Analysis

Effective connectivity analysis was performed using the Independent Multi‐sample Greedy Equivalence Search (IMaGES) and the Linear non‐Gaussian Orientation, Fixed Structure (LOFS) with Tetrad software (Version 5.6; http://www.phil.cmu.edu/tetrad/) [Ramsey et al., 2010, 2011]. This software uses Bayes network methods [Spirtes et al., 1993] to perform effective connectivity analysis on fMRI data. The method reveals the direction of temporal influence among the regions within the network as well as the connection strength between regions. As the method is search‐based, there is no need for a pre‐defined model of connections. The search is constrained to pre‐defined regions and penalized for over‐fitting, thus the risk of false‐positive connections is minimized.

Time series from the 14 regions selected‐based activation from previous studies [Leung et al., 2000; Peterson et al., 1999; Wolf et al., 2014] of interest activated by the Stroop interference task (reported in Table 2) were extracted from all study subjects. Then, time series from each study group were given as input to the IMaGES algorithm. This algorithm starts with an empty graph and searches forward, one new connection at a time, until the set of connections that optimally represent the entire group of subjects is determined. This is accomplished by selecting the model with the highest Bayesian Information criterion (BIC) score across all datasets. As we are trying to estimate causal relationships among variables, “triangular” graphs (i.e., graphs where three regions are circularly connected) were avoided by increasing the penalty function in the BIC score [Ramsey et al., 2010, 2011]. After the IMaGES algorithm identified a directed acyclic graph for the set of regions, the LOFS algorithm [Ramsey et al., 2011] was used to determine the dominant direction (i.e., the causality) of the connection between two regions, eliminating bidirectional edges. Finally, after the connections were detected and oriented, a structural equation modeling (SEM) estimator was used to calculate the goodness‐of‐fit of each dataset to the outcome connectivity model of the LOFS algorithm. The parameter values of the SEM model (representing connection strengths for each study subject) were estimated using a regression optimizer and were exported from Tetrad as t‐scores for formal statistical analysis [Dobryakova et al., 2016].

Structural MRI Analysis

T2 lesion volumes (LV) were measured using Jim 6.0 software (http://www.xinapse.com). Normalized brain volume (NBV) was measured using Structural Imaging Evaluation of Normalized Atrophy (SIENAx) software [Smith et al., 2002].

Statistical Analysis

Between‐group comparisons of demographic and structural MRI data were performed using two‐sample t‐tests and the Mann‐Withney U test, as appropriate, while behavioral measures (i.e., RT and accuracy) were compared between groups using repeated‐measures ANOVA models. Effective connectivity coefficients estimated by SEM regression were averaged within each study group and were reported, together with their average standard errors, in Table 3 for descriptive purposes. Two‐sample t‐tests were performed to compare connection strengths (represented as t‐scores) of effective connectivity coefficients between groups. Finally, in PPMS patients, univariate Spearman's rho correlations across groups were performed between t‐scores of connectivity coefficients and behavioral as well as structural MRI measures.

Table 3.

Mean connectivity values (standard error of the mean) of causal connections between brain regions observed in HC and primary progressive multiple sclerosis (PPMS) groups during the performance of the Stroop task

Average effective connectivity coefficients (SE)
Directed connections HC PPMS
dACC => R PPL 0.32 (0.05) —
L CN => dACC 0.34 (0.06) —
L CN => R CN 0.68 (0.04) 0.72 (0.04)
L DLFPC => dACC — 0.47 (0.04)
L DLPFC => PCC — −0.71 (0.05)
R DLPFC => L DLPFC 0.69 (0.04) 0.78 (0.03)*
L DLPFC => R IFG — 0.54 (0.03)
L DLPFC => R INS — 0.30 (0.04)
L DLPFC =>L VMPFC — 0.34 (0.03)
R PPL => L PPL 0.82 (0.06) 0.82 (0.04)*
L VMFPC => PCC −0.47 (0.07) —
L VMFPC => R IFG 0.28 (0.04) —
L VMFPC => R VMPFC 0.66 (0.03) 0.47 (0.05)*
L VMPFC => R DLPFC 0.65 (0.04) —
L VMPFC => R CN — 0.09 (0.04)
PCC => L PPL 0.16 (0.03) 0.36 (0.03)
R DLPFC => dACC 0.10 (0.05) —
R DLPFC => L INS 0.13 (0.05) —
R IFG => L IFG 0.64 (0.04) 0.65 (0.04)
R INS => L INS — 0.38 (0.05)
R VMPFC => R INS 0.29 (0.06) —
Open in a new tab

Connections marked with * are reversed in PPMS patients compared with HC.

Abbreviations: L = left; R = right; SE = standard error; DLPFC= dorsolateral prefrontal cortex; VMPFC = ventromedial prefrontal cortex; dACC = dorsolateral anterior cingulate; CN = caudate nucleus; PCC = posterior cingulate cortex; INS = insula; PPL = posterior parietal lobule; IFG = inferior frontal gyrus.

RESULTS

Behavioral Results

Accuracy

Behavioral results are presented in Table 1 and Figure 1A. A group (HC vs. PPMS) × trial type (congruent vs. incongruent trials) interaction was observed in accuracy for congruent and incongruent trial (interaction: F(1, 34) = 5.22, P = 0.03, η = 0.13). An independent sample t‐test showed that the interaction was driven by the PPMS group performing significantly worse on the incongruent trials (t(34) = 3.48, P < 0.001) than on the congruent (t(34) = 4.64, P < 0.05) trials, relative to the HC group. In addition, overall accuracy was better in the HC group (main effect of accuracy: F(1,34) = 9.82, P = 0.004, η = 0.2).

Figure 1.

Figure 1

Open in a new tab

A. Accuracy for the congruent and incongruent trials during the Stroop task. Error bars correspond to the standard error of the mean. B. RT for the congruent and incongruent trials during the Stroop task. Error bars correspond to the standard error of the mean. [Color figure can be viewed at http://wileyonlinelibrary.com]

Reaction Times

As expected, both groups showed an interference effect. That is, both groups responded faster to correct congruent trials compared to correct incongruent trials (trial type main effect: (F(1, 34)=44.56, P <0.001, η = 0.57) (Fig. 1B). The group x trial type interaction was not significant.

Structural MRI Results

Compared to HC, PPMS patients showed significant brain atrophy (Table 1).

Effective Connectivity Analysis

HC group

As can be seen in Table 3 and Figure 2A, the connectivity pattern in HC was characterized by reciprocal connections between the left and the right hemisphere, especially in the frontal lobe. The HC group had 6 left‐to‐right hemisphere connections (left caudate nucleus (CN) => dorsal anterior cingulate cortex (dACC), left CN => right CN, left ventromedial prefrontal cortex (VMPFC) => posterior cingulate cortex (PCC), left VMPFC => right inferior frontal gyrus (IFG), left VMPFC => right VMPFC, left VMPFC => right dorsolateral prefrontal cortex (DLPFC)), and 7 right‐to‐left hemisphere connections (right DLPFC => left DLPFC, right posterior parietal lobule (PPL) => left PPL, right DLPFC => dACC, right DLPFC => left insula, right IFG => left IFG, right insula => left insula, right VMPFC => right insula). The left VMPFC and the right DLPFC had the majority of outgoing connections. Specifically, the left VMPFC showed 4 outgoing connections, while the right DLPFC showed three outgoing connections, suggesting that these regions play a major role during Stroop task performance in HC. Communication with the parietal regions in HCs was delivered via two primary routes: (1) from the left VMPFC; and (2) from the dACC, sending information to the left PPL through the right PPL. The dACC received top‐down information from the right DLPFC and bottom‐up information from the left CN.

Figure 2.

Figure 2

Open in a new tab

Effective connectivity graphs representing information flow during Stroop task performance for the HC (A) and PPMS (B) groups. VMPFC: ventromedial prefrontal cortex; DLPFC: dorsolateral prefrontal cortex; dACC: dorsal anterior cingulate cortex; IFG: inferior frontal gyrus; INS: insula; CN: caudate nucleus; PCC: posterior cingulate cortex; PPL: posterior parietal lobule. [Color figure can be viewed at http://wileyonlinelibrary.com]

PPMS group

PPMS showed a very different connectivity pattern during Stroop task performance compared to the HC (Figs. 2B and 3; Table 3). In contrast to what has been observed in HC connectivity pattern, where the left VMPFC showed greater number of connections, a shift was observed in PPMS patients, where the left VMPFC connections were lost. Compared to HC, PPMS patients lost eight connections in total with three of those connections emanating from the left VMPFC. In PPMS patients, the greater number of connections was observed from the left DLPFC. In contrast to HC, where the left DLPFC showed no connections emanating from it, the left DLPFC acted as the primary connecting region, exerting influence on the majority of frontal and parietal regions of the network (Fig. 3).

Figure 3.

Figure 3

Open in a new tab

A depiction of extra (red), reversed (yellow), and missing (green) connections in the PPMS participants. [Color figure can be viewed at http://wileyonlinelibrary.com]

There were seven extra connections in PPMS patients, all emanating from the left DLPFC. As can be seen from Figure 3 (green arrows), PPMS patients lost top‐down connections that were present in HC. Furthermore, the left hemisphere appeared to contribute more significantly to Stroop task performance, as compared to HC, as most of the connections emanated from it. PPMS patients had eight left‐to‐right hemisphere connections and three right‐to‐left hemisphere connections.

Connectivity strength differences

PPMS patients and HC shared six common connections: (1) left caudate nucleus => right caudate nucleus; (2) right IFG => left IFG; (3) left VMPFC => right VMPFC; (4) PCC => left PPL; (5) right DLPFC => left DLPFC; and (6) right PPL => left PPL. While common in the two groups, three of these connections (i.e., connections 3, 5, 6) were reversed in PPMS patients (Fig. 3). Of the six shared connections, significant between‐group differences in strength were observed between three. The following connections were stronger in PPMS patients: PCC => left PPL (t(15.7)=2.65, P = 0.02, d = 0.98) and right PPL => left PPL (t(14.54)=2.34, P = 0.03, d = 0.88). Conversely, the left VMPFC => right VMPFC connection was stronger in HC (t(34)=2.14, P = 0.04, d = 0.77).

Correlation analysis

Of the three shared connections, which showed group differences in strength (i.e., connections 3, 4, 6), the PCC => lPPL connection showed a negative correlation with the Stroop performance accuracy on incongruent trials (r = −0.35, P < 0.05). Because the lVMPFC => rVMPFC connection was only reversed in the PPMS group, the correlation for each group was performed separately. No significant correlations between the strength of this connection and behavioral measures were observed in any group. In addition, the lDLPFC => PCC connection, an extra connection in the PPMS group, showed a positive correlation with accuracy on the incongruent trials (r = 0.69, P < 0.01).

In PPMS, no correlations were found between effective connectivity alterations and structural MRI measures (T2 LV and brain atrophy).

DISCUSSION

The current study examined the effective connectivity pattern in PPMS subjects during performance of the Stroop task in comparison to HC. Results showed that, compared to the HC group, the PPMS group exhibited: (1) behavioral deficits, and (2) robust alterations of the effective connectivity pattern, which manifested in the form of the loss of top‐down connections, reversal of connections, and increased number of prefrontal connections during Stroop task performance. Moreover, extra connections were associated with behavioral alterations. These three major points are discussed below.

Behavioral Deficits

Performance on the Stroop task was evaluated through measures of performance accuracy, as well as response time for congruent and incongruent trials. The PPMS group exhibited abnormal performance on both measures during the Stroop task, suggesting a deficiency in cognitive control and the use of alternative strategies during task performance [Price and Friston, 1999]. These behavioral results further suggest the presence of abnormalities in neuronal architecture. The observed behavioral pattern is typical of the PPMS participants that have been shown to perform worse than HC on tasks of cognitive control and response inhibition [Ruet et al., 2013].

Connectivity Pattern Alterations

The PPMS group exhibited abnormal effective connectivity compared to the HC group. Specifically, the effective connectivity analysis revealed that PPMS patients had seven extra connections that were not present in the connectivity pattern of HC. Out of the seven extra connections, five of them were emanating from the left DLPFC, in contrast to HC where the majority of efferent connections emanated from the left VMPFC. The DLPFC is generally implicated in a variety of executive processes, such as working memory, planning and information integration [Hoshi, 2006; Krawczyk, 2002]. It is clear from the PPMS connectivity pattern that the primary node in PPMS shifted to the left DLPFC, and its hyperconnectivity might reflect an attempt of the PPMS brain to dedicate more resources to task performance. Hyperconnectivity is a common finding in clinical populations and is usually interpreted as either adaptive or maladaptive cerebral reorganization [Chiaravalloti et al., 2015]. Given the present findings, observed prefrontal hyperconnectivity is difficult to interpret as adaptive due to observed deficits in performance described in the previous section, thus more likely reflecting a maladaptive process.

Prefrontal hyperconnectivity in the PPMS group might also be related to the loss of the eight connections that were present in the HC connectivity pattern. As can be seen from Figure 3, the loss of connections made the PPMS connectivity pattern more reliant on the left DLPFC than the HC connectivity pattern, which was dominated by connections from the left VMPFC. Lost connections cannot be related to behavioral performance, as there is no connection strength associated with an absent connection. However, lost connections might potentially explain worse behavioral measures in the PPMS group, when compared to the HC group. Future studies should investigate this possibility, as well as whether lost functional connections are related to structural abnormalities and/or contribute to functional remapping [Tomassini et al., 2012] caused by disease progression.

The PPMS and HC groups shared 6 connections: (1) left caudate nucleus => right caudate nucleus; (2) right IFG => left IFG; (3) left VMPFC => right VMPFC; (4) PCC => left PPL; (5) right DLPFC => left DLPFC; and (6) right PPL => left PPL. While the shared connections represent qualitative similarities between the PPMS and HC groups, quantitative differences in connection strengths were also observed. Connection strength reflects the degree of influence of one region on another. The connections between the PCC and left PPL and between the right PPL and left PPL were stronger in the PPMS group than in the HC group. The greater influence of one region on the other in the networks of the PPMS group might contribute to diminished cognitive control and response inhibition during Stroop task performance observed in the PPMS group. Similar results were reported in a study that examined effective connectivity during Stroop task performance in relapsing remitting, secondary progressive, and benign MS [Dobryakova et al, 2016]. Furthermore, three of these connections (left VMPFC => right VMPFC, right DLPFC => left DLPFC, right PPL => left PPL) were reversed in the PPMs group, indicating a switch in information flow between the regions. This reversal of connections might also be interpreted as maladaptive in light of poor performance of the PPMS group compared to the HC group [Dobryakova et al., 2016], however, more research is needed to better understand the significance and consequences of reversed connections.

Behavior Correlations with Connection Strengths

Behavioral correlations with connection strengths were observed with two connections that the HC and PPMS groups shared. Specifically, the PCC => left PPL connection showed a negative correlation with performance accuracy on incongruent trials. As mentioned above, this connection was stronger in the PPMS group and suggests that stronger influence of the PCC on the left PPL might be associated with worse accuracy during incongruent Stroop trials. Greater influence of one region on the other might contribute to cognitive control and response inhibition deficits observed in the PPMS group, and can be interpreted as maladaptive cerebral re‐organization.

An extra connection in the PPMS group between the left DLPFC and the PCC showed a positive association with performance accuracy on incongruent trials of the Stroop, suggesting that this long‐range top‐down connection is compensatory or adaptive. However, as the PPMS group exhibited worse performance accuracy compared to the HC group, the compensatory influence of the left DLPFC on the PCC did not have a strong enough impact on participants' performance.

Limitations

The small sample size in this study, especially the PPMS group, is an obvious limitation. However, significant effects in behavioral parameters, effective connectivity measures, and brain‐behavior interactions were still identified. These findings, which probably only represent large effect sizes, warrant replication with larger samples.

Another limitation of the current study is the absence of neuropsychological measures to characterize the cognitive status of the PPMS group. Future studies should examine PPMS with and without cognitive impairment to gain a better understanding of how cognitive impairment influences effective connectivity patterns in the brain.

CONCLUSION

The current study examined information flow in PPMS and HC patients utilizing a data‐driven, effective connectivity approach. Behavioral deficits of the PPMS group in Stroop task performance suggested the presence of neural alterations in the PPMS group compared to the HC group. This was confirmed by results of the effective connectivity analysis that revealed hyperconnectivity, connection loss, and reversal of connections in the PPMS connectivity patterns when compared to HC connectivity patterns. These effective connectivity patterns, in conjunction with observed behavioral deficits, can be characterized as maladaptive cerebral re‐organization in the PPMS brain. Furthermore, the current study confirms that there are structural as well as functional changes in brain connectivity in patients with PPMS. More studies are needed to replicate the current findings. In addition, future studies should focus on examination of changes in effective connectivity after cognitive rehabilitation; this approach will provide valuable evidence about the effects of cognitive rehabilitation on changes in information flow.

ACKNOWLEDGMENTS

Massimo Filippi serves on scientific advisory board for Teva Pharmaceutical Industries; has received compensation for consulting services and/or speaking activities from Bayer Schering Pharma, Biogen Idec, Excemed and Teva Pharmaceutical Industries; and receives research support from Bayer Schering Pharma, Biogen Idec, Merck Serono, Teva Pharmaceutical Industries, Italian Ministry of Health, Fondazione Italiana Sclerosi Multipla, Cure PSP, and the Jacques and Gloria Gossweiler Foundation (Switzerland). John DeLuca serves on an advisory board for Biogen Idec, and for Novartis, and receives speaker's fees from EMD Serono.

REFERENCES

  1. Camp SJ, Stevenson VL, Thompson AJ, Miller DH, Borras C, Auriacombe S, Brochet B, Falautano M, Filippi M, Hérissé‐Dulo L, Montalban X, Parrcira E, Polman CH, De Sa J, Langdon DW (1999): Cognitive function in primary progressive and transitional progressive multiple sclerosis: A controlled study with MRI correlates. Brain 122: 1341–1348. [DOI] [PubMed] [Google Scholar]
  2. Ceccarelli A, Rocca MA, Valsasina P, Rodegher M, Falini A, Comi G, Filippi M (2010): Structural and functional magnetic resonance imaging correlates of motor network dysfunction in primary progressive multiple sclerosis. Eur J Neurosci 31:1273–1280. [DOI] [PubMed] [Google Scholar]
  3. Chiaravalloti ND, Genova HM, DeLuca J (2015): Cognitive rehabilitation in multiple sclerosis: The role of plasticity. Front Neurol 6:67. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Ciccarelli O, Toosy AT, Marsden JF, Wheeler‐Kingshott CM, Miller DH, Matthews PM, Thompson AJ (2006): Functional response to active and passive ankle movements with clinical correlations in patients with primary progressive multiple sclerosis. J Neurol 253:882–891. [DOI] [PubMed] [Google Scholar]
  5. Dedovic K, D'Aguiar C, Pruessner JC (2009): What stress does to your brain: A review of neuroimaging studies. Can J Psychiatry 54:6–15. [DOI] [PubMed] [Google Scholar]
  6. Dobryakova E, Rocca MA, Valsasina P, Ghezzi A, Colombo B, Martinelli V, Comi G, DeLuca J, Filippi M (2016): Abnormalities of the executive control network in multiple sclerosis phenotypes: An fMRI effective connectivity study. Hum Brain Mapp 37:2293–2304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dogonowski AM, Siebner HR, Soelberg Sørensen P, Paulson OB, Dyrby TB, Blinkenberg M, Madsen KH (2013a): Resting‐state connectivity of pre‐motor cortex reflects disability in multiple sclerosis. Acta Neurol Scand 128:328–335. [DOI] [PubMed] [Google Scholar]
  8. Dogonowski AM, Siebner HR, Sørensen PS, Wu X, Biswal B, Paulson OB, Dyrby TB, Skimminge A, Blinkenberg M, Madsen KH (2013b): Expanded functional coupling of subcortical nuclei with the motor resting‐state network in multiple sclerosis. Mult Scler 19:559–566. [DOI] [PubMed] [Google Scholar]
  9. Filippi M, Rocca MAA, Falini A, Caputo D, Ghezzi A, Colombo B, Scotti G, Comi G (2002): Correlations between structural CNS damage and functional MRI changes in primary progressive MS. Neuroimage 15:537–546. [DOI] [PubMed] [Google Scholar]
  10. Friston KJ (1994): Functional and effective connectivity in neuroimaging: A synthesis. Hum Brain Mapp 2:56–78. [Google Scholar]
  11. Friston KJ, Holmes AP, Poline JB, Grasby PJ, Williams SC, Frackowiak RS, Turner R (1995): Analysis of fMRI time‐series revisited. Neuroimage 2:45–53. [DOI] [PubMed] [Google Scholar]
  12. Hoshi E (2006): Functional specialization within the dorsolateral prefrontal cortex: A review of anatomical and physiological studies of non‐human primates. Neurosci Res 54:73–84. [DOI] [PubMed] [Google Scholar]
  13. Krawczyk DC (2002): Contributions of the prefrontal cortex to the neural basis of human decision making. Neurosci Biobehav Rev 26:631–664. [DOI] [PubMed] [Google Scholar]
  14. Leung H, Skudlarski P, Gatenby JC, Peterson BS, Gore JC (2000): An event‐related functional MRI study of the Stroop color word interference task. Cereb Cortex 10:552–560. [DOI] [PubMed] [Google Scholar]
  15. Lowe MJ, Beall EB, Sakaie KE, Koenig KA, Stone L, Marrie RA, Phillips MD (2008): Resting state sensorimotor functional connectivity in multiple sclerosis inversely correlates with transcallosal motor pathway transverse diffusivity. Hum Brain Mapp 29:818–827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Miller DH, Leary SM (2007): Primary‐progressive multiple sclerosis. Lancet Neurol 6:903–912. [DOI] [PubMed] [Google Scholar]
  17. Montalban X, Sastre‐Garriga J, Filippi M, Khaleeli Z, Téllez N, Vellinga MM, Tur C, Brochet B, Barkhof F, Rovaris M, Miller DH, Polman CH, Rovira A, Thompson AJ (2009): Primary progressive multiple sclerosis diagnostic criteria: A reappraisal. Mult Scler 15:1459–1465. [DOI] [PubMed] [Google Scholar]
  18. Peterson BS, Skudlarski P, Gatenby JC, Zhang H, Anderson AW, Gore JC (1999): An fMRI study of Stroop word‐color interference: Evidence for cingulate subregions subserving multiple distributed attentional systems. Biol Psychiatry 45:1237–1258. [DOI] [PubMed] [Google Scholar]
  19. Price CJ, Friston KJ (1999): Scanning patients with tasks they can perform. Hum Brain Mapp 8:102–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Ramsey JD, Hanson SJ, Hanson C, Halchenko YO, Poldrack R. a, Glymour C (2010): Six problems for causal inference from fMRI. Neuroimage 49:1545–1558. [DOI] [PubMed] [Google Scholar]
  21. Ramsey JD, Hanson SJ, Glymour C (2011): Multi‐subject search correctly identifies causal connections and most causal directions in the DCM models of the Smith et al. simulation study. Neuroimage 58:838–848. [DOI] [PubMed] [Google Scholar]
  22. Riccitelli G, Rocca MA, Pagani E, Rodegher ME, Rossi P, Falini A, Comi G, Filippi M (2011): Cognitive impairment in multiple sclerosis is associated to different patterns of gray matter atrophy according to clinical phenotype. Hum Brain Mapp 32:1535–1543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Ridderinkhof KR, van den Wildenberg WPM, Segalowitz SJ, Carter CS (2004): Neurocognitive mechanisms of cognitive control: The role of prefrontal cortex in action selection, response inhibition, performance monitoring, and reward‐based learning. Brain Cogn 56:129–140. [DOI] [PubMed] [Google Scholar]
  24. Roberts KL, Hall DA (2008): Examining a supramodal network for conflict processing: A systematic review and novel functional magnetic resonance imaging data for related visual and auditory Stroop tasks. J Cogn Neurosci 20:1063–1078. [DOI] [PubMed] [Google Scholar]
  25. Rocca MA, Valsasina P, Ceccarelli A, Absinta M, Ghezzi A, Riccitelli G, Pagani E, Falini A, Comi G, Scotti G, Filippi M (2009): Structural and functional MRI correlates of Stroop control in benign MS. Hum Brain Mapp 30:276–290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rocca MA, Riccitelli G, Rodegher M, Ceccarelli A, Falini A, Falautano M, Meani A, Comi G, Filippi M (2010): Functional MR imaging correlates of neuropsychological impairment in primary‐progressive multiple sclerosis. AJNR Am J Neuroradiol 31:1240–1246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Rocca MA, Bonnet MC, Meani A, Valsasina P, Colombo B, Comi G, Filippi M (2012): Differential cerebellar functional interactions during an interference task across multiple sclerosis phenotypes. Radiology 265:864–873. [DOI] [PubMed] [Google Scholar]
  28. Ruet A, Deloire M, Charré‐Morin J, Hamel D, Brochet B (2013): Cognitive impairment differs between primary progressive and relapsing‐remitting MS. Neurology 80:1501–1508. [DOI] [PubMed] [Google Scholar]
  29. Smith SM, Zhang Y, Jenkinson M, Chen J, Matthews PM, Federico A, De Stefano N (2002): Accurate, robust, and automated longitudinal and cross‐sectional brain change analysis. Neuroimage 17:479–489. [DOI] [PubMed] [Google Scholar]
  30. Spirtes P, Glymour C, Scheines R (1993): Causation, Prediction, and Search, Lecture Notes in Statistics. New York, NY: Springer. doi: 10.1007/978-1-4612-2748-9 [DOI] [Google Scholar]
  31. Tomassini V, Matthews PM, Thompson AJ, Fuglø D, Geurts JJ, Johansen‐Berg H, Jones DK, Rocca MA, Wise RG, Barkhof F, Palace J (2012): Neuroplasticity and functional recovery in multiple sclerosis. Nat Rev Neurol 8:635–646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Tur C, Penny S, Khaleeli Z, Altmann DR, Cipolotti L, Ron M, Thompson AJ, Ciccarelli O (2011): Grey matter damage and overall cognitive impairment in primary progressive multiple sclerosis. Mult Scler 17:1324–1332. [DOI] [PubMed] [Google Scholar]
  33. Wolf D, Zschutschke L, Scheurich A, Schmitz F, Lieb K, Tüscher O, Fellgiebel A (2014): Age‐related increases in Stroop interference: Delineation of general slowing based on behavioral and white matter analyses. Hum Brain Mapp 35:2448–2458. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES