bs_bs_banner

Addiction Biology

ORIGINAL ARTICLE

doi:10.1111/adb.12199

Brain function during cognitive flexibility and white matter integrity in alcohol-dependent patients, problematic drinkers and healthy controls Jochem M. Jansen1*, Ruth J. van Holst1,2,3*, Wim van den Brink1, Dick J. Veltman1,4, Matthan W. A. Caan5 & Anna E. Goudriaan1,6 Academic Medical Centre, Department of Psychiatry, Amsterdam Institute for Addiction Research, The Netherlands1, Donders Institute for Cognition, Brain and Behaviour, Radboud University, The Netherlands2, Department of Neurology, Radboud University Medical Centre, The Netherlands3, VU University Medical Center, The Netherlands4, Academic Medical Centre, Department of Radiology, University of Amsterdam, The Netherlands5 and Arkin Mental Health, The Netherlands6

ABSTRACT Cognitive flexibility has been associated with prefrontal white matter (WM) integrity in healthy controls (HCs), showing that lower WM integrity is associated with worse performance. Although both cognitive flexibility and WM integrity have been found to be aberrant in alcohol-dependent (AD) patients, the relationship between the two has never been tested. In this study, we investigated the association between WM tract density and cognitive flexibility in patients with AD (n = 26) and HCs (n = 22). In order to assess the influence of AD severity, we also included a group of problematic drinkers (PrDs; n = 23) who did not meet the AD criteria. Behavioral responses and brain activity during a cognitive flexibility task were measured during functional magnetic resonance imaging. Probabilistic fiber tracking was performed between the dorsolateral prefrontal cortex and the basal ganglia; two crucial regions for task switching. Finally, the task-related functional connectivity between these areas was assessed. There were no significant group differences in the task performance. However, compared with HCs, AD patients and PrDs showed decreased WM integrity and increased prefrontal brain activation during task switching. Evidence is presented for a compensatory mechanism, involving recruitment of additional prefrontal resources in order to compensate for WM and neural function impairments in AD patients and PrDs. Although present in both alcohol groups, the PrDs were more successful in invoking this compensatory mechanism when compared to the AD patients. We propose that this may therefore serve as a protective factor, precluding transition from problematic drinking into alcohol dependence. Keywords:

Alcohol dependence, cognitive flexibility, fiber tracking, fMRI.

Correspondence to: Jochem M. Jansen, Academic Medical Centre, Department of Psychiatry, University of Amsterdam, Room PA3-220, 1100 DD Amsterdam, The Netherlands. E-mail: [email protected] or [email protected]

INTRODUCTION Alcohol and drug dependence are chronic relapsing brain disorders characterized by altered reward circuitry and weakened cognitive control systems that often lead to impairments in cognitive functioning, including impaired cognitive flexibility (Baler & Volkow 2006; Koob & Volkow 2010). Cognitive flexibility is a prerequisite for goal-directed behavior and involves executive decision making preceding switching behavior (Lezak 2004). It has been defined as the ability to adjust one’s

thinking from old to new situations (Moore & Malinowski 2009). Previous studies have shown impaired cognitive flexibility (Vanes et al. 2014) in almost all substance use disorders (for a review, see Volkow et al. 2002) and in problematic gamblers (for a review, see van Holst et al. 2010). These impairments in cognitive flexibility are related to the compulsive use of drugs and/or alcohol, which is one of the key elements in the development and persistence of substance dependence (Everitt & Robbins 2005; Stalnaker et al. 2009).

*These authors contributed equally to this paper and should be referred to as joint first authors. © 2014 Society for the Study of Addiction

Addiction Biology

2

Jochem M. Jansen et al.

Lesion and neuro-imaging studies have shown that the dorsolateral prefrontal cortex (dlPFC) is essential for cognitive flexibility (Badre & Wagner 2006; Nyhus & Barcelo 2009). Neuro-imaging studies on cognitive flexibility in healthy subjects using cognitive switching paradigms observed increased activation in dlPFC, ventrolateral prefrontal cortex (vlPFC), including the inferior frontal gyrus (IFG), anterior cingulate cortex (ACC), posterior parietal cortex and basal ganglia during correct task switching (Sohn et al. 2000; Smith et al. 2004; Ravizza & Carter 2008). In addition, many studies have reported on impaired prefrontal functioning during cognitive tasks in substance-dependent patients (Baler & Volkow 2006; Goldstein & Volkow 2011). However, neural processes underlying aberrant cognitive switching in alcohol use disorders are still not well examined. Previous studies have also shown that prefrontal white matter (WM) integrity is related to cognitive functioning and cognitive flexibility (Bennett et al. 2011; van Schouwenburg et al. 2014). Alcohol is neurotoxic and inflicts damage to the brain, and WM in the prefrontal cortex in particular (Kril et al. 1997; Harper 2007; Yeh et al. 2009). A decline in WM integrity in alcoholdependent (AD) patients has previously been linked to cognitive deficits, including (working) memory, attention and responsiveness to reward (Harris et al. 2008; Chanraud et al. 2009; Pfefferbaum et al. 2010; Schulte et al. 2010, 2012). However, there are no studies that have investigated the relationship between prefrontal WM integrity and cognitive flexibility in a sample of AD subjects and problematic drinkers (PrDs). As previous studies strongly associate the dlPFC and basal ganglia with cognitive flexibility (van Schouwenburg et al. 2012), it is hypothesized that impaired switching behavior in substance dependence is related to damaged fiber bundles connecting the prefrontal cortex and basal ganglia, including ventral striatum and putamen (van Schouwenburg et al. 2014). WM integrity can be investigated non-invasively using diffusion tensor imaging (DTI), which provides fractional anisotropy (FA) and mean diffusivity (MD) values to describe the macroscopic characteristics of fiber bundles (Hagmann et al. 2006). Decreases in FA are thought to reflect a decline in fiber density and coherence, whereas increases in MD values mirror cell membrane damage (Beaulieu 2002). In AD patients, decreases in FA and increases in MD values have been reported (Pfefferbaum & Sullivan 2004; Harper 2007, 2009; Pfefferbaum et al. 2009). Similar to findings in substance dependence, there is a decrease in cognitive functioning and WM integrity in healthy aging of the brain (for a review, see Goh & Park 2009; Park & Reuter-Lorenz 2009). Neuro-imaging studies have reliably reported increased prefrontal activa© 2014 Society for the Study of Addiction

tion during cognitive tasks in elderly subjects (Park & Reuter-Lorenz 2009; Cooper, Worthy & Maddox 2013). High-cognitive functioning elderly exhibit higher activations when compared to low-cognitive functioning elderly (Park & Reuter-Lorenz 2009). The scaffolding theory of aging and cognition (STAC) states that increased activation in cognitively demanding tasks in healthy aging is a compensatory mechanism for decreased WM integrity and neural functioning (Park & Reuter-Lorenz 2009). Increased prefrontal activations are sometimes accompanied by decreased activity in more posterior parts of the brain (Davis et al. 2008). Higher activity during cognitive and emotional tasks has previously been reported and explained as a compensatory mechanism in substancedependent populations (Yucel & Lubman 2007; Goldstein & Volkow 2011; Charlet et al. 2014). For example, higher activations during high-working memory load and affective face recognition were associated with lower relapse risk (Charlet et al. 2014). These higher activations are likely to be associated with alcohol-induced WM damage and support the explanation that higher activations in cognitively demanding situations reflect a compensatory mechanism in patients with AD. The aim of this study is to investigate cognitive flexibility and underlying neuronal mechanisms in a sample of AD patients, PrDs and healthy controls (HCs). The inclusion of PrDs enables the assessment of WM integrity and cognitive flexibility across a broad range of alcoholrelated problems. On a behavioral level, we hypothesize that AD patients and PrDs perform worse compared to HC, as indicated by fewer correct switch trials and higher switch costs. In line with the STAC, we hypothesize that both the AD and the PrD groups show a decline in WM integrity together with an increase in prefrontal neural activity during cognitive switching when compared to HCs. In analogy with high and low functioning elderly, it is expected that PrDs show higher compensatory activation of the prefrontal cortex than AD patients. To test our hypothesis, we applied a cognitive switching task, while measuring the BOLD response in the brain, and performed diffusion tensor fiber tracking between the dlPFC and the basal ganglia.

MATERIALS AND METHODS Participants AD subjects were recruited from Dutch addiction treatment centers where they received cognitive behavioral therapy. HCs and PrDs were recruited through advertisements in local newspapers. All participants were male; HCs did not drink more than 21 standard units (10 g) of alcohol per week. The Ethical Review Board of the Academic Medical Centre approved the study and all subjects Addiction Biology

Brain function and WM in AUD

provided written informed consent and were remunerated for their participation. The number of subjects varied per analysis due to missing data in either one or more modalities; the exact numbers per analysis can be found in Table 2 and are reported throughout the Results section. AD subjects were diagnosed according to DSM-IV-TR alcohol dependence criteria with Section J of the Dutch version of the Clinical International Diagnostic Inventory (World Health Organization 1997). A measure of alcohol problem severity was obtained for all participants by administering the alcohol use disorder identification test (AUDIT; Bush et al. 1998). Furthermore, to ensure that all participants were detoxified from alcohol, AD patients had to be abstinent for at least 2 weeks (ranging from 2 weeks to 12 months) to be included in the study. The PrDs were abstinent for at least 24 hours, which was confirmed with urine and breathing tests. For a more elaborative description of the PrD group, see Starcke et al. (2013). Urine samples were analyzed by a clinical toxicology laboratory (ATAL Medical Diagnostics, Amsterdam, the Netherlands) and none of the tests were positive for alcohol. Exclusion criteria for all groups included lifetime diagnosis of schizophrenia or psychotic episodes; 12-month diagnosis of manic disorder, obsessive compulsive disorder and post-traumatic stress disorder; other substance use disorders than alcohol dependence (except for nicotine); treatment for mental disorders in the past 12 months (except for alcohol dependence); use of psychotropic medication; difficulty reading Dutch, age under 18 years; IQ below 80 (measured by the Dutch Adult Reading Test; Schmand et al. 1991); positive urine screen for alcohol, amphetamines, benzodiazepines, opioids or cocaine; history or current treatment for neurological disorders; major internal disorders; brain trauma; or exposure to other neurotoxic factors than alcohol. Assessments and statistical analyses Functional magnetic resonance imaging (fMRI) paradigm In order to test cognitive flexibility, we applied an fMRI compatible self-paced switch task adapted from Sohn et al. (2000). During each trial, a stimulus consisting of one letter and one digit was shown, the color of which indicated the task to be performed: vowel/consonant (blue) or odd/even (red) judgments. Letters were taken from the set [a, e, i, u, b, c, d, f] and digits were taken from the set [2, 4, 6, 8, 3, 5, 7, 9]. Two consecutive trials never contained the same letter or digit. Color–task and stimulus–response associations were counterbalanced across subjects. Task switching (color change) occurred randomly after four to six trials to avoid predictability. First trials immediately after a color switch were defined © 2014 Society for the Study of Addiction

3

as ‘switch trials’, all other trials were defined as ‘repeat trials’. For example, on a switch trial, the subject had to switch from making a number judgment to a letter judgment or vice versa. If the target was blue, the criterion dimension was letter, and the criterion response was a left key press for a vowel and a right key press for a consonant. The task ended after 32 task switches and after 160 repeat trials, in a single run that took around 21 minutes to be completed. The task was self-paced with a maximum response window of 4000 milliseconds per trial and an interstimulus interval of 500 milliseconds. If the subject did not respond within 4000 milliseconds, the trial was considered an error and the next trial appeared. In addition, the paradigm included six 30-second baseline blocks, during which a fixation cross was shown and in which the subject had to remain still and wait until the task resumed, i.e. a passive baseline condition. Subjects performed a training session outside the scanner. There was no provision of feedback during the task. Behavioral variables of interest included (1) percentage correct task switches; (2) percentage of correct task repeats; and (3) switch cost [reaction times (RTs) of switch trials minus RTs of repeat trials]. Behavioral analysis. Demographic and behavioral data were analyzed using univariate analysis of variance (ANOVA and ANCOVAs) and Tukey’s post hoc tests in SPSS 20.0 (IBM Corp., Armonk, NY, USA). Non-normally distributed data (i.e. age, AUDIT scores and number of drinks per week) were analyzed using Kruskal–Wallis tests, and post hoc tests were conducted with Mann– Whitney tests. A paired-sample t-test was used to investigate differences between switch and repeat conditions across groups. Furthermore, an ANOVA test was used to assess group differences in the task performance. Finally, the effect of abstinence duration was assessed in a correlation analysis with task performance, brain activity and WM integrity only in AD patients because PrDs were mostly abstinent for less than 2 weeks. All analyses were performed two-tailed with α set at 0.05. fMRI analysis. Imaging data were obtained using a 3.0 T Intera whole-body fMRI scanner (Philips Medical Systems, Best, the Netherlands) with a phased array SENSE RF eight-channel receiver head coil. Task stimuli were projected on a screen behind the subject’s head at the end of the scanner table. The screen was visible for the subject through a mirror mounted above the subject’s head. Two magnet-compatible response boxes were used to record the subject’s responses. A total of 35 axial slices (voxel size 2.29 × 2.29 × 3 mm) T2*-weighted echo planar images (EPIs) were obtained, without interslice gap, a matrix size of 96 × 96 mm2, repetition time/echo time = 2.3 seconds/30 milliseconds and a bandwidth of Addiction Biology

4

Jochem M. Jansen et al.

90 kHz. These images are sensitive to blood oxygenation level-dependent (BOLD) contrast and covered the entire brain, except for the inferior regions of the cerebellum. A T1-weighted structural scan was made for co-registration with the fMRI data (voxel size 1 × 1 × 1 mm; 170 slices). Functional imaging analyses were performed with SPM5 (Statistical Parametric Mapping, Wellcome Trust Centre for Neuroimaging, London, UK). Images were manually reoriented and subsequently slice-timed, and realigned and unwarped. Next, images were warped to MNI space using each subject’s co-registered T1 image, and spatially smoothed using an 8-mm full width at half maximum Gaussian kernel. fMRI data were analyzed in the context of the general linear model using delta functions convolved with a canonical hemodynamic response function to model responses to each type of stimulus (switch trials and repeat trials). Contrast images containing parameter estimates were entered into a second-level (random effects) analysis. Main effects across groups for each contrast were analyzed using one-sample t-test implemented in SPM5 and are reported at P < 0.05 corrected for multiple comparisons at the voxel level across the whole brain, according to the family-wise error (FWE) method. Peak significant voxels per group were extracted with Marsbar v0.43 in order to assess the relationship between activity levels, task performance and DTI measurements (Brett et al. 2002). Group interactions were investigated using specific a priori regions of interest (ROIs) with a threshold set at P < 0.05, FWE corrected for multiple comparisons (Friston et al. 1996). The dlPFC, vlPFC, ACC, posterior parietal cortex and basal ganglia were defined as a priori ROIs, given their role in cognitive switching (Sohn et al. 2000; Smith et al. 2004; Ravizza & Carter 2008). The basal ganglia ROI, including the putamen, was defined using the WFU PickAtlas Tool v2.4 (Maldjian et al. 2003) that incorporates the automatic anatomical labeling atlas (Tzourio-Mazoyer et al. 2002). Activity in the dlPFC, posterior parietal cortex and vlPFC was detected by using ROI templates of the dlPFC, parietal cortex, and the middle orbitofrontal cortex and inferior orbitofrontal cortex, respectively. These ROIs were defined using the online Voxel of Interest Database (Fox & Lancaster 1994). Second, for exploratory purposes, we reported brain regions that were not predicted a priori but which met a threshold of P < 0.001, uncorrected with a cluster size of 10 contiguous voxels (Supporting Information Appendix S1). Functional connectivity. Functional connectivity analyses were conducted with the Conn toolbox in Matlab (Whitfield-Gabrieli & Nieto-Castanon 2012). Input for the functional connectivity analyses were the first level © 2014 Society for the Study of Addiction

switch versus repeat contrasts for each subject. Per subject, the correlation coefficient between the dlPFC and basal ganglia ROIs was computed per hemisphere and analyzed in SPSS. Diffusion tensor imaging: acquisition, pre-processing and fiber tracking Diffusion weighted EPIs were acquired along 32 directions with a b-value of 1000 seconds/mm2 and one acquisition without diffusion weighting (b = 0). All acquisitions were made with the following parameters: TR = 4.862 milliseconds, TE = 94 milliseconds, 38 axial interleaved slices with a 3-mm slice thickness with no gap, with a 112 × 110 mm2 matrix (0.898 × 0.898 mm in-plane resolution). The pre-processing of the DTI data was performed using in-house developed software, written in Matlab (The MathWorks Inc., Natick, MA, USA) and was executed on the Dutch Grid using a web interface to the e-Bioinfra gateway (Olabarriaga, Glatard & de Boer 2010; Shahand et al. 2011). We corrected for head motion and deformations induced by eddy currents by an affine registration of the diffusion weighted images (DWIs) to the non-diffusion weighted image. The gradient directions were corrected by the rotation component of the transformation. The DWIs were resampled isotropically. Rician noise in the DWIs was reduced by an adaptive noise filtering method (Caan et al. 2010). Diffusion tensors were estimated with a non-linear least squares procedure. From the resulting tensors, FA and MD maps were computed. Additional processing was required to model crossing fibers (BEDPOSTX, Bayesian Estimation of Diffusion Parameters Obtained using Sampling Techniques). Probabilistic fiber tracking was performed using [FMRIB’s Software Library (FSL) software (Behrens et al. 2007)]. Seed regions for the fiber tracking were based on the fMRI results by saving suprathreshold voxels in the dlPFC and basal ganglia masks as binary masks, reoriented and transformed back to single subject space with SPM8. These ROI-masks were resliced with the BEDPOSTX output file and subsequently used as seed and waypoint masks in fiber tracking. Fiber tracking was performed bilaterally. To improve intersubject correspondence along the individually tracked fibers, we non-rigidly registered (Ashburner 2007) the non-diffusion weighted images (b0) and warped the tracked fibers accordingly, with in-house written software in MATLAB (7.10.0, The MathWorks Inc.). The fiber with minimal distance to all fibers of all subjects was chosen as the reference fiber. The normal plane along this fiber was used to define correspondence to the fibers of all subjects. The FA profiles with 2-mm distance sampling were generated on the Addiction Biology

Brain function and WM in AUD

bilateral dlPFC-basal ganglia tracts and resulted in 10 mean FA and MD points along the tracts. Subsequent analyses were performed in SPSS using a repeated measures ANCOVA test per tract to check for group differences in mean FA and MD values, and partial correlation analysis to assess the relationship with task performance. In order to reduce the number of comparisons, the behavioral measurements of task switching were only correlated with those segments that differed most in FA and MD in the combined AD and PrD group, compared with the HC group. Greater differences are thought to reflect a higher degree of compromised WM integrity and these segments are therefore of interest. The difference between the combined alcohol group compared with the HC group was greatest in the most prefrontal segments (corresponding to number 10) of the left (for MD) and right hemisphere (for FA), indicating that WM integrity was mostly compromised in these areas. Partial correlation analysis between these areas and the task performance measurements were carried out. All DTI analyses were corrected for age because age has a well-known negative effect on WM integrity (Bennett et al. 2011).

RESULTS Demographics Table 1 summarizes the demographic and clinical characteristics for the three groups: AD, PrDs and HCs. There were no significant differences in age or IQ, but all groups significantly differed in their AUDIT scores and number of drinks per week. Post hoc comparisons revealed that the AD group had significantly higher scores for the AUDIT and on the number of standard drinks per week com-

5

pared with the PrD and the HC groups. Furthermore, the PrD group had higher AUDIT scores and more standard drinks per week compared with the HC group (AD > PrD > HC). Analysis of the effect of abstinence duration in AD patients revealed no significant effects on the ask performance, brain activity or WM integrity. Behavioral performance on the switch task Across groups, we found that subjects responded significantly faster during repeat trials (M = 1118.7, SD = 26.6) than to switch trials [M = 1551.0, SD = 316.9, t(66) = 17.25, P < 0.001]. In addition, across groups, subjects made more correct responses during repeat trials (M = 85.0 percent, SD = 13.4) than during switch trials [M = 76.2 percent, SD = 16.7, t(68) = −10.12, P < 0.001]. There were no significant differences between groups in the percentage of correct task switches, percentage of erroneous task repeats or switch costs (see Table 2). fMRI results Main effect task switch versus task repeat Increased activity during switch versus repeat trials in all groups was found in the whole brain analysis for the bilateral basal ganglia, ventrolateral and dorsolateral PFC, right middle temporal cortex, left anterior cingulate cortex and right premotor cortex (see Supporting Information Appendix S1). Group interaction task switch versus repeats Group interactions were analyzed with ROIs and the activity differed significantly between groups on their switch versus repeat brain activity, mainly in the dlPFC.

Table 1 Sample characteristics. Values for age, IQ, AUDIT, drinks per week, DSM-IV criteria for alcohol abuse and dependence are denoted as mean (standard deviation) and [min : max].

Age IQ AUDIT Number of standard drinks/week DSM-IV criteria alcohol abuse DSM-criteria alcohol dependence

Healthy controls (n = 22)

Problem drinkers (n = 21)

Alcohol dependent (n = 26)

38.5 (10.3) [22:53] 104.5 (12.5) [80:123] 5.1 (4.4) [0:15] 7.3 (7.2) [0:25] 0.14 (0.36) [0.0:1.0] 0.05 (0.22) [0.0:1.0]

42.1 (13.8) [22:61] 106.7 (14.3) [73:128] 18.5 (5.3) [11:33] 46.7 (31.5) [0:144] 0.8 (0.95) [0.0:3.0] 1.3 (1.46) [0.0:6.0]

44,3 (9.2) [21:59] 101.0 (15.1) [71:125] 27.9 (5.1) [11:36] 96.7 (61.2) [0.5:264] 1.9 (0.85) [1.0:3.0] 6.65 (1.56) [3.0:9.0]

F (d.f.)

Significance

3.08 (2)

ns

2.12 (2)

ns

51.83 (2)

P < 0.001

50.90 (2)

P < 0.001

30.32 (2)

P < 0.001

100.72 (2)

P < 0.001

AUDIT = alcohol use disorder identification test; d.f. = degrees of freedom; ns, non-significant. © 2014 Society for the Study of Addiction

Addiction Biology

6

Jochem M. Jansen et al.

Table 2 Behavioral measurements of task switching. This table shows the results of the ANCOVA test with age as covariate.

Percentage correct switches Switch cost (RT)a Percentage erroneous repeats

Healthy controls (n = 22)

Problem drinkers (n = 21)

Alcohol dependent (n = 26)

F value (d.f.)

Significance

78.1 (19.6) 136.7 (20.8) 2.0 (1.2)

79.0 (17.0) 128.1 (27.2) 2.1 (1.3)

74.2 (16.8) 124 (29.3) 2.5 (0.8)

0.55 (2,66) 1.16 (2,61) 1.81 (2,66)

ns ns ns

The number of participants is smaller for the switch cost analysis due to missing data (healthy controls = 19, problem drinkers = 20, alcohol dependent = 25). d.f. = degrees of freedom; ns = non-significant; RT = reaction time.

a

Table 3 Between group comparisons switch versus repeat functional magnetic resonance imaging. This table reports the differential activations between the groups for the pre-selected regions of interest. Significance levels are FWE-corrected, reported results with a P-value smaller than 0.05 are considered significant, whereas a P-value of P = 0.06 is considered trend significant.

Group comparison

Brain region

Hemisphere

Z-value

Number of voxels

x

y

z

P (FWE-corr)

HC > AD AD > HC AD > PrD PrD > HC – PrD > AD –

ns dlPFC ns dlPFC – dlPFC vlPFC

– Left – Left – Left Left

– 3.51 – 3.63 3.64 3.52 2.90



– −36 – −33 −36 −45 −45

– 30 – 36 33 30 15

– 18 – 27 30 30 −6

– 0.06 – 0.04 0.04 0.06 0.01

4 – 28 – 27 3

AD = alcohol-dependent patient; dlPFC = dorsolateral prefrontal cortex; few-corr = family-wise error correction; HC = healthy control; ns = non-significant; PrD = problematic drinker; vlPFC = ventrolateral prefrontal cortex.

Specifically, the PrD group particularly showed more activity in the left dlPFC than the HC group, and more activity in the left dlPFC than the AD group at a (trend) significant level (see Table 3). Task performance correlated with task performance only in HCs. The activity in the left dlPFC peak voxel was positively correlated with percentage correct switch trials (r = 0.48, n = 18, P < 0.05) and negatively correlated with erroneous repeats (r = −0.49, n = 17, P < 0.05). Functional connectivity fMRI time courses were extracted from the bilateral dlPFC and basal ganglia masks and tested for correlational changes between the regions during switch and repeat trials. Both left [t(45) = 4.27, P < 0.001] and right [t(45) = 4.85, P < 0.001] hemisphere connectivity between these ROIs were significantly higher during switch trials compared with repeat trials (left; r = 0.15 P < 0.001, right; r = 0.15, P < 0.001). Functional connectivity did not differ between groups and did not correlate with AUDIT scores or switch task performance. Fiber tracking The repeated measures ANCOVA revealed no significant interaction between group and segment for MD and FA © 2014 Society for the Study of Addiction

values in either hemisphere. The FA and MD values in the left and right hemisphere all revealed significant effects for segment, except for left MD values. This indicates that MD and FA values differ along the fiber tract. Main effect of group was only significant for the left and right MD values (see Table 4). Post hoc analyses for these significant group effects in MD revealed that HCs had significantly lower MD values than AD patients (left; P < 0.05, right; P < 0.01) and PrD (left; P < 0.05, right; P < 0.01), whereas AD patients and PrD did not significantly differ in MD values (see Fig. 1: HC < PrD = AD). Mean FA and MD values are reported in Supporting Information Appendix S2.

Relationship between FA values and task performance The results of the correlation analyses revealed that for the overall group (HCs + PrDs + ADs), prefrontal WM FA values positively correlated with the percentage correct switches and negatively correlated with erroneous repeats (see Table 5), whereas MD values were not correlated with task performance. The results of the analysis per group indicate that the relationships between FA and task performance were restricted to the HC group (see Table 6). Moreover, in PrD, MD values negatively correlated with percentage correct switches. FA or MD values were unrelated to functional connectivity. The MD and FA Addiction Biology

Brain function and WM in AUD

7

Table 4 Results for fiber tracking analysis for both left and right hemisphere. This table reports the results of the one-way repeated measures ANCOVAs that were conducted for FA and MD values in both hemispheres. Analyses are corrected for age. Left hemisphere (n = 55)

FA MD

Right hemisphere (n = 55)

Segment (repeated)

Group (between subjects)

Segment (repeated)

Group (between subjects)

F(9) = 10.94, P < 0.001 F(9) = 0.80, P = ns

F(2) = 0.96, P = ns F(2) = 5.87, P < 0.01

F(9) = 4.17, P < 0.05 F(9) = 6.79, P < 0.01

F(2) = 0.23, P = ns F(2) = 7.68, P < 0.001

FA = fractional anisotropy; MD = mean diffusivity; ns = non-significant.

Table 5 The results of the correlation analyses of the whole group. Results indicate that FA values are correlated with task performance. Significant results are displayed in bold.

Correlations

Percentage RT switch Percentage correct switches cost error repeata (n = 51) (n = 47) (n = 50)

Right FA segment 10 0.35* Left MD segment 10 0.01

0.17 −0.16

−0.37* 0.11

*Significant at P < 0.01, tested for two-tailed significance. aLogtransformed. FA = fractional anisotropy; MD = mean diffusivity; RT = reaction time.

values did not correlate with activity levels in any of the groups.

DISCUSSION The current study reports on the underlying functional and structural brain differences related to cognitive flexibility in HC, PrDs and AD patients. Unexpectedly, no group differences in cognitive flexibility were found, but the results of this study confirmed the previously reported role of the dlPFC, vlPFC, ACC, posterior parietal cortex and basal ganglia in switching behavior (Sohn et al. 2000; Smith et al. 2004; Ravizza & Carter 2008). Consistent with our hypotheses, PrD and AD groups revealed compromised WM integrity, especially MD, and this was associated with increased dlPFC activity during task switching. These findings are in line with the STAC theory, which states that higher prefrontal activation in cognitive demanding tasks in the presence of neural damage reflects a compensatory mechanism. Our findings resonate with this compensatory hypothesis: PrDs showed higher activations compared with the AD and HC groups (PrD > AD > HC), whereas both AD patients and PrDs had significantly compromised WM integrity when compared to controls (Goldin et al. 2008). In healthy aging, some studies have found that increased prefrontal activation is accompanied by decreased activation in more posterior parts of the brain; these results were not found in this study (Davis et al. 2008). Possibly, the © 2014 Society for the Study of Addiction

difference in results reflects differences between damage related to both healthy aging and the neurotoxic effects of alcohol. Based on the AUDIT scores, PrDs experienced less problems in daily life compared with AD patients, and PrDs and AD patients may therefore be considered to represent high and low functioning phenotypes, respectively. The current results suggest a compensatory mechanism in alcohol use disorders, in which additional prefrontal resources are recruited in order to compensate for impaired WM integrity and neural functioning. WM integrity, especially MD values, differed between groups and indicates a loss of WM integrity in PrDs and AD patients. FA values did not differ significantly between groups but were only related to task performance in HC. The absence of decreased FA values in AD compared with HCs is not in line with the previous research, which is however inconsistent, as both decrease (Pfefferbaum & Sullivan 2004; Pfefferbaum et al. 2009, 2010) and increase (De Bellis et al. 2008; Cardenas et al. 2013) in FA values compared with HCs have been reported. The association between WM integrity and cognitive flexibility is in line with a previous fiber tracking study on cognitive flexibility in HCs (van Schouwenburg et al. 2014). The MD values negatively correlated with task performance only in PrDs, whereas MD values were not correlated with task performance in the AD group. The negative association between lower MD values and higher task performance indicates that those PrDs with less damage to prefrontal WM show better switching performance. In AD patients, no such relationship was apparent as task performance did not correlate with any of the DTI measurements. This might indicate that WM damage in AD patients is too high to show an association between DTI and functional performance; however, a similar pattern of WM damage and higher functional activity during the switch task was found. These results are likely to present a compensatory mechanism in AD patients and PrDs where decreased WM integrity and related signal transfer is accompanied by increased neural recruitment. However, as the current study only assessed one specific fiber tract, it is not possible to make inferences on other WM tracts. Future studies should assess whether neural scaffolding is related to additional brain areas and other WM tracts. Addiction Biology

8

Jochem M. Jansen et al.

Figure 1 Fractional anisotropy (FA) and mean diffusivity (MD) profiles for prefrontal white matter.The fiber tract is divided in ten segments, segment 1 is the most subcortical segment and segment 10 is the most prefrontal segment, as is depicted at the top of the figure. Different colors within the brain correspond to different segments. For the graphs, differently colored lines represent the different groups. MD values are in mm2/s. Error bars denote standard error of the mean

These results should be viewed in light of some limitations. The non-significant difference between the groups on task performance might be related to practice effects, task design and/or sample sizes. First, all subjects practiced outside the scanner, which may have obscured performance differences. Second, as the focus of our experiment was to assess the differential neural © 2014 Society for the Study of Addiction

involvement in cognitive flexibility, the current version of the task was adapted for efficiently measuring the BOLD response while at the same time minimizing the amount of trials to keep the paradigm as short as possible. The small amount of switch trials may therefore have been less suited to assess behavioral differences. Based on previous studies on diminished cognitive flexibility in Addiction Biology

Brain function and WM in AUD

9

Table 6 The results of the correlation analysis per group. Results indicate that the FA values correlate with task performance only in the control group, whereas MD values only correlate with task performance in the problematic drinkers. Significant results are displayed in bold. Percentage error repeata

Percentage correct switches

RT switch cost

Correlations

HC (n = 15)

PrD (n = 13)

AD (n = 17)

HC (n = 13)

PrD (n = 12)

AD (n = 16)

HC (n = 14)

PrD (n = 12)

AD (n = 17)

Right FA segment 10 Left MD segment 10

0.65** 0.37

0.42 −0.53*

0.20 −0.03

−0.32 −0.03

0.22 −0.26

0.14 0.09

−0.66** −0.22

−0.36 0.37

−0.22 0.18

*Significant at P < 0.05, tested for two-tailed significance. **Significant at P < 0.01, tested for two-tailed significance. aLog-transformed. AD = alcoholdependent patient; FA = fractional anisotropy; HC = healthy control; MD = mean diffusivity; PrD = problematic drinker; RT = reaction time.

substance dependence (Hildebrandt et al. 2006; Pitel et al. 2009; van der Plas et al. 2009) and combined with our reported differences regarding neural involvement during cognitive switching, we argue that these neural differences are indicative of subtle impairments in cognitive flexibility in both AD subjects and PrDs. However, future studies should consider using a different switch paradigm [e.g. as was used in the work of de Ruiter et al. (2009)], which is powerful enough to detect behavioral differences to assess cognitive flexibility, and relate those results to WM integrity. Furthermore, there was no relationship between neural activity levels and DTI measurements. This would have underscored the results of this study, but this non-significant result does not dismiss the proposed compensatory mechanism in AD patients and PrDs. Although the non-significant result could be an indication that activity is unrelated to WM integrity, it may also reflect some methodological issue. The fiber tracking was based on the whole dlPFC ROI, whereas we correlated the DTI measurements with the peak activation voxel and the location of the peak voxel was slightly different for the three groups. As mentioned in the Materials and Methods section, the PrDs and AD patients differed in their period of abstinence based on the inclusion criteria of the study. We could therefore only perform correlation analyses for the AD group. Although abstinence did not have any relationship with task performance, fMRI or DTI measurements, the available data were not very specific and future studies should use a more precise way of determining abstinence duration. Furthermore, the AD patients all received cognitive behavioral therapy treatment, which may have had an effect on the brain function (Dichter, Felder & Smoski 2010; Zatorre, Fields & Johansen-Berg 2012). This may have influenced the results of the study, but unfortunately, data on the exact period of treatment were not collected. Finally, sample sizes, although adequate for detecting BOLD differences, may have limited the possibility to detect significant differences in behavioral performance. © 2014 Society for the Study of Addiction

CONCLUSION The current study reports on the underlying functional and structural brain differences associated with cognitive flexibility between HCs, PrDs and AD patients. Evidence is reported for preserved cognitive flexibility in PrDs and AD patients by recruiting more prefrontal resources in order to compensate for reduced WM integrity. Furthermore, PrDs showed more recruitment of compensatory prefrontal resources compared with AD patients. This compensatory mechanism could indicate a protective mechanism against a decline in cognitive functioning and/or transition into compulsive drug use, but future research should confirm these hypotheses. Acknowledgement This research was funded by The European Foundation for Alcohol Research (ERAB), grant no. EA 10 27 “Changing the vulnerable brain: a neuromodulation study in alcohol dependence” to AEG, WB and DJ. Authors Contribution Authors RJvH and JMJ have contributed equally to this paper and should be referred to as joint first authors. RvJ, AG and WvdB designed the study. RvJ was responsible for the data collection. RvJ and JJ were responsible for the data analysis. DV provided assistance for the fMRI analysis, whereas MC did so for the DTI data analysis. Then JJ interpreted the results and wrote the paper. All authors have critically reviewed content and approved final version submitted for publication.

References Ashburner J (2007) A fast diffeomorphic image registration algorithm. Neuroimage 38:95–113. Badre D, Wagner AD (2006) Computational and neurobiological mechanisms underlying cognitive flexibility. Proc Natl Acad Sci U S A 103:7186–7191. Addiction Biology

10

Jochem M. Jansen et al.

Baler RD, Volkow ND (2006) Drug addiction: the neurobiology if disrupted self-control. Trends Mol Med 12:559–566. Beaulieu C (2002) The basis of anisotropic water diffusion in the nervous system—a technical review. NMR Biomed 15:435– 455. Behrens TEJ, Berg HJ, Jbabdi S, Rushworth MFS, Woolrich MW (2007) Probabilistic diffusion tractography with multiple fibre orientations: what can we gain? Neuroimage 34:144– 155. Bennett IJ, Madden DJ, Vaidya CJ, Howard JH, Jr., Howard DV (2011) White matter integrity correlates of implicit sequence learning in healthy aging. Neurobiol Aging 32:2317.e1– 2317.e12. Brett M, Anton JL, Valabregue R, Poline JB (2002) Region of interest analysis using the MarsBar toolbox for SPM 99. Neuroimage 16:S497. Bush K, Kivlahan DR, McDonell MB, Fihn SD, Bradley KA (1998) The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch Intern Med 158:1789– 1795. Caan MW, Khedoe G, Poot D, den Dekker A, Olabarriaga S, Grimbergen K, van Vliet L, Vos F (2010) Adaptive noise filtering for accurate and precise diffusion estimation in fiber crossings. Med Image Comput Comput Assist Interv 13 (Pt 1):167–174. Cardenas VA, Greenstein D, Fouche JP, Ferrett H, Cuzen N, Stein DJ, Fein G (2013) Not lesser but greater fractional anisotropy in adolescents with alcohol use disorders. Neuroimage Clin 2:804–809. Chanraud S, Reynaud M, Wessa M, Penttila J, Kostogianni N, Cachia A, Artiges E, Delain F, Perrin M, Aubin HJ, Cointepas Y, Martelli C, Martinot JL (2009) Diffusion tensor tractography in mesencephalic bundles: relation to mental flexibility in detoxified alcohol-dependent subjects. Neuropsychopharmacology 34:1223–1232. Charlet K, Beck A, Jorde A, Wimmer L, Vollstädt-Klein S, Gallinat J, Walter H, Kiefer F, Heinz A (2014) Increased neural activity during high working memory load predicts low relapse risk in alcohol dependence. Addict Biol 19:402–414. Charlet K, Schlagenhauf F, Richter A, Naundorf K, Dornhof L, Weinfurtner CE, König F, Walaszek B, Schubert F, Müller CA, Gutwinski S, Seissinger A, Schmitz L, Walter H, Beck, A, Gallinat J, Kiefer F, Heinz A (2014) Neural activation during processing of aversive faces predicts treatment outcome in alcoholism. Addict Biol 19:439–451. Cooper JA, Worthy DA, Maddox WT (2013) Scaffolding across the lifespan in history-dependent decision making. Psychol Aging 28:505–514. Davis SW, Dennis NA, Daselaar SM, Fleck MS, Cabeza R (2008) Que PASA? The posterior-anterior shift in aging. Cereb Cortex 18:1201–1209. De Bellis MD, Van Voorhees E, Hooper SR, Gibler N, Nelson L, Hege SG, MacFall J (2008) Diffusion tensor measures of the corpus callosum in adolescents with adolescent onset alcohol use disorders. Alcohol Clin Exp Res 32:395–404. de Ruiter MB, Veltman DJ, Goudriaan AE, Oosterlaan J, Sjoerds Z, van den Brink W (2009). Response perseveration and ventral prefrontal sensitivity to reward and punishment in male problem gamblers and smokers. Neuropsychopharmacology 34:1027–1038. Dichter GS, Felder JN, Smoski MJ (2010) The effects of brief behavioral activation therapy for depression on cognitive © 2014 Society for the Study of Addiction

control in affective contexts: an fMRI investigation. J Affect Disord 126:236–244. Everitt BJ, Robbins TW (2005) Neural systems of reinforcement for drug addiction: from actions to habits to compulsion. Nat Neurosci 8:1481–1489. Fox PT, Lancaster JL (1994) Neuroscience on the net. Science 266:994–996. Friston KJ, Holmes A, Poline JB, Price CJ, Frith CD (1996) Detecting activations in PET and fMRI: levels of inference and power. Neuroimage 4 (3 Pt 1):223–235. Goh JO, Park DC (2009) Neuroplasticity and cognitive aging: the scaffolding theory of aging and cognition. Restor Neurol Neurosci 27:391–403. Goldin PR, McRae K, Ramel W, Gross JJ (2008) The neural bases of emotion regulation: reappraisal and suppression of negative emotion. Biol Psychiatry 63:577–586. Goldstein RZ, Volkow ND (2011) Dysfunction of the prefrontal cortex in addiction: neuroimaging findings and clinical implications. Nat Rev Neurosci 12:652–669. Hagmann P, Jonasson L, Maeder P, Thiran JP, Wedeen VJ, Meuli R (2006) Understanding diffusion MR imaging techniques: from scalar diffusion-weighted imaging to diffusion tensor imaging and beyond. Radiographics 26 (Suppl. 1):s205–s223. Harper C (2007) The neurotoxicity of alcohol. Hum Exp Toxicol 26:251–257. Harper C (2009) The neuropathology of alcohol related brain damage. Alcohol Alcohol 44:136–140. Harris GJ, Jaffin SK, Hodge SM, Kennedy D, Caviness VS, Marinkovic K, Papadimitriou GM, Makris N, Oscar-Berman M (2008) Frontal white matter and cingulum diffusion tensor imaging deficits in alcoholism. Alcohol Clin Exp Res 32:1001– 1013. Hildebrandt H, Brokate B, Hoffmann E, Kröger B, Eling PATM (2006) Conditional responding is impaired in chronic alcoholics. J Clin Exp Neuropsychol 28:631–645. Koob GF, Volkow ND (2010) Neurocircuitry of addiction. Neuropsychopharmacology 35:217–238. Kril JJ, Halliday GM, Svoboda MD, Cartwright H (1997) The cerebral cortex is damaged in chronic alcoholics. Neuroscience 79:983–998. Lezak MD (2004) The behavioral geography of the brain. In: Lezak MD, ed. Neuropsychological Assessment, 4th edn, pp. 39–85. New York: Oxford University Press. Maldjian JA, Laurienti PJ, Burdette JB, Kaft RA (2003) An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage 19:1233–1239. Moore A, Malinowski P (2009) Meditation, mindfulness and cognitive flexibility. Conscious Cogn 18:176–186. Nyhus E, Barcelo F (2009) The Wisconsin Card Sorting Test and the cognitive assessment of prefrontal executive functions: a critical update. Brain Cogn 71:437–451. Olabarriaga SD, Glatard T, de Boer PT. (2010) A virtual laboratory for medical image analysis. IEEE Trans Inf Technol Biomed 14:979–985. Park DC, Reuter-Lorenz P (2009) The adaptive brain: aging and neurocognitive scaffolding. Annu Rev Psychol 60:173– 196. Pfefferbaum A, Sullivan EV (2004) Disruption of brain white matter microstructure by excessive intracellular and extracellular fluid in alcoholism: evidence from diffusion tensor imaging. Neuropsychopharmacology 30:423–432. Pfefferbaum A, Rosenbloom M, Rohlfing T, Sullivan EV (2009) Degradation of association and projection white matter Addiction Biology

Brain function and WM in AUD

systems in alcoholism detected with quantitative fiber tracking. Biol Psychiatry 65:680–690. Pfefferbaum A, Rosenbloom MJ, Fama R, Sassoon SA, Sullivan EV (2010) Transcallosal white matter degradation detected with quantitative fiber tracking in alcoholic men and women: selective relations to dissociable functions. Alcohol Clin Exp Res 34:1201–1211. Pitel AL, Rivier J, Beaunieux H, Vabret F, Desgranges B, Eustache F (2009) Changes in the episodic memory and executive functions of abstinent and relapsed alcoholics over a 6 month period. Alcohol Clin Exp Res 33:490–498. Ravizza SM, Carter CS (2008) Shifting set about task switching: behavioral and neural evidence for distinct forms of cognitive flexibility. Neuropsychologia 46:2924–2935. Schmand B, Bakker D, Saan R, Louman J (1991) [The Dutch Reading Test for Adults: a measure of premorbid intelligence level]. Tijdschr Gerontol Geriatr 22:15–19. Schulte T, Müller-Oehring EM, Pfefferbaum A, Sullivan EV (2010) Neurocircuitry of emotion and cognition in alcoholism: contributions from white matter fiber tractography. Dialogues Clin Neurosci 12:554–560. Schulte T, Müller-Oehring EM, Sullivan EV, Pfefferbaum A (2012) White matter fiber compromise contributes differentially to attention and emotion processing impairment in alcoholism, HIV-infection, and their comorbidity. Neuropsychologia 50:2802–2822. Shahand S, Santcroos M, Mohammed Y, Korkhov V, Luyf ACM, van Kampen A, Olabarriaga SD (2011) Front-ends to biomedical data analysis on grids. Proceedings of the HealthGrid Conference. Bristol, UK. Smith AB, Taylor E, Brammer M, Rubia K (2004) Neural correlates of switching set as measured in fast, event-related functional magnetic resonance imaging. Hum Brain Mapp 21:247–256. Sohn MH, Ursu S, Anderson JR, Stenger VA, Carter CS (2000) Inaugural article: the role of prefrontal cortex and posterior parietal cortex in task switching. Proc Natl Acad Sci U S A 97:13448–13453. Stalnaker TA, Takahashi Y, Roesch MR, Schoenbaum G (2009) Neural substrates of cognitive inflexibility after chronic cocaine exposure. Neuropharmacology 56 (Suppl. 1):63–72. Starcke K, Holst RJ, van den Brink W, Veltman DJ, Goudriaan AE (2013) Physiological and endocrine reactions to psychosocial stress in alcohol use disorders: duration of abstinence matters. Alcohol Clin Exp Res 37:1343–1350. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N (2002) Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 15:273–289. van der Plas EA, Crone EA, van den Wildenberg WP, Tranel D, Bechara A (2009) Executive control deficits in substance-

© 2014 Society for the Study of Addiction

11

dependent individuals: a comparison of alcohol, cocaine, and methamphetamine and of men and women. J Clin Exp Neuropsychol 31:706–719. van Holst RJ, van den Brink W, Veltman DJ, Goudriaan AE (2010) Why gamblers fail to win: a review of cognitive and neuroimaging findings in pathological gambling. Neurosci Biobehav Rev 34:87–107. van Schouwenburg MR, O’Shea J, Mars RB, Rushworth MF, Cools R (2012) Controlling human striatal cognitive function via the frontal cortex. J Neurosci 32:5631–5637. van Schouwenburg MR, Onnink AMH, ter Huurne N, Kan CC, Zwiers MP, Hoogman M, Franke B, Buitelaar JK, Cools R (2014) Cognitive flexibility depends on white matter microstructure of the basal ganglia. Neuropsychologia 53:171– 177. Vanes LD, van Holst RJ, Jansen JM, van den Brink W, Oosterlaan J, Goudriaan AE (2014) Contingency learning in alcohol dependence and pathological gambling: learning and unlearning reward contingencies. Alcohol Clin Exp Res 38:1602–1610. Volkow ND, Fowler JS, Wang GJ, Goldstein RZ (2002) Role of dopamine, the frontal cortex and memory circuits in drug addiction: insight from imaging studies. Neurobiol Learn Mem 78:610–624. Whitfield-Gabrieli S, Nieto-Castanon A (2012) Conn: a functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect 2:125–141. World Health Organization (1997) Composite International Diagnostic Interview—Version 2.1. Geneva: World Health Organization. Yeh PH, Simpson K, Durazzo TC, Gazdzinski S, Meyerhoff DJ (2009) Tract-Based Spatial Statistics (TBSS) of diffusion tensor imaging data in alcohol dependence: abnormalities of the motivational neurocircuitry. Psychiatry Res 173:22–30. Yucel M, Lubman DI (2007) Neurocognitive and neuroimaging evidence of behavioural dysregulation in human drug addiction: implications for diagnosis, treatment and prevention. Drug Alcohol Rev 26:33–39. Zatorre RJ, Fields RD, Johansen-Berg H (2012) Plasticity in gray and white: neuroimaging changes in brain structure during learning. Nat Neurosci 15:528–536.

SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Appendix S1 The main effect of task switch versus task repeat for the whole group Appendix S2 Mean FA and MD values per group and hemisphere

Addiction Biology

Brain function during cognitive flexibility and white matter integrity in alcohol-dependent patients, problematic drinkers and healthy controls.

Cognitive flexibility has been associated with prefrontal white matter (WM) integrity in healthy controls (HCs), showing that lower WM integrity is as...
370KB Sizes 0 Downloads 8 Views