OUP user menu

Nicotinic antagonist effects on functional attention networks

Renate Thienel, Bianca Voss, Thilo Kellermann, Martina Reske, Sarah Halfter, Abigail J. Sheldrick, Katrin Radenbach, Ute Habel, Jon Nadim Shah, Ulrich Schall, Tilo Kircher
DOI: http://dx.doi.org/10.1017/S1461145709990551 1295-1305 First published online: 1 November 2009

Abstract

Cholinergic neurotransmission has been implicated in memory and attention. We investigated the effect of the non-competitive nicotinic antagonist mecamylamine on three components of attention processes (i.e. alerting, orienting and executive control) in 12 healthy male subjects whilst performing the Attention Network Task (ANT) in a magnetic resonance imaging (MRI) scanner. Participants received 15 mg mecamylamine in a single blind and placebo-controlled randomized procedure 90 min prior to obtaining functional MRI data. Our results confirm previous reports of beneficial effects of cueing (alerting and orienting) and detrimental effects of conflict (executive control) on reaction times when performing the ANT. The functional MRI data confirmed distinct neural networks associated with each of the three attention components. Alerting was associated with increased left temporal lobe activation while orienting increased bilateral prefrontal, right precuneus and left caudate activation. Executive control activated anterior cingulate and precuneus. Mecamylamine slowed overall response time and down-regulated brain activation associated with orienting and to some extent brain activation associated with executive control when compared to placebo. These findings are consistent with nicotinic modulation of orienting attention by cueing and executive control when responding to conflicting information. The latter nicotine antagonist effect may be mediated via cholinergic modulation of dopamine neurotransmission in mesolimbic pathways.

Key words
  • Attention Network Task
  • functional magnetic resonance imaging
  • mecamylamine
  • nicotinic antagonist
  • pharmacological challenge

Introduction

A wide range of attention processes are mediated by frontal and parietal cortical networks (Posner & Dehaene, 1994). They can be subdivided into distinct neural systems, subserving alerting, orienting and executive control of attention processing (Posner & Petersen, 1990). Alerting refers to achieving and maintaining an ‘alert state’ which mainly involves noradrenergic up-modulation of frontal and parietal regions in the right hemisphere (Witte & Marrocco, 1997) while executive control, modulated by dopaminergic input, refers to ‘conflict resolution’ by engaging the anterior cingulate and the lateral prefrontal cortex (Bush et al. 2000; MacDonald et al. 2000; Marrocco & Davidson, 1998). The former is proposed to monitoring and the latter to resolving conflicts (Botvinick et al. 2001; MacDonald et al. 2000).

The focus of the current study is the cholinergic neuromodulation of orienting; i.e. the selection of information from sensory input, a cognitive process that has been hypothesized to engage the temporo-parietal junction and the inferior frontal gyrus in the right hemisphere (Corbetta et al. 2000). Converging evidence suggests nicotinic neurotransmission as the major mediator of orienting attention processes (see Ochoa & Lasalde-Dominicci, 2007 for review).

In humans, the highest levels of nicotine-binding sites have been indentified in frontal, cingulate and insular cortices, as well as in thalamus and basal ganglia by positron emission tomography (PET; Nyback et al. 1994). Functional magnetic resonance imaging (fMRI) further demonstrated increased brain activity in these brain regions (namely cingulate and frontal lobe, including the dorsolateral, orbital and mediofrontal cortex) as well as nucleus accumbens, amygdala and limbic thalamus in response to nicotine challenge (Stein et al. 1998). The latter structures are closely linked to the mesolimbic dopamine system (Corrigall et al. 1992; Nisell et al. 1994, 1995; Pontieri et al. 1996), thus mediating the rewarding and addictive effects of nicotine in smokers (Benowitz, 1992).

While nicotine effects on attention have been widely studied, the current experiments investigated the role of the non-competitive nicotine receptor antagonist mecamylamine on sub-components of attention. It is known from animal studies that mecamylamine impairs orienting of attention behaviourally (Stewart et al. 2001) and also suppresses the regional cerebral blood flow increase to nicotine challenge in cingulate, insular, medial prefrontal, and orbito-frontal cortical regions as well as in amygdala and dorsomedial hippocampus (Gozzi et al. 2006). In humans, mecamylamine dose-dependently slows reaction time across several cognitive domains (Newhouse et al. 1992), suggestive of widespread effects on cognition. However, the specific effects of mecamylamine on the sub-components of attention processing are unknown.

We employed the Attention Network Task (ANT) which has been developed in Posner's laboratory and adapted to the MRI environment (Fan et al. 2002) in order to probe the functional aspects of alerting, orienting and executive control of attention with event-related fMRI (Fan et al. 2005). Using the ANT, Fan et al. (2005) reported thalamic, anterior and posterior cortical activation associated with alerting, predominantly anterior cingulate activation associated with executive control, and parietal and frontal eye-field activation with orienting.

We predicted reduced frontal and parietal activation for the orienting component of the ANT with mecamylamine vs. placebo and detrimental effects on response times for spatial cueing. However, nicotine-mediated dopamine release in mesolimbic pathways also suggests some effects on executive control. For instance, Dumas et al. (2008) reported reduced frontal lobe activation with mecamylamine vs. placebo whilst their subjects were performing the N-back task in the MRI scanner. This observation suggests an indirect effect of mecamylamine on dopamine-mediated working-memory function. Hence we predicted decreased activation with mecamylamine associated with executive control and detrimental effects on response times when processing conflicting information in the ANT. Nicotine antagonistic effects on alerting should not be present.

Methods

The study was approved by the RWTH Aachen Medical Research Ethics Committee. All subjects gave written informed consent and received a small honorarium for participation.

Subjects

Twelve right-handed (Edinburgh Handedness Inventory Score 9±0.9), non-smoking, healthy male adults without a history of mental illness [including substance addiction or current abuse; SCID-I (First et al. 1997)] or any other relevant neurological or medical condition past or present took part (standard MRI exclusion criteria also applied). Participants' mean age was 26±2.1 yr (range 23–29 yr) and estimated mean verbal IQ was 118±13.8 [Mehrfachwahl-Wortschatztest (Multiple choice vocabulary test); Lehrl et al. 2005].

Study design

Mecamylamine was administered employing a single-blind crossover design alternating with scopolamine and placebo at least 1 wk apart. Drugs were given 90 min prior to each of three MRI sessions in a randomized, double-dummy fashion following procedures described elsewhere (Curran et al. 1991; Ebert et al. 1998; Newhouse et al. 1992; Safer & Allen, 1971): (1) 15 mg mecamylamine orally and saline intravenously; (2) 0.4 mg scopolamine intravenously and placebo orally; (3) or oral placebo and saline intravenously. Scopolamine results are reported elsewhere (Thienel et al. in press).

Stimuli and task

Subjects performed a modified version of the ANT (Fan et al. 2002, 2005; Konrad et al. 2005). The task (see Fig. 1) consisted of an initial cueing condition (0.15 s), followed by 0.4-s inter-stimulus interval with a central fixation cross presented and concluded with a target response interval of 1.5 s. Inter-trial intervals (i.e. target interval onset to cueing condition onset) ranged from 1.9 to 6.4 s (normally distributed around a mean of 3.35 s; Dale, 1999) whilst the central fixation cross was presented.

Fig. 1

The task consisted of an initial cueing condition (0.15 s), followed by 0.4-s inter-stimulus interval with a central fixation cross presented, and concluded with a target response interval of 1.5 s. Inter-trial intervals (i.e. target interval onset to cueing condition onset) ranged from 1.9 to 6.4 s whilst the central fixation cross was presented. The cueing condition consisted of no cue (central fixation cross), a centre cue or a spatial cue. The spatial cue predicted the left or right location of a target arrow on the screen. The target arrow was pointing to the left or right and was always presented with four flanker arrows (two above and two below the target arrow), either pointing in the same direction as the target arrow (congruent condition) or in the opposite direction (incongruent condition). The subjects were asked to make a button press response with their right index or middle finger dependent on the direction the target arrow was pointing.

The cueing condition consisted of no cue (central fixation cross), a centre cue or a spatial cue. The spatial cue predicted the left or right location of a target arrow on the screen (4° to the right or 4° to left of the central fixation cross; Fig. 1). The target arrow was pointing to the left or right and was always presented with four flanker arrows (two above and two below the target arrow), either pointing in the same direction as the target arrow (congruent condition) or in the opposite direction (incongruent condition; Fig. 1).

The subjects were asked to make a button press response with their right index or middle finger dependent on the direction the target arrow was pointing. Subjects were presented with 10 repetitions of the 24 possible stimulus combinations (three cue conditions, two flanker conditions, two target presentation sides, and two target pointing directions) in a randomized and counterbalanced sequence of 240 trials.

Data acquisition

MRI data were acquired on a 1.5 T Siemens Symphony (Jülich Research Centre) employing echo-planar imaging (EPI). Repetition time was 3 s with a flip angle of 90° and an echo time of 60 ms. Images consisted of 30 slices of 4 mm thickness (gap between slices was 0.4 mm) and resulting in an in-plane resolution of 64×64. The field of view was 200×200 mm2 resulting in a voxel size of 3.125×3.125×4 mm3. For each subject a series of 286 images was acquired, discarding the first three images to account for T1 stabilization effects.

Data analyses

MRI data were analysed using SPM2 (http://www.fil.ion.ucl.ac.uk/spm/). Images were aligned to the first volume using rigid-body transformations. Acquisition delays of individual slices were corrected by the 15th slice of each volume as reference in time before transforming into standard MNI space by normalizing to an EPI template. Data were spatially interpolated to a voxel size of 2×2×2 mm3 and spatially smoothed using an isotropic Gaussian kernel of 10 mm full width at half maximum.

Correctly performed trials were entered into the analysis with each trial assigned to each level of the three factors of interest (i.e. congruency, cueing, and visual field) resulting in 12 different conditions. First-level analyses were performed separately for each drug condition. Onset of each event was synchronized with the appearance of the target stimulus and modelled as event. T contrasts were calculated for each cell of the two factors congruency and cueing, which were subsequently entered into repeated-measures ANOVA at the second level. Potential reaction time confounds were tested with reaction time as covariate of no interest.

T tests were performed (Fan et al. 2005) contrasting (1) central vs. no cue (alerting), (2) spatial vs. central cue (orienting) and (3) incongruent vs. congruent (executive control) for the placebo condition, followed by contrasting the resulting T maps for alerting, orienting and executive control for mecamylamine vs. placebo, respectively. Non-sphericity correction was applied, thus accounting for hetero-scedasticity and covarying conditions. Resulting statistical maps were thresholded at p<0.01 (voxel level) and a minimum cluster size of 54 continuous voxels equivalent to a Monte-Carlo-corrected threshold of p<0.05 (1000 iterations; Slotnick et al. 2003). Anatomical labels were generated in SPM [automatic anatomic labelling (http://www.fil.ion.ucl.ac.uk/spm/ext/#AAL; Tzourio-Mazoyer et al. 2002].

Reaction times were analysed using a full-factorial linear mixed-effects model with the fixed factors congruency, cueing and drug (Pinheiro & Bates, 2000). Repeated measures and therefore non-independence of observations were accounted for by incorporating subjects as a random-effects factor with variance components modelling correlated measures.

Results

Subjective drug effects and behavioural data

Participants did not report any subjective placebo or drug effects with the exception of one participant who reported mild sedation with mecamylamine. Overall performance rate for correct responses was 98%. Mean reaction time (±s.e.m.) across all task conditions was 650±7 ms.

In the placebo condition, mean reaction time over all task conditions was 638±11 ms. Alerting (central vs. no cue) and orienting (spatial vs. central cue) facilitated response time by 31±7 ms and 75±11 ms, respectively [F(2, 55)>40.8, p<0.0001], while reaction times increased by 77±8 ms for incongruent vs. congruent flanker presentations [executive control: F(1, 55)=245.2, p<0.0001]. Cue interacted with flanker type [F(2, 55)=5.4, p<0.008] with the central cue increasing response times for the incongruent flanker condition (Fig. 2).

Fig. 2

Mean reaction times (s.e.m.) for correctly performed trials recorded in non-cued, centrally cued or spatially cued and congruent (Con) or incongruent (Incon) flanker conditions for placebo (left) and mecamylamine (right).

Mecamylamine did not affect overall correct response rate when compared to placebo (F<1.0). Nicotine antagonism slowed total mean response time across all task conditions by 24±17 ms [F(1, 121)=5.9, p=0.016], but without interacting with the different attention components. Hence, as under the placebo condition, alerting (central vs. no cue) and orienting (spatial vs. central cue) manipulations reduced response times by 29±6 ms and 77±6 ms, respectively [F(2, 55)=99.9, p<0.0001], and incongruent vs. congruent flanker conditions (assessing executive control) slowed response times by 70±9 ms [F(1, 55)=24.1, p<0.0001], without interacting with drug condition for each of the attention components (F<1.0). Similar to placebo, there was also a significant interaction of cue by flanker type [Fig. 2; F(2, 55)=4.8, p=0.01].

Functional imaging data

Consistent with equal reaction time effects for all stimulus types, reaction time as covariate of no interest did not alter any activation clusters and was therefore excluded from further analyses.

Alerting was associated with a blood oxygenation level dependent (BOLD) increase in the left interior temporal gyrus in the placebo condition (Table 1, Fig. 3). Mecamylamine increased the alerting effect in left superior orbito-frontal and right precentral, middle temporal, and middle occipital gyri (Table 1, Fig. 2).

Fig. 3

For the alerting contrast (central cue >no cue) the glass brain and intersecting sagittal, coronal, and transaxial slices (SPM-2 T1 template) show (a) left inferior temporal gyrus activation with placebo and (b) increased left superior orbitofrontal, right middle temporal, occipital and precentral gyri activation with mecamylamine >placebo (Monte Carlo-corrected threshold of p<0.05 for clusters >54 continuous voxels). (c) The parameter plots show an increase in the left superior orbito-frontal gyrus after no cue vs. central cue conditions under placebo, while the opposite pattern is induced by mecamylamine.

View this table:
Table 1

Orienting was associated with increased BOLD bilaterally in prefrontal cortex, right precuneus, and left caudate (Table 2, Fig. 4). The prefrontal areas were identified as regions activated by placebo and drug×placebo interaction (i.e. reduced activation with mecamylamine vs. placebo). Parameter plots of the left superior frontal gyrus confirm drug×stimulus interactions due to mecamylamine reversing the activation increase in response to spatial cues with placebo as well as the inverse effect in response to central cues.

Fig. 4

For the orienting contrast (spatial cue >central cue) the glass brain and intersecting sagittal, coronal, and transaxial slices (SPM-2 T1 template) show (a) bilateral superior and right inferior frontal gyrus, right precuneus and left caudate activation with placebo and (b) increased bilateral superior and right inferior frontal, right superior, and occipital gyrus and left thalamus activation with mecamylamine <placebo (Monte Carlo-corrected threshold of p<0.05 for clusters >54 continuous voxels). (c) The parameter plots show an increase in the left superior frontal gyrus after spatial cues and a reduction after central cues, while the opposite pattern is induced by mecamylamine.

View this table:
Table 2

Other areas which were not activated under placebo, but exhibited a decrease in activation in the orienting condition with mecamylamine, included the right superior occipital gyrus and the left thalamus while parahippocampal gyrus, posterior cingulate and rolandic operculum in the left hemisphere. (Table 2 and Fig. 4).

Executive control was associated with increased BOLD bilaterally in anterior cingulate, right superior frontal and parietal gyri, left gyrus rectus, right angular and left inferior occipital gyrus and bilateral precuneus (Table 3, Fig. 5). Regions showing activation with placebo and as a drug×placebo interaction were indentified in the left hemisphere for gyrus rectus, precuneus and superior parietal areas with reduced activation with mecamylamine vs. placebo. Parameter plots of the left precuneus confirm drug×stimulus interactions due to mecamylamine differentially reversing the activation increase in response to incongruent flankers as seen with placebo while not interacting with the processing of congruent flankers. Other areas with decreased executive control effects included the right gyrus rectus, vermis, left calcarine and right superior parietal gyrus while increased executive control effects were present in the left inferior parietal gyrus (Table 3, Fig. 4).

Fig. 5

For the executive control contrast (incongruent flanker >congruent flanker) the glass brain and intersecting sagittal, coronal, and transaxial slices (SPM-2 T1 template) show (a) bilateral precuneus, angular gyrus and anterior cingulate, left gyrus rectus, inferior occipital gyrus and right superior parietal gyrus activation with placebo and (b) increased left precuneus, right superior parietal gyrus, bilateral gyrus rectus, left calcarine gyrus, and vermis activation with mecamylamine <placebo (Monte Carlo-corrected threshold of p<0.05 for clusters >54 continuous voxels). (c) The parameter plots show that incongruent cues under placebo lead to a BOLD increase in the left precuneus. This effect is selectively reversed by mecamylamine.

View this table:
Table 3

Discussion

Our behavioural data confirm beneficial performance effects of cueing and detrimental effects of conflict on attention processing when performing the ANT (Fan et al. 2002). Our functional brain-imaging data further confirm separable neural networks associated with alerting, orienting or executive control of attention processing (Fan et al. 2005) with specific effects of mecamylamine on individual attention networks, despite a lack of drug-specific effects on ANT components other than an overall increase of response time.

Previous studies (Corbetta et al. 2000; Fan et al. 2005) reported right-hemispheric activations in the temporo-parietal junction and the inferior frontal gyrus when assessing for alerting (central vs. no cue). Our placebo data did not confirm these previous reports other than an increased BOLD response in left interior temporal gyrus with alerting. These inconsistencies with previous findings may be due to differences in our study design when presenting the central cue as a fixation aid instead of a fixation cross, thus diminishing the ‘alerting salience’ of the cueing stimulus.

However, largely consistent with previous reports (Corbetta et al. 2000; Fan et al. 2005) orienting (spatial vs. central cue) activated a predominantly right-hemispheric neural network comprised of inferior and superior frontal cortical areas, precuneus and caudate under placebo conditions. Corbetta and colleagues (2000) suggested the intraparietal sulcus to be mainly involved in orienting following cue presentation while re-orienting (invalid vs. valid cue) towards visual targets presented at unattended locations appears to involve the right temporo-parietal junction and the right inferior frontal gyrus. However, re-orienting was not investigated in the current study. Notwithstanding, our data confirm right-hemispheric inferior frontal cortex activation for orienting but differ in parietal activation patterns by involvement of the right precuneus rather than the intraparietal sulcus as in Corbetta et al. (2000). However, according to Corbetta et al. (2000) the precuneus was reported to show increased activation when subjects were engaged in target detection during the orienting process.

Executive control (incongruent vs. congruent flanker conditions) activated anterior cingulate, precuneus, and superior parietal and frontal gyri consistent with previous findings (Fan et al. 2005). In particular, the anterior cingulate is engaged in conflict resolution (Botvinick et al. 2001; Bush et al. 2000) which is captured by the ANT when introducing incongruent flanker conditions while executive function tasks generally activate fronto-parietal cortical areas (Perfetti et al. 2007; Specht et al. 2008). Nicotinic blockade in our study lead to a decrease in activation in main parts of the executive control network, comprising precuneus, superior parietal gyrus and gyrus rectus.

In line with Fan et al. (2002) we found an interaction between the cueing and congruency condition, with an increase in the amount of flanker interference after the central cue was presented. Together with our functional brain-imaging data this further confirms separable but partly interacting neural networks associated with alerting, orienting or executive control of attention (Fan et al. 2005).

Nicotine antagonism by mecamylamine slowed overall response time but did not specifically modulate orienting and executive control performance of the ANT as predicted by animal research for the orienting of attention processes (Stewart et al. 2001). Ceiling effects in task performance may have concealed differential behavioural effects, rendering our sample size retrospectively too small when investigating a very well performing cohort, as in this case.

Notwithstanding, our functional imaging findings support the predictions, i.e. nicotine antagonism by mecamylamine significantly reduced brain activation in the orienting as well as executive control networks. Conversely, mecamylamine induced an up-regulation in brain regions usually not associated with alerting, which may suggest a non-specific compensatory drug effect. This may also explain increased BOLD response to mecamylamine in areas usually not activated in the placebo condition when assessing for orienting and executive control.

Cerebrovascular coupling may explain some of these non-specific drug effects. It refers to the effect of a drug on global and local blood flow; i.e. increases in cerebral blood flow occur independently of changes in neuronal metabolic activity, thus introducing a potential confound to BOLD signal changes. Nicotine, for instance, appears to affect global blood flow and cerebral oxygen uptake to some extent (Ghatan et al. 1998; but see Jacobsen et al. 2002 for negative findings). Mecamylamine as a nicotinic antagonist certainly exerts antihypertensive properties, thus affecting global blood flow. However, potential neurovascular coupling does not explain the specificity of our findings in respect to neural networks and attention conditions, as it cannot explain the differential responses to specific stimuli such as spatial cue vs. central cue for example.

In summary, our study demonstrates that individual functional networks of attention can be selectively down-regulated by mecamylamine challenge. Our findings are consistent with the notion of direct cholinergic modulation of orienting attention (Davidson et al. 1999; Giessing et al. 2006; Thiel et al. 2005) via nicotinic antagonism. Our findings also demonstrate nicotine antagonist effects on the executive control network of attention, which is consistent with nicotinic modulation of dopamine neurotransmission in frontal attention networks (Jones et al. 2001; Klink et al. 2001). As such, pharmacological challenge of attention processes has the potential to investigate more specifically clinical populations that are impaired in attention, thus allowing for selectively testing attention sub-processes and their response to agonist and antagonist intervention by functional brain imaging. This is of particular clinical importance, given that most psychopharmacological intervention impacts on cholinergic and/or monoaminergic neurotransmission, either directly or indirectly, due to their spectrum of receptor interaction.

Such an approach would also allow investigation of pharmacological effects on attention in clinical conditions, such as attention deficit hyperactivity disorder (ADHD) or schizophrenia. For instance, ADHD (combined type) is associated with significantly slowed response times to non-spatial cues (alerting) and cues with spatial conflict (Oberlin et al. 2005) whereas schizophrenia subjects present with significant deficits in orienting attention and executive control but intact alerting (Depatie et al. 2002; Levin et al. 1996; Sacco et al. 2005). Intervention in both conditions targets dopamine transmission, either by increasing dopamine via indirect dopamine agonists, thus putatively improving executive control in ADHD, or via D2 dopamine receptor blockade in the case of psychosis, thus potentially aggravating deficits in executive control in schizophrenia. Nicotine abuse may counter some of these adverse antipsychotic side-effects on attention, such as on executive control, which coincides with a high rate of cigarette smoking in schizophrenia patients (for review see Berman & Meyer-Lindenberg, 2004). Pharmacological fMRI can provide a tool to investigate these effects in relation to sub-components of attention and alterations in neurotransmission.

Acknowledgements

We thank radiographers Cordula Kemper and Dorothe Krug at the Jülich Research Centre for help with fMRI scanning, and Helen Stain and Brian Kelly from the Centre for Rural & Remote Mental Health in Orange for encouraging data analysis and presentation. We gratefully acknowledge the participation of our subjects. This work was supported by the Deutsche Forschungsgemeinschaft (KFO 112).

Statement of Interest

None.

References

View Abstract