Abstract

We examined whether the known noradrenergic attenuation of the alerting effect (the beneficial effect of a warning cue) results from an underlying effect of noradrenaline on temporal orienting (orienting toward a particular moment in time). Following a within-subjects, counterbalanced design, 10 healthy human volunteers received placebo, 200 μg clonidine or 1 mg guanfacine (α2 agonists) in three separate testing sessions. Subjects were scanned by fMRI while performing attentional orienting tasks containing spatially informative, temporally informative, non-informative or no cues. The alerting effect primarily activated left-lateralized prefrontal, premotor and parietal regions. Clonidine, but not guanfacine, impaired behavioural measures of the alerting effect while attenuating actvity in the left temporo-parietal junction. Replicating previous results, the temporal orienting task activated left parietal and frontal cortex, while parietal cortex was activated bilaterally during spatial orienting. Of these networks, clonidine, but not guanfacine, attenuated left prefrontal cortex and insula activity during temporal orienting and attenuated right superior parietal cortex activity during spatial orienting,. To complement these neuroanatomical changes, clonidine produced selective behavioural effects on both temporal and spatial orienting. The anatomical dissociation between the effects of clonidine during temporal orienting versus alerting suggests that noradrenergic modulation of the alerting effect does not result only from an underlying effect on temporal orienting. Furthermore, we have demonstrated lateralized neuroanatomical substrates for the noradrenergic modulation of human attentional orienting in the spatial and temporal domains.

Introduction

Arousal and selective attention are multidimensional psychological processes that interact closely with one another, both behaviourally and neuroanatomically. There is accumulating evidence that the noradrenergic (NA) system may contribute to both arousal and selective attention, as well as play an important role in mediating the interaction between these functions (Arnsten and Constant, 1992; Smith and Nutt, 1996; Coull et al., 1997).

Visual spatial orienting tasks, originally developed by Posner (Posner et al., 1980), have provided an experimental paradigm with which to investigate both selective directed focal spatial expectations as well as general, unfocused alerting. Fernandez-Duque and Posner have recently demonstrated that the spatial orienting and alerting effects in the Posner task provide independent indices of attention and arousal, respectively (Fernandez-Duque and Posner, 1997). In spatial orienting tasks subjects detect peripheral visual targets that appear after either informative or neutral cues. Informative cues predict, with a high probability, the target location. Neutral cues signal the imminent appearance of the target, but provide no spatial information. Subjects are typically fastest to respond to targets whose location has been correctly predicted by the cue (‘valid’ trials) and slowest to detect targets that appear incorrectly in the non-predicted location (‘invalid’ trials). Reaction times to detect targets after neutral cues are usually intermediate (Posner et al., 1980). The advantage that valid predictive cues confer over neutral ones (‘validity effect’) represents the effect of covert focused spatial orienting. The disadvantage that invalid predictive cues bring relative to neutral ones (‘invalidity effect’) represents the cost of recovery, after engaging attention at the wrong location. Neutral cues also confer a behavioural advantage, when compared with conditions where no cue precedes the target (‘alerting effect’), which has been linked to the beneficial arousing effect of a neutral warning cue.

Witte and Marrocco have investigated the effects of NA agents on the spatial orienting and alerting responses using the Posner attentional orienting task in monkeys and compared them with the effects of cholinergic agents (Witte and Marrocco, 1997a). Infusion of the α2 noradrenergic agonist clonidine or guanfacine impaired the alerting affect while having no effect on the spatial validity effect. In contrast, infusion of the cholinergic agonist nicotine impaired spatial orienting, but had no effect on alerting (Witte and Marrocco, 1997b). This dissociation was interpreted to reflect NA modulation of the benefit produced by a temporal cue, versus cholinergic modulation of the ability to benefit from a spatial cue (Marrocco et al., 1994). Contrary to the null result of clonidine upon spatial orienting obtained by Witte and Marrocco (Witte and Marrocco, 1997a), Clark and co-workers have reported a significant effect of clonidine upon selective spatial orienting in humans (Clark et al., 1989). Clonidine attenuated the cost of being invalidly cued to the wrong spatial location. The ability of NA modulation to affect spatial orienting in addition to arousal functions thus remains unresolved.

We have recently designed a variant of the Posner orienting task that specifically manipulates the deployment of attentional resources toward specific time intervals (Coull and Nobre, 1998; Nobre et al., 1999). We found evidence for hemispheric specialization of right versus left parietal cortex for spatial versus temporal attentional orienting, respectively, suggesting that these two processes can be distinguished. In the temporal orienting conditions of the task informative cues predict with high probability the exact time interval at which the target will appear. The introduction of a selective temporal orienting condition allows specific expectations about time to be disentangled from more general arousing effects of warning cues. For example, the benefit derived from a temporal warning cue during the alerting effect could be due to directed temporal orienting, but other possibilities include a general increase in phasic arousal or even the ability to prepare a motor response in advance. Our task therefore enables further and clearer fractionation between selective spatial expectations, selective temporal expectations and other alerting effects that do not direct the subject toward specific locations or time intervals.

We imaged the brain activity in healthy human volunteers under different states of NA modulation while they detected peripheral targets during conditions of spatial orienting, temporal orienting, neutral alerting and in the absence of cues. Behavioural data and brain imaging data from functional magnetic resonance imaging (fMRI) were collected simultaneously. We used both clonidine and guanfacine to manipulate the NA system. The reasons for this were two-fold. First, both drugs had been used by Witte and Marrocco (Witte and Marrocco, 1997a) and we wished to replicate this study as closely as possible. Secondly, and more theoretically, guanfacine (high selectivity for the α2A receptor sub-type) is thought to be more effective than clonidine in modulating cognitive function, whereas clonidine is more effective in inducing hypotension or sedation, functions associated with the α2B and α2C sub-types, respectively (Arnsten et al., 1996). We included both drugs in our study to test the hypothesis that guanfacine would have more selective effects on cognition than clonidine.

Our experiment sought to clarify the possible role of NA on directed spatial attentional orienting and to delineate the underlying neuroanatomical substrates of this effect. We expected to find an effect of clonidine on human spatial attentional orienting, particularly during invalid trials (Clark et al., 1989) and we examined whether the more selective α2 agonist guanfacine would have a similar effect. Based on previous imaging studies of spatial orienting, we predicted that modulation of right parietal cortex may underly this psychopharmacological effect (Corbetta et al., 1993; Nobre et al., 1997). Secondly, we investigated whether NA modulation of the alerting effect was associated with an effect upon selective temporal expectations, as opposed to an effect on more generic processes such as unfocused arousal or motor response preparation. By testing healthy human volunteers while they were being scanned with fMRI, we could probe the underlying neuroanatomical correlates of the NA modulation of temporal orienting and of the alerting effect. Overlap in the anatomical substrates of such modulations would suggest a shared functional substrate (Coull et al., 1999). Specifically, we looked for NA modulation of areas previously implicated in temporal orienting [e.g. left parietal and prefrontal cortices (Coull and Nobre, 1998; Coull et al., 2000)] during the alerting effect condition. This is the first time that neurochemical modulation of the neuroanatomical systems underlying attentional orienting and alerting has been imaged in the human brain.

Materials and Methods

Subjects

Ten healthy, right-handed volunteers (mean age 25.1 years, range 18–45; five male, five female) took part in the experiment. Subjects were physically fit and none were taking medication. The study was approved by the local hospital ethics committee and written informed consent was obtained prior to the study.

Experimental Task

The basic visual display consisted of a central cueing stimulus (1° eccentricity) and two peripheral boxes (7° eccentricity), inside which the target (x) appeared. The subjects' task was to covertly detect the peripheral target stimuli as rapidly as possible, while avoiding mistakes. There were four active cueing conditions that manipulated subjects' expectations of where or when target stimuli would appear within the experimental display. The spatial cue (S) predicted the target location, the temporal cue (T) predicted the target onset time, the neutral cue (X) gave no predictive information about target location or onset time but did warn the subject that a target was about to appear and the no-cue condition (O) gave neither predictive nor warning information. Figure 1 illustrates the stimuli used in the four task conditions.

The central cue was a compound stimulus consisting of a diamond and two concentric circles. For spatial or temporal orienting trials, part of the cue brightened (for 100 ms) to provide the subject with different types of predictive information. During the spatial condition, the left (or right) side of the diamond brightened to inform the subject that the target was likely to appear in the left (or right) peripheral box. This represented a validly cued trial. If the cue predicted that the target would appear on for example the left, but it actually appeared on the right, then this constituted an invalidly cued trial. During valid trials in the temporal condition, brightening of the inner circle indicated that the target would appear within a short time interval (600 ms) while brightening of the outer circle represented a longer time interval (1400 ms). However, if the cue predicted that the target would appear for example at 600 ms but it actually appeared at 1400 ms, then this constituted an invalidly cued trial. During the neutral condition the entire cue brightened, providing no specific spatial or temporal information. For no-cue trials the central stimulus remained unchanged with the first change in the stimulus display being appearance of the target.

Trials started with onset of the cue stimulus. In all trials the target appeared briefly (50 ms) at one peripheral location (left or right) after one of the two cue–target intervals (600 or 1400 ms). In the no-cue condition trial timing proceeded as if a cue had been presented at trial onset. Following the presentation of the target, a randomized delay of 1200, 1400 or 1600 ms was included before the beginning of the next trial. These delays occurred in equal proportion across all trial types. Therefore, the interval between onsets of successive trials was randomized between 1950 and 3150 ms, depending on the length of both the cue–target interval and the inter-trial interval. Subjects indicated covert detection of the target by pressing a response button with their right index finger. The computer recorded reaction times to target stimuli.

Prior to scanning subjects practiced each of the four experimental conditions in order to familiarize themselves with the tasks and cue–target contingencies. During this training period the proportion of trials with valid cues in the focused attention conditions (S or T) was 80% and the proportion of trials with invalid cues was 20%. During scanning 100 trials were presented for each of the four experimental conditions (400 trials in total). The proportion valid:invalid trials during the focused attention conditions (S or T) was 88:8. The remaining 4% of the trials were catch trials, in which no targets appeared, in order to prevent subjects from responding automatically, without attending to the target stimulus.

The high proportion of valid trials emphasized measures of the neural correlates of spatial and temporal orienting during brain imaging. Our previous neuroimaging findings have shown that invalid trials introduce additional brain activations related to emotion and/or response inhibition (Nobre et al., 1999). Higher proportions of valid:invalid trials have already been employed in brain imaging studies (Corbetta et al., 1993; Nobre et al., 1997), leading to reliable activation of brain areas and behavioural effects.

The four experimental conditions were presented successively, in counterbalanced order and in alternation with a baseline condition (B). During the baseline condition subjects fixated a static display of the central cue stimulus. The visual fixation condition provided a common baseline level of brain activity between presentations of the active tasks.

Drug Treatment

Each subject was tested on three experimental sessions, each separated by at least 1 week. Subjects received (orally) either placebo, 200 μg clonidine or 1 mg guanfacine in each of the sessions, in a within-subjects, counterbalanced, double blind design. Two tablets were given in each case. After administration of the experimental treatment subjects waited for 90 min before being tested. All sessions were identical apart from the specific treatment being given and the order of task conditions. Clonidine has well-established anti-hypertensive properties and so blood pressure was monitored for subject safety. Measurements were taken just prior to ingestion of the tablets (t = 0), 90 min post-ingestion (t = 90) and once more at the end of the testing session (t 130). Several studies have shown that α2 agents have parallel effects on blood pressure and cognitive performance (Coull et al., 1996; Jäkälä et al., 1999a,b) and so blood pressure may additionally provide an approximate index of central NA activity (Langer et al., 1980; Beckett and Finch, 1982).

fMRI Scanning

Scans were acquired using a 2 Tesla Magnetom VISION (Siemens, Erlangen, Germany) whole body MRI system, which was equipped with a head coil. Echo-planar imaging (EPI) was used to obtain T2*-weighted fMRI images in the axial plane. The acquired image volume consisted of 32 × 3 mm thick slices, which covered the entire cortex, apart from the most ventral parts of the temporal lobe. This sequence resulted in an inter-scan interval (TR) of 3.25 s.

During a single session experimental conditions were presented 10 times each in a counterbalanced manner (both within and between subjects), in alternation with the baseline condition. A blocked design was chosen in favour of an event-related approach for two main reasons. It minimized the testing time required in the scanner, since drug action had to be optimized, and increased the signal-to-noise ratio in the activation images, allowing us to test for pharmacological interactions with task performance. Each repetition of active task conditions lasted for 10 trials and eight scans (26 s). Invalid and catch trials were dispersed pseudo-randomly throughout the session. Each baseline condition lasted for six scans (19.5 s). The total number of images was 548, with 80 images for each of the four experimental conditions and 228 images in total for the intervening baseline condition. Images were acquired continuously during task performance. One break was introduced midway through the session to avoid subject fatigue and because of limitations in scanning capacity. Total scanning time was ~30 min. A structural MRI image was also acquired (using a standard T1-weighted scanning sequence) to allow anatomically specific localization of significant areas of brain activation.

Data Analysis

Blood Pressure (BP) and Behavioural Reaction Times (RTs)

Repeated-measures ANOVAs were used to test the effects of the NA agonists on blood pressure and behavioural performance. Systolic and diastolic BP were analysed separately using 2-way (3 × 3) repeated- measures ANOVAs, with drug treatment (placebo, clonidine or guanfacine) and time (pre-drug, post-drug or post-testing) as within- subject factors. Behavioural analyses tested several effects within attentional orienting and alerting. RTs faster than 50 ms or slower than 1000 ms were excluded. The remaining RTs were log-transformed in order to minimize the positively skewed nature of the data.

Attentional Orienting

Separate behavioural analyses tested the effects of drugs on the benefit of valid cues, on the cost of invalid cues and on the validity effect. Drug effects on benefits of spatially or temporally valid cues were assessed in two 2-way (2 × 3) ANOVAs, one for spatial and one for temporal conditions. These ANOVAs tested for cue type [valid (V) or neutral (X)] and drug treatment (placebo, clonidine or guanfacine). Drug effects on the cost of spatially or temporally invalid cues were assessed in two 2-way (2 × 3) ANOVAs, one for spatial and one for temporal conditions. These ANOVAs tested for cue type [invalid (I) or neutral (X)] and drug treatment (placebo, clonidine or guanfacine). The effects of the drugs on the validity effect in spatial or temporal orienting were assessed in two independent 3-way (3 × 2 × 2) ANOVAs. These ANOVAs tested for the effect of drug treatment (placebo, clonidine or guanfacine), cue type [valid (V) or invalid (I)] and side of target presentation (left or right) for the spatial condition or cue–target interval (short or long) for the temporal condition.

Alerting Effect

The effect of clonidine and guanfacine on the benefit that a warning cue has on RT (the ‘alerting effect’) was assessed in a 2-way (2 × 3) ANOVA which tested for the effect of cue type [neutral (X) or no cue (O)] and drug treatment (placebo, clonidine or guanfacine).

fMRI Data

All functional images for each subject were realigned to the first image, in order to correct for head movement between scans. The structural MRI was then co-registered to the functional images, in order to put both functional and structural images into the same space. All images were then spatially normalized into a standard space (Talairach and Tournoux, 1988), by matching each image to a standardized MNI template (Montreal Neurological Institute) using both linear and non-linear 3-dimensional transformations as implemented in SPM99 (Friston et al., 1995a). Functional images were smoothed to accommodate inter-subject differences in anatomy, using isotropic Gaussian kernels of 8 mm. Functional images were analysed using statistical parametric mapping (Friston et al., 1991) (SPM99, http://www.fil.ion.ucl.ac.uk/spm, Wellcome Department of Cognitive Neurology, London, UK). Condition and subject effects were estimated according to the general linear model at each voxel in brain space (Friston et al., 1995b). Images were adjusted for both global intensity, using proportional scaling, and for low frequency physiological drifts, using a high pass filter of 462 s (twice the length of one full testing cycle). A Gaussian temporal smoothing kernel of 4 s (at full-width half maximum) was applied to the data to minimize temporal autocorrelations.

Linear contrasts were used to test hypotheses about regionally specific condition effects, which produced statistical parametric maps of the t statistic generated for each voxel (SPM{t}). The SPM{t} was transformed to a map of corresponding Z values, thresholded at a Z value of 3.09 (P = 0.001 uncorrected for multiple comparisons), and the resulting foci were characterized in terms of both spatial extent and peak height. The significance of each region was estimated using distributional approximations from the theory of Gaussian fields. Two independent statistical analyses were conducted. These were aimed at identifying regions that were selectively involved in each type of attentional cueing (first analysis) and how these task-induced activations were altered by administration of the drugs (second analysis).

Task-related Activations

For the first analysis scans from the placebo session only were entered into a fixed effects analysis in order to examine the extent to which the current study replicated the results of our previous study using a similar paradigm (Coull and Nobre, 1998). Brain regions selectively involved in spatial and temporal orienting of attention were measured by contrasting the spatial (S) or temporal (T) orienting condition to the neutral cue (X) condition. We interrogated only those brain regions which were also present in the comparison of the appropriate orienting task to baseline [i.e. the S/T – B contrast (P < 0.05) was used to mask the entire brain volume]. We used a significance threshold corrected for multiple comparisons for all other regions and uncorrected for multiple comparisons for regions previously implicated in spatial and temporal orienting (Coull and Nobre, 1998; Nobre et al., 1999). The boundaries for hypothesized regions were defined based upon our previous findings for spatial and temporal orienting using a blocked design fMRI analysis (Coull and Nobre, 1998). A margin of 6 mm (half the FWHM) was added to the locations of the previously reported activations to define these boundaries. For example, previously described regions included the posterior parietal cortex in the right (x = 30 to 54, y = –30 to –60, z = 15 to 54 mm) and left (x = –15 to –48, y = –24 to –78, z = 30 to 51) hemispheres. In addition, we included a region of interest centred over the orbitofrontal cortex based upon a previous PET study (Nobre et al., 1999) to account for the inclusion of invalid trials in the present study (x = ±14 to ±50, y = 40 to 62, z = –32 to 6 mm). A margin of 8 mm (half the FWHM from the PET study) defined this boundary.

Brain regions selectively involved in the alerting effect were obtained by contrasting the neutral cue (X) condition with the no-cue (O) condition. We interrogated only those regions which were also present in the comparison of neutral trials to baseline [i.e. the X – B contrast (P < 0.05) was used to mask the entire brain volume]. The alerting effect has been less well characterized both in terms of the contributing brain areas and its functional components. We adopted a liberal strategy and used an uncorrected significance threshold for brain areas previously implicated in temporal orienting, arousal/alertness and motor preparation. This enabled us to test the potential contribution of these functional components with a higher sensitivity. Hypothesized areas included the posterior parietal cortex, thalamus, premotor areas and basal ganglia and were defined according to previously published results (Deiber et al., 1996; Kinomura et al., 1996; Coull and Nobre, 1998; Jueptner and Weiller, 1998; Portas et al., 1998; Sturm et al., 1999; Coull et al., 2000).

NA Modulation of Task Activations

In the second analysis we examined the interaction between drug and task condition by estimating the effects of clonidine or guanfacine on brain regions involved in all four active conditions compared with baseline. Scans from the three drug treatment sessions were subjected to a random effects analysis, in order to compare task-induced changes in activity between treatment groups. We performed 10 separate single subject analyses in which the linear contrasts of interest were calculated in each subject for placebo (pla), clonidine (clo) and guanfacine (gua) sessions separately. Therefore, analysis of each subject's data set would result in the following 12 contrast images: (Spla – Bpla), (Sclo – Bclo), (Sgua – Bgua), (Tpla – Bpla), (Tclo – Bclo), (Tgua – Bgua), (Xpla – Bpla), (Xclo – Bclo), (Xgua – Bgua), (Opla – Bpla), (Oclo – Bclo) and (Ogua – Bgua). Pairs of contrast images (e.g. Spla – Bpla and Sclo – Bclo) were then compared in a paired t-test, thereby effecting a random effects model for the drug by task interactions. In order to isolate drug modulation of directed orienting or alerting, rather than of sensori-motor processes, we looked for drug by task interactions only in those brain areas which were also associated with orienting (S or T compared with X) or alerting (X compared with O) under placebo conditions. In other words, the relevant contrast from the fixed effects placebo analysis (P < 0.05, uncorrected for multiple comparisons) was used to mask the entire brain volume when examining drug effects. This masking procedure also had the advantage of defining a priori a set of regions within which to interrogate the data for drug by task interactions. This allowed us to adopt a significance threshold uncorrected for multiple comparisons.

Results

Blood Pressure

Systolic BP

A significant main effect of time of measurement [F(2,18) = 34.68, P < 0.0001] was qualified by a significant interaction between drug treatment and time of measurement [F(4,36) = 3.30, P < 0.05]. To remove any group differences in pre-drug BP, post hoc analyses considered the change in BP from baseline (t = 0) to pre-scanning (t = 90) and from baseline to post-scanning (t = 130). Compared to placebo, clonidine decreased systolic BP pre-scanning [t(9) = 2.65, P < 0.05] and also tended to lower systolic BP post-scanning [t(9) = 2.03, P < 0.08]. These data suggest that this dose of clonidine attenuated NA activity throughout the scanning session (Fig. 2). Guanfacine had no effect on systolic BP either pre- or post-scanning [t(9) = 0.72, NS; t(9) = 1.47, NS].

Diastolic BP

A significant main effect of time of measurement [F(2,18) = 34.68, P < 0.0001] and a trend towards significance for drug [F(2,18) = 3.08, P < 0.1] was qualified by a significant interaction between drug treatment and time of measurement [F(4,36) = 3.54, P < 0.05]. Post hoc analyses of the change in BP from baseline to pre- and post-scanning showed that clonidine had no effect on diastolic BP, either pre- or post-scanning [t(9) = 0.21, NS; t(9) = 0.001, NS], as compared with placebo. However, guanfacine increased diastolic BP pre-scanning [t(9) = –2.61, P < 0.05], but was no different to placebo post-scanning [t(9) = 0.001, NS].

Behavioural Data (RTs)

Alerting Effect

RTs in the neutral cue (X) condition were significantly faster than those in the no-cue (O) condition [F(1,9) = 52.32, P < 0.0001] (see Table 1). A significant main effect of drug [F(2,18) = 5.62, P < 0.05] was qualified by the interaction between cue type and drug, which tended towards significance [F(2,18) = 3.30, P = 0.06].1 This interaction showed that clonidine [F(1,9) = 9.23, P < 0.05] but not guanfacine [F(1,9) = 3.28, NS] slowed RTs for neutral trials more than no-cue trials, as compared with placebo (Fig. 3). This represents a clonidine-induced reduction in the alerting effect via a reduction in the benefit that a warning cue normally confers on RT.

Temporal Orienting

Benefit of Valid Cues.

RTs for validly cued (V) trials were significantly faster than those in which a neutral (temporally uninformative) cue (X) was used [F(1,9) = 24.76, P = 0.0001] (Table 1). Although there was a main effect of drug [F(2,18) = 8.10, P < 0.005], such that clonidine [F(1,9) = 13.70, P = 0.05] but not guanfacine [F(1,9) = 1.57, NS] slowed RTs compared with placebo, there was no significant interaction between drug and cue type [F(2,18) = 2.51, NS]. In other words, the drugs had no effect on the benefit of being validly cued to a particular target onset time.

Cost of Invalid Cue.

There was no significant difference between invalidly cued (I) trials and neutral (X) trials [F(1,10) = 0.01, NS] (Table 1). Although there was a main effect of drug [F(2,18) = 7.21, P = 0.005], such that clonidine [F(1,9) = 14.85, P < 0.05] and guanfacine [F(1,9) = 5.17, P < 0.05] significantly slowed RTs compared with placebo, there was no significant interaction between drug and cue type [F(2,18) = 0.13, NS]. Therefore, neither clonidine nor guanfacine had an effect on the cost of being invalidly cued to a particular target onset time.

Validity Effect.

RTs for validly cued (V) trials were significantly faster than invalidly cued (I) trials [F(1,7) = 17.88, P < 0.005] (see Table 1). A significant 2-way interaction between validity of cue type and cue–target interval [F(1,7) = 11.73, P < 0.05] was qualified by a significant 3-way interaction between validity of cue type, cue–target interval and drug treatment [F(2,14) = 4.86, P < 0.05]. Figure 4a shows that clonidine slowed RTs during trials in which subjects were cued to expect a short cue–target interval but actually were presented with a long one. Post hoc analysis of the validity effect (invalid – valid) for short and long cue–target intervals confirmed that being invalidly cued to a long target interval is less costly than being invalidly cued to a short interval in placebo subjects [t(9) = 2.31, P < 0.05] (Coull and Nobre, 1998; Coull et al., 2000). However, following administration of clonidine, these two types of invalid trial are equally costly [t(9) = 0.68, NS].

Spatial Orienting

Benefit of Valid Cues.

RTs for validly cued (V) trials were generally faster than those in which a neutral (spatially uninformative) cue (X) was used [F(1,9) = 56.41, P < 0.0001]. Although there was a main effect of drug [F(2,18) = 9.25, P < 0.005], such that clonidine [F(1,9) = 12.9, P < 0.01] but not guanfacine [F(1,9) = 3.0, NS] slowed RTs compared to placebo, there was no significant interaction between drug and cue type [F(2,18) = 2.20, NS]. In other words, the drugs had no effect on the benefit of being validly cued to a particular spatial location.

Cost of Invalid Cues.

RTs for invalidly cued (I) trials were generally slower than those in which a neutral (spatially uninformative) cue (X) was used [F(1,9) = 13.31, P = 0.005] (Table 1). However, there were no significant main effects of drug nor interaction between drug and cue type. Therefore neither clonidine nor guanfacine had an effect on the cost of being invalidly cued to a particular spatial location.

Validity Effect.

RTs for validly cued (V) trials were generally faster than invalidly cued (I) trials [F(1,8) = 54.18, P < 0.0001] (Table 1). A significant interaction between drug and side of target presentation [F(2,16) = 4.17, P < 0.05] was also qualified by a significant 3-way interaction between drug, side of presentation and cue validity [F(2,16) = 4.32, P < 0.05]. Figure 4b shows that clonidine speeded RTs for invalid trials in which subjects were cued to the right, but the target actually appeared on the left. Post hoc analysis of the validity effect (invalid – valid) for left and right visual fields confirmed that clonidine attenuated the cost of invalidly cued targets presented in the left visual field [t(8) = –2.99, P < 0.05] but had no effect on invalid targets presented in the right visual field [t(9) = 0.46, NS].

fMRI Data

Alerting Effect

Placebo Only.

The brain areas involved in the alerting effect were captured by the comparison of trials with a neutral cue (X) to those with no cue (O). The contrast revealed significant activation of left insula, left dorsal premotor cortex, supplementary motor area, left inferior and superior parietal cortex and right visual cortex during performance of the neutral cue condition (Fig. 5 and Table 2).

Clonidine versus Placebo.

There were no significant clonidine-induced changes in activity of areas associated with the alerting effect. However, we noted drug-induced attenuation of task-associated activity [(X – B)pla – clo masked by (X – O)pla] in the left temporo-parietal junction (BA39) at a sub-threshold level (Fig. 6a and Table 3). Inspection of Figures 5c and 6a suggests that this focus falls within the inferior parietal cortex activation noted during analysis of the placebo scans (see above).

Guanfacine versus Placebo.

There were no significant guanfacine-induced changes in activity of areas associated with the alerting effect.

Temporal Orienting

Placebo Only.

Brain areas selectively involved in focused temporal orienting were isolated by the comparison of trials with a temporally informative cue (T) to those with a neutral cue (X). This contrast revealed significant activation of dorsolateral prefrontal cortex bilaterally, orbitofrontal cortex bilaterally, left intra-parietal sulcus, left superior parietal cortex, right visual cortex and left globus pallidus (Table 2).

Clonidine versus Placebo.

There were no significant clonidine-induced augmentations of activity of areas associated with temporal orienting. However, administration of clonidine attenuated activity associated with performance of the temporal orienting task [(T – B)pla – clo masked by (T – X)pla] in left dorsolateral prefrontal cortex, left ventral premotor cortex and left insula (Fig. 6b and Table 3).

Guanfacine versus Placebo.

There were no significant guanfacine-induced changes in activity of areas associated with temporal orienting.

Spatial Orienting

Placebo Only.

Brain areas involved in focused spatial orienting were defined by the comparison of trials with a spatially informative cue (S) to those with a neutral cue (X). The contrasts revealed significant activation of bilateral intra-parietal sulcus and superior parietal cortex (BA 7) during performance of the spatial orienting task. Bilateral visual cortex, right ventrolateral prefrontal cortex and right middle temporal cortex (BA 37) were also activated (Table 2).

Clonidine versus Placebo.

There were no significant clonidine-induced augmentations of activity in areas associated with spatial orienting. However, administration of clonidine attenuated activity associated with performance of the spatial orienting task [(S – B)pla – clo masked by (S – X)pla] in right superior parietal cortex (Fig. 6c and Table 3).

Guanfacine versus Placebo.

There were no significant guanfacine-induced changes in activity of areas associated with spatial orienting.

Discussion

We investigated whether manipulations of the NA system in humans differentially affected behavioural and neuroanatomical measures of alertness and focal spatial and temporal attentional orienting. Firstly, the results replicated our previously published findings of differential involvement of right and left parietal cortex in spatial and temporal orienting, respectively, and showed that the alerting effect activates a left-lateralized fronto-parietal network of areas. Secondly, a 200 μg dose of clonidine, an α2 NA agonist, produced selective effects on behavioural and neuroanatomical indices of spatial or temporal attentional orienting and the alerting effect. In contrast, a 1 mg dose of guanfacine, a selective α2 agonist, had no effect on either reaction times or regional brain activity. These results point to modulation of distinct aspects of human attentional orienting and alerting by clonidine, which are differentially mediated by discrete brain structures. The selective nature of both the behavioural and neuroanatomical effects argues against a simple sedative explanation of the data.

Modulation of the Alerting Effect by Clonidine

Clonidine has previously been shown to impair the alerting effect in monkeys (Witte and Marrocco, 1997a). We confirm this result in healthy human volunteers. The alerting effect was measured by comparing neutral cue to no-cue trials and clonidine was found to reduce the benefit that a neutral warning cue confers to RT performance by lengthening neutral trial RTs. There was no effect of the drug on no-cue trials. In general, trials in which a warning cue precedes a target are thought to index phasic alertness, while trials without a warning cue index tonic (or ‘intrinsic’) alertness (Posner, 1978; Sturm et al., 1999). Therefore, our data suggest that clonidine has a greater effect on phasic, rather than tonic, indices of alertness. However, the alerting effect may also tap into more specific cognitive processes. For example, the presentation of a neutral cue allows subjects to orient attention towards an expected time point and also to prepare a motor response in advance. In fact, Marrocco and co-workers have already suggested that the effect of clonidine on the alerting effect could represent an effect on temporal processes involved in sensori-motor readiness, rather than phasic alertness per se (Marrocco et al., 1994). Examination of our functional neuroimaging data can help identify the cognitive processes likely to be involved in the alerting effect. Comparison of neutral to no-cue trials in the placebo session activated a predominantly left-lateralized network of areas, including frontal and parietal cortices and SMA. These are regions which have been implicated in temporal orienting (Coull and Nobre, 1998; Coull et al., 2000) and motor preparation (Passingham, 1993; Deiber et al., 1996; Rushworth et al., 1997). There was no activation of thalamus or right-sided fronto-parietal cortices, regions that have been heavily implicated in alertness and arousal (Kinomura et al., 1996; Paus et al., 1997; Portas et al., 1998; Robertson et al., 1998; Sturm et al., 1999). This suggests that the alerting effect primarily indexes temporal orienting and motor preparation, rather than arousal or phasic alertness.

Inspection of drug by task interactions showed that clonidine had very little effect on task-associated regional activity, with only modest attenuation of activity in left superior temporal/ inferior parietal cortex (BA 39). A region slightly more dorsal to this one, in left BA 40, has been implicated in temporal orienting of attention (Coull et al., 2000). However, in the present study clonidine-induced impairment of temporal orienting performance was parallelled by attenuation of activity in left frontal areas, rather than left parietal or temporal cortices. Therefore, the lack of overlap between the effects of clonidine on alerting (left temporo-parietal region) versus temporal orienting (left frontal cortex) suggests that modulation of the alerting effect by clonidine is unlikely to be due to an underlying effect on temporal orienting processes.

Modulation of Temporal Attentional Orienting by Clonidine

Clonidine significantly impaired temporal attentional orienting by slowing RTs to invalidly cued trials in which the target appeared at an unexpectedly delayed time point. There were no effects of the drug during validly cued trials or invalid trials in which the target appeared sooner than expected. We have previously observed that the cost of being invalidly cued is significantly smaller for trials in which the target is unexpectedly delayed compared with those in which it was unexpectedly premature (Coull and Nobre, 1998; Miniussi et al., 1999; Nobre et al., 1999; Coull et al., 2000). Clonidine, however, renders both cue types equally disruptive. A specific comparison of unexpectedly delayed targets to unexpectedly premature targets has been shown to activate right prefrontal cortex in an event- related fMRI study (Coull et al., 2000). We have suggested that this represents a voluntary reorientation (or ‘endogenous shifting’) of attention to the delayed time point and may be subserved by any number of processes, including working memory and voluntary control of attention. Clonidine has previously been shown to affect both working memory and executive processes (Coull et al., 1995a; Arnsten, 1998) and so the selective effect of clonidine on unexpectedly delayed targets could represent modulation of either of these processes.

Resolving the cognitive mechanism underpinning the effects of clonidine on delayed targets may by aided by examining the functional imaging results. Task-associated activity was attenuated by clonidine in left insula, left prefrontal cortex and left ventral premotor cortex. These areas have been shown previously, and in the current study, to be involved in temporal orienting of attention (Coull and Nobre, 1998; Coull et al., 2000). Using these observations to guide our interpretation of the behavioural data, it appears that the effect of clonidine on unexpectedly delayed targets involves impairment of temporal orienting which, of course, would be necessary for reorienting to a later time point. The lack of effect on right prefrontal cortex does not mean that we must exclude the possibility that clonidine impairs endogenous shifting of attention, as suggested earlier. We were unable to isolate the neural correlates of unexpectedly delayed targets (which index endogenous shifts) specifically, since we have used a block design fMRI study in which both types of temporally invalid trial type are mixed with validly cued trials during a scanning epoch. Furthermore, the number of unexpectedly delayed targets per epoch is relatively small, which may make it difficult to show NA modulation of areas selectively engaged by this particular trial type.

Modulation of Spatial Attentional Orienting by Clonidine

We also sought to clarify whether clonidine impairs spatial attentional orienting. Clark and co-workers reported that clonidine attenuated the cost of invalid spatial cues in humans (Clark et al., 1989), whereas Witte and Marrocco found no effect of clonidine on any aspect of spatial orienting in monkeys (Witte and Marrocco, 1997a). The latter suggested that the comparison of invalid trials to neutral, rather than valid, trials in the Clark et al. study (Clark et al., 1989) may have biased their interpretation. Since clonidine has been shown to slow RTs during neutral trials, this could give the effect of speeded invalid trial RTs. However, comparing invalid with valid trials directly, as did Witte and Marrocco (Witte and Marrocco, 1997a), we have partially replicated the results of Clark and co-workers (Clark et al., 1989) using central cues in healthy human volunteers. Clonidine speeded RTs to invalidly cued targets, but only those presented in the left, not right, visual field.

A larger validity effect (invalid – valid) was observed in the left visual field (LVF) as compared with the right (RVF) following administration of placebo. This bias has been previously observed in healthy volunteers and has been suggested to represent a right hemisphere advantage for spatially valid cues (Mangun et al., 1994). The selective effect of clonidine on LVF targets is consistent with the clinical neglect literature. This suggests that while both right and left hemispheres mediate shifts of attention to the RVF, only the right hemisphere is recruited for attentional shifts to the LVF (Heilman et al., 1980; Mesulam, 1981; Weintraub and Mesulam, 1987). As such, stimuli in the LVF are more sensitive to disruption. The effect of clonidine on such an explicitly spatial aspect of task performance is mirrored by clonidine-induced attenuation of activity in right superior parietal cortex during spatial orienting, an area which has been heavily implicated in spatial attentional orienting (Corbetta et al., 1993; Nobre et al., 1997). This suggests that the modulation of task performance by clonidine is due to an effect of the drug specifically on spatial attentional orienting processes, rather than to non-specific sensori-motor processes or arousal.

A similarly asymmetrical impairment has been noted in studies of children with attention deficit disorder (ADD), who are generally slower to respond to targets presented in the LVF (Swanson et al., 1991; Nigg et al., 1997; McDonald et al., 1999), particularly if they have been invalidly cued (Swanson et al., 1991). More pertinently, this asymmetry can be attenuated, or even reversed, by administration of the catecholaminergic agonist methylphenidate (Swanson et al., 1991; Nigg et al., 1997). In effect therefore, the administration of methylphenidate to these children speeded RTs to invalid targets in the LVF. This mirrors our own result, suggesting that these effects in ADD children may be mediated more by noradrenergic than dopaminergic mechanisms.

The null effect of clonidine on spatial orienting in monkeys (Witte and Marrocco, 1997) may be due to species differences or, alternatively, to their use of peripheral (or ‘exogenous’), rather than central (or ‘endogenous’), cues (Robbins and Everitt, 1995; Fernandez-Duque and Posner, 1997; Witte and Marrocco, 1997). However, recent neuroimaging studies have suggested that there is little difference in exogenous versus endogenous attentional cueing in humans (Nobre et al., 1997; Kim et al., 1999; Rosen et al., 1999). Another possibility is that dose-dependent, or species- dependent, effects on pre- and post-synaptic receptors could differentially alter levels of NA in each of the study types, resulting in behavioural discrepancies. The Witte and Marrocco study used far lower doses of clonidine in monkeys than were used with human volunteers in the present study. Our replication of the results of Clark and co-workers (Clark et al., 1989) in humans suggests that species and or dose differences may be the largest contributing factor to these anomalies.

Finally, it is notable that while clonidine slowed RTs overall, it actually speeded RTs during invalidly cued spatial orienting trials. However, similar results have already been observed. Clark and co-workers also found clonidine-induced speeding of invalidly cued trials (Clark et al., 1989), while Witte and Marrocco found clonidine-induced slowing of RTs during neutral compared to no-cue trials (Witte and Marrocco, 1997). More compellingly, within the same subjects and with the same drug dose, Coull et al. found clonidine-induced impairment of an attentional RT task with concurrent improvement of a working memory task (Coull et al., 1995a,b). The reverse pattern was seen with the α2 antagonist idazoxan in a group of patients with frontal dementia (Coull et al., 1996). These results suggests that a single dose of an α2 agent can have opposing effects depending on the underlying pattern of cognitive activation and the brain structures being recruited for task performance. For example, tasks relying on frontal cortex activation may show improvements with a particular dose of clondine, while tasks dependant upon parietal cortex functioning show impairments with the same dose (Arnsten, 2000). Studies which are designed explicitly to assess these possibilities are needed before firm conclusions can be drawn.

Guanfacine (1 mg) has no Behavioural or Neuroanatomical Effects

We observed no significant effects of guanfacine on any aspect of performance or regional brain activity. Since we tested only one dose of guanfacine we cannot be certain whether our data represent a lack of effect of guanfacine on orienting or the alerting effect or merely represents the use of a dose insufficient to alter NA activity. Previous investigations have shown selective effects of guanfacine on tasks related to frontal lobe function, such as the delayed response task in monkeys (Arnsten and Contant, 1992) or the Tower of London planning task in humans (Jäkälä et al., 1999a), but no effect on a choice reaction time attentional task in humans, performance of which was significantly impaired by clonidine (Jäkälä et al., 1999b). Our own data may represent a similar dissociation in the effects of clonidine, but not guanfacine, on attentional tasks and, ultimately, may reflect differential roles for α2 receptor sub-types on attentional orienting or even attention in general. Guanfacine and clonidine show high affinities for α2A and α2B receptor sub-types, respectively, and α2A receptors are densely located in prefrontal cortex (Aoki et al., 1994; Arnsten et al., 1996). Therefore, in order to see effects of guanfacine on either behaviour or regional brain activity, it may be necessary to use tasks which tax prefrontal function more selectively (e.g. working memory or executive function tasks), rather than using orienting tasks that primarily activate parietal cortex. This is clearly an avenue for future research.

Another explanation for the lack of effect of guanfacine is that clonidine, but not guanfacine, has a high affinity for imidazoline receptors and so the effects that we have observed may be due to an action on imidazoline, rather than α2, receptors. Alternatively, previous studies by Jäkälä and co-workers (Jäkälä et al., 1999a,b) used a higher dose of guanfacine than has been used in the present study and so our null result may simply reflect an insufficient dose of the drug. Without appropriate control tasks or drugs or a dose–response curve it is premature to choose amongst these alternatives.

Conclusion

We have shown selective effects of clonidine on alerting and on both spatial and temporal aspects of attentional orienting performance in healthy human volunteers, whilst simultaneously measuring regional brain activity with fMRI to help elucidate the cognitive and neuroanatomical processes underlying these effects. Clonidine-induced deficits in the alerting effect and temporal orienting were accompanied by attenuation of task- associated activity in distinct cortical regions. We suggest that this neuroanatomical distinction reflects differential functional effects of the drug. Therefore, NA modulation of the alerting effect is unlikely to be simply the result of an underlying effect on temporal orienting. However, clonidine does impair temporal orienting performance and attenuates task-associated activity in areas previously implicated in orientation of attention to specific time intervals. We have also shown that attenuation of right superior parietal cortex activity underlies the NA modulation of spatial attentional orienting. This suggests that the effects of the drug on this task are due to processes related to orienting to one or other side of space, rather than non-specific sensori-motor or arousal processes.

Notes

  1. As we were trying to replicate a result already obtained in monkeys (Witte and Marrocco, 1997) we accept this result as significant according to a one-tailed t-test.

This work was funded by The Wellcome Trust.

Address correspondence to J.T. Coull, Wellcome Department of Cognitive Neurology, Institute of Neurology, 12 Queen Square, London WC1N 3BG, UK. Email: j.coull@fil.ion.ucl.ac.uk.

Table 1

Mean (± SE) RTs for each cue type during all three drug treatments

 Neutral No-cue Spatial valid Spatial invalid Temporal valid Temporal invalid 
Placebo 321.45 ± 10.5 369.69 ± 11.4 288.81 ± 11.5 400.79 ± 23.75 305.70 ± 16.8 320.17 ± 15.9 
Clonidine 352.73 ± 17.2 386.02 ± 15.7 320.25 ± 18.84 399.36 ± 29.56 334.61 ± 23.7 361.13 ± 24.8 
Guanfacine 336.18 ± 14.1 377.16 ± 13.5 293.91 ± 13.2 384.78 ± 26.1 305.10 ± 15.9 327.10 ± 21.5 
 Neutral No-cue Spatial valid Spatial invalid Temporal valid Temporal invalid 
Placebo 321.45 ± 10.5 369.69 ± 11.4 288.81 ± 11.5 400.79 ± 23.75 305.70 ± 16.8 320.17 ± 15.9 
Clonidine 352.73 ± 17.2 386.02 ± 15.7 320.25 ± 18.84 399.36 ± 29.56 334.61 ± 23.7 361.13 ± 24.8 
Guanfacine 336.18 ± 14.1 377.16 ± 13.5 293.91 ± 13.2 384.78 ± 26.1 305.10 ± 15.9 327.10 ± 21.5 
Table 2

Selective areas of activation associated with the alerting effect and temporal and spatial orienting, as estimated during the placebo session only

Brain area Coordinates (x, y, zZ score 
Z scores for hypothesized regions are significant at a threshold of P < 0.001, uncorrected for multiple comparisons. Z scores indicated by * are significant at a threshold of P < 0.05, corrected for multiple comparisons. 
Alerting effect (X-O)     
 Left anterior insula –57  27 4.81* 
 Left dorsal premotor cortex (BA 6) –30  15  60 3.72 
 –45  51 3.64 
 Supplementary motor area  12  63 4.71* 
 Left superior parietal cortex (BA 7) –30 –63  54 4.31 
 Left inferior parietal cortex (BA 40) –48 –57  24 3.84 
 Right lateral visual cortex (BA 19)  51 –72  –3 4.53 
Temporal orienting (T-X)     
 Left and right dorsolateral prefrontal cortex (BA 46) –42  39  24 5.22* 
  45  45  18 4.19 
 Left and right orbitofrontal cortex (BA10/11) –36  57  –6 3.13 
  48  51 –12 3.12 
 Left intra-parietal sulcus –36 –36  66 4.67* 
 –54 –39  39 4.50 
 Left superior parietal cortex (BA 7) –18 –63  63 4.59* 
 Right lateral visual cortex (BA 19)  36 –87 5.05* 
 Left globus pallidus  –6  15  –3 3.14 
Spatial orienting (S-X)     
 Right ventrolateral prefrontal cortex (BA 47)  48  21 –12 4.76* 
 Right middle temporal cortex (BA 37)  60 –57  –9 4.70* 
 Left and right intra-parietal sulcus –33 –48  54 8.21* 
  39 –42  57 5.49* 
 Left and right superior parietal cortex –18 –63  63 5.09* 
  24 –66  60 5.11* 
 Left and right lateral visual cortex (BA 19) –45 –78 4.82* 
  36 –81 3.99 
Brain area Coordinates (x, y, zZ score 
Z scores for hypothesized regions are significant at a threshold of P < 0.001, uncorrected for multiple comparisons. Z scores indicated by * are significant at a threshold of P < 0.05, corrected for multiple comparisons. 
Alerting effect (X-O)     
 Left anterior insula –57  27 4.81* 
 Left dorsal premotor cortex (BA 6) –30  15  60 3.72 
 –45  51 3.64 
 Supplementary motor area  12  63 4.71* 
 Left superior parietal cortex (BA 7) –30 –63  54 4.31 
 Left inferior parietal cortex (BA 40) –48 –57  24 3.84 
 Right lateral visual cortex (BA 19)  51 –72  –3 4.53 
Temporal orienting (T-X)     
 Left and right dorsolateral prefrontal cortex (BA 46) –42  39  24 5.22* 
  45  45  18 4.19 
 Left and right orbitofrontal cortex (BA10/11) –36  57  –6 3.13 
  48  51 –12 3.12 
 Left intra-parietal sulcus –36 –36  66 4.67* 
 –54 –39  39 4.50 
 Left superior parietal cortex (BA 7) –18 –63  63 4.59* 
 Right lateral visual cortex (BA 19)  36 –87 5.05* 
 Left globus pallidus  –6  15  –3 3.14 
Spatial orienting (S-X)     
 Right ventrolateral prefrontal cortex (BA 47)  48  21 –12 4.76* 
 Right middle temporal cortex (BA 37)  60 –57  –9 4.70* 
 Left and right intra-parietal sulcus –33 –48  54 8.21* 
  39 –42  57 5.49* 
 Left and right superior parietal cortex –18 –63  63 5.09* 
  24 –66  60 5.11* 
 Left and right lateral visual cortex (BA 19) –45 –78 4.82* 
  36 –81 3.99 
Table 3

Clonidine-induced attenuations of task-induced regional activity, as compared with placebo, for the alerting effect and temporal and spatial orienting

Brain area Coordinates (x, y, zZ score 
Z scores for hypothesized regions are significant at a threshold of P < 0.001 (uncorrected for multiple comparisons), except *, which is significant at a threshold of P = 0.001 (uncorrected for multiple comparisons). 
Alerting effect     
 Left temporo-parietal junction (BA39) –51 –57 18 2.99* 
Temporal orienting     
 Left dorsolateral prefrontal cortex (BA 9) –39  21 30 3.44 
 Left anterior insula –51  18 –3 3.17 
 Left ventral premotor cortex (BA 6) –60  –3 24 4.16 
Spatial orienting     
 Right superior parietal cortex (BA 7)  21 –66 54 3.17 
Brain area Coordinates (x, y, zZ score 
Z scores for hypothesized regions are significant at a threshold of P < 0.001 (uncorrected for multiple comparisons), except *, which is significant at a threshold of P = 0.001 (uncorrected for multiple comparisons). 
Alerting effect     
 Left temporo-parietal junction (BA39) –51 –57 18 2.99* 
Temporal orienting     
 Left dorsolateral prefrontal cortex (BA 9) –39  21 30 3.44 
 Left anterior insula –51  18 –3 3.17 
 Left ventral premotor cortex (BA 6) –60  –3 24 4.16 
Spatial orienting     
 Right superior parietal cortex (BA 7)  21 –66 54 3.17 
Figure 1.

(a) A typical trial which, in this example, contains a neutral cue. Stimuli are white, presented on a black background. Discrete elements (according to condition) of the central cue brighten for 100 ms and act to guide subjects attention. The cue–target interval is either 600 or 1400 ms (short/long cue–target interval) and then the target appears for 50 ms in either the left or right box. The central attentional cue can be one of four types. (b) The spatial cue directs the subjects' attention to the left or right, the temporal cue directs attention to a short or long cue–target interval and the neutral cue provides neither spatial nor temporal information. In the no-cue condition, the central cue remains unchanged for the duration of the trial.

Figure 1.

(a) A typical trial which, in this example, contains a neutral cue. Stimuli are white, presented on a black background. Discrete elements (according to condition) of the central cue brighten for 100 ms and act to guide subjects attention. The cue–target interval is either 600 or 1400 ms (short/long cue–target interval) and then the target appears for 50 ms in either the left or right box. The central attentional cue can be one of four types. (b) The spatial cue directs the subjects' attention to the left or right, the temporal cue directs attention to a short or long cue–target interval and the neutral cue provides neither spatial nor temporal information. In the no-cue condition, the central cue remains unchanged for the duration of the trial.

Figure 2.

Systolic and diastolic blood pressure following administration of placebo, 200 μg clonidine or 1 mg guanfacine at three discrete time points: at the time of administration of the drug (t = 0), immediately before scanning (t = 90 min) and immediately after scanning (t = ~130 min).

Figure 2.

Systolic and diastolic blood pressure following administration of placebo, 200 μg clonidine or 1 mg guanfacine at three discrete time points: at the time of administration of the drug (t = 0), immediately before scanning (t = 90 min) and immediately after scanning (t = ~130 min).

Figure 3.

Reaction times (RT) demonstrating the alerting effect in placebo subjects (the beneficial effect that a neutral warning cue confers to RT) and the attenuation of this benefit following administration of 200 μg clonidine, but not 1 mg guanfacine.

Figure 3.

Reaction times (RT) demonstrating the alerting effect in placebo subjects (the beneficial effect that a neutral warning cue confers to RT) and the attenuation of this benefit following administration of 200 μg clonidine, but not 1 mg guanfacine.

Figure 4.

Reaction times (RT) during (a)temporal orienting and (b) spatial orienting, following administration of placebo, 200 μg clonidine or 1 mg guanfacine. RTs are shown separately for targets appearing after a 600 ms (short SOA) or 1400 ms (long SOA) cue-target interval during temporal orienting and for targets appearing in the left visual field (LVF) or right visual field (RVF) during spatial orienting. Target onset time or location were either correctly (valid) or incorrectly (invalid) predicted by the cue. The significant simple main effects is indicated by *.

Figure 4.

Reaction times (RT) during (a)temporal orienting and (b) spatial orienting, following administration of placebo, 200 μg clonidine or 1 mg guanfacine. RTs are shown separately for targets appearing after a 600 ms (short SOA) or 1400 ms (long SOA) cue-target interval during temporal orienting and for targets appearing in the left visual field (LVF) or right visual field (RVF) during spatial orienting. Target onset time or location were either correctly (valid) or incorrectly (invalid) predicted by the cue. The significant simple main effects is indicated by *.

Figure 5.

Neuroanatomical correlates of the alerting effect: brain regions significantly more activated during neutral cue than no-cue trials rendered onto (a) lateral and (b) dorsal views of a standardized template brain. Superior and inferior parietal cortex activations are shown in greater detail in (c) and are superimposed over the averaged structural MRI of all 10 subjects.

Figure 5.

Neuroanatomical correlates of the alerting effect: brain regions significantly more activated during neutral cue than no-cue trials rendered onto (a) lateral and (b) dorsal views of a standardized template brain. Superior and inferior parietal cortex activations are shown in greater detail in (c) and are superimposed over the averaged structural MRI of all 10 subjects.

Figure 6.

Clondine attenuates task-related activity in the (a) left temporo-parietal junction during the alerting effect (neutral minus no-cue), (b) left anterior insula during temporal orienting (temporal minus neutral) and (c) right superior parietal cortex during spatial orienting (spatial minus neutral). Areas of attenuated activity are shown superimposed over the averaged structural MRI of all 10 subjects. Changes in BOLD signal at each of these regions are shown in the corresponding plots. Data are shown for each of the comparisons of interest, following administration of placebo or clonidine. In all cases clonidine decreases activity more during the task of interest than during the control condition.

Figure 6.

Clondine attenuates task-related activity in the (a) left temporo-parietal junction during the alerting effect (neutral minus no-cue), (b) left anterior insula during temporal orienting (temporal minus neutral) and (c) right superior parietal cortex during spatial orienting (spatial minus neutral). Areas of attenuated activity are shown superimposed over the averaged structural MRI of all 10 subjects. Changes in BOLD signal at each of these regions are shown in the corresponding plots. Data are shown for each of the comparisons of interest, following administration of placebo or clonidine. In all cases clonidine decreases activity more during the task of interest than during the control condition.

References

Aoki C, Go C-G, Venkatesan C, Kurose H (
1994
) Perikaryal and synaptic localization of α2A adrenergic receptor-like immunoreactivity.
Brain Res
 
650
:
181
–204.
Arnsten AFT (
1998
) Catecholamine modulation of prefrontal cortical cognitive function.
Trends Cogn Sci
 
2
:
436
–447.
Arnsten AFT (
2000
) Through the looking glass: differential noradrenergic modulation of prefrontal cortical function.
Neural Plast
 
7
:
133
–146.
Arnsten AFT, Contant TA (
1992
) Alpha-2 adrenergic agonists decrease distractibility in aged monkeys performing the delayed response task.
Psychopharmacology
 
108
:
159
–169.
Arnsten AFT, Steere JC, Hunt RD (
1996
) The contribution of α2 noradrenergic mecahnisms to prefrontal cognitive function: potential significance for attention-deficit hyperactivity disorder.
Arch Gen Psychiat
 
53
:
448
–455.
Beckett PJ, Finch L (
1982
) The α1 and α2-adrenoceptor involvement in the central cardiovascular action of clonidine in the conscious renal hypertensive cat.
Eur J Pharmacol
 
82
:
155
–160.
Clark CR, Geffen GM, Geffen LB (
1989
) Catecholamines and the covert orientation of attention in humans.
Neuropsychologia
 
27
:
131
–139.
Corbetta M, Miezin FM, Shulman GL, Petersen SE (
1993
) A PET study of visuospatial attention.
J Neurosci
 
13
:
1202
–1226.
Coull JT, Nobre AC (
1998
) Where and when to pay attention: the neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI.
J Neurosci
 
18
:
7426
–7435.
Coull JT, Middleton HC, Robbins TW, Sahakian BJ (
1995
) Contrasting effects of clonidine and diazepam on tests of working memory and planning.
Psychopharmacology
 
120
:
311
–321.
Coull JT, Middleton HC, Robbins TW, Sahakian BJ (
1995
) Clonidine and diazepam have differential effects on tests of attention and learning.
Psychopharmacology
 
120
:
322
–332.
Coull JT, Hodges JR, Sahakian BJ (
1996
) The α2 antagonist idazoxan remediates certain attentional and executive dysfunction in patients with dementia of frontal type.
Psychopharmacology
 
123
:
239
–249.
Coull JT, Frith CD, Dolan RJ, Frackowiak RSJ, Grasby PM (
1997
) The neural correlates of the noradrenergic modulation of human attention, arousal and learning.
Eur J Neurosci
 
9
:
589
–598.
Coull JT, Frith CD, Dolan RJ (
1999
) Dissociating neuromodulatory effects of diazepam on episodic memory encoding and executive function.
Psychopharmacology
 
145
:
213
–222.
Coull JT, Frith CD, Büchel C, Nobre AC (
2000
) Orienting attention in time: behavioural and neuroanatomical distinction between exogenous and endogenous shifts.
Neuropsychologia
 
38
:
808
–819.
Deiber M-P, Ibañez V, Sadato N, Hallett M (
1996
) Cerebral structures participating in motor preparation in humans: a positron emission tomography study.
J Neurophysiol
 
75
:
233
–247.
Fernandez-Duque D, Posner MI (
1997
) Relating the mechanisms of orienting and alerting.
Neuropsychologia
 
35
:
477
–486.
Friston K, Frith C, Liddle P, Frackowiak R (
1991
) Comparing functional (PET) images: the assessment of significant change.
J Cereb Blood Flow Metab
 
11
:
690
–699.
Friston KJ, Ashburner J, Poline J-B, Frith CD, Heather JD, Frackowiak RSJ (
1995
) Spatial registration and normalisation of images.
Hum Brain Map
 
2
:
165
–189.
Friston KJ, Holmes AP, Worsley KJ, Poline J-P, Frith CD, Frackowiak RSJ (
1995
) Statistical parametric maps in functional imaging: a general linear approach.
Hum Brain Map
 
2
:
189
–210.
Heilman KM, Van Den Abell T (
1980
) Right hemisphere dominance for attention: the mechanism underlying hemispheric asymmetries of inattention (neglect).
Neurology
 
30
:
327
–330.
Jäkälä P, Riekkinen M, Sirviö J, Koivisto E, Kejonen K, Vanhanen M, Riekkinen P (
1999
) Guanfacine, but not clonidine, improves planning and working memory performance in humans.
Neuropsychopharmacology
 
20
:
460
–470.
Jäkälä P, Riekkinen M, Sirviö J, Koivisto E, Riekkinen P (
1999
) Clonidine, but not guanfacine, impairs choice reaction time performance in young healthy volunteers.
Neuropsychopharmacology
 
21
:
495
–502.
Jueptner M, Weiller C (
1998
) A review of differences between basal ganglia and cerebellar control of movement as revealed by functional imaging studies.
Brain
 
121
:
1437
–1449.
Kim Y-H, Gitelman DR, Nobre AC, Parrish TB, LaBar KS, Mesulam MM (
1999
) The large-scale neural network for spatial attention displays multifunctional overlap but differential asymmetry.
NeuroImage
 
9
:
269
–277.
Kinomura S, Larsson J, Gulyas B, Roland PE (
1996
) Activation by attention of the human reticular formation and thalamic intralaminar nuclei.
Science
 
271
:
512
–515.
Langer SZ, Cavero I, Massinham R (
1980
) Recent developments in noradrenergic neurotransmission and its relevance to the mechanism of action of certain antihypertensive agents.
Hypertension
 
2
:
372
–382.
McDonald S, Bennett KMB, Chambers H, Castiello U (
1999
) Covert orienting and focusing of attention in children with attention deficit hyperactivity disorder.
Neuropsychologia
 
37
:
345
–356.
Mangun GR, Luck SJ, Plager R, Loftus W, Hillyard SA, Handy T, Clark VP, Gazzaniga MS (
1994
) Monitoring the visual world—hemispheric asymmetries and subcortical processes in attention.
J Cogn Neurosci
 
6
:
267
–275.
Marrocoo RT, Witte EA, Davidson MC (
1994
) Arousal systems.
Curr Opin Neurobiol
 
4
:
166
–170.
Mesulam MM (
1981
) A cortical network for directed attention and unilateral neglect.
Arch Neurol
 
10
:
304
–325.
Miniussi C, Wilding EL, Coull JT, Nobre AC (
1999
) Orienting attention in time: modulation of brain potentials.
Brain
 
122
:
1507
–1518.
Nigg JT, Swanson JM, Hinshaw SP (
1997
) Covert visual spatial attention in boys with attention deficit hyperactivity disorder: lateral effects, methylphenidate response and results for parents.
Neuropsychologia
 
35
:
165
–176.
Nobre AC, Gitelman DR, Sebestyen GN, Meyer J, Frackowiak RSJ, Frith CD, Mesulam MM (
1997
) Functional localisation of the system for visuospatial attention using positron emission tomography.
Brain
 
120
:
515
–533.
Nobre AC, Coull JT, Frith CD, Mesulam MM (
1999
) Orbitofrontal cortex is activated during breaches of expectation in tasks of visual attention.
Nature Neurosci
 
2
:
11
–12.
Passingham R (1993) The frontal lobes and voluntary action. Oxford University Press, Oxford.
Paus T, Zatorre RJ, Gotman J, Caramanos Z, Petrides M, Evans AC (
1997
) Time-related changes in neural systems underlying attention and arousal during the performance of an auditory vigilance task.
J Cogn Neurosci
 
9
:
392
–408.
Portas CM, Rees G, Howseman AM, Josephs O, Turner R, Frith CD (
1998
) A specific role for the thalamus in mediating the interaction of attention and arousal in humans.
J Neurosci
 
18
:
8979
–8989.
Posner MI (1978) Chronometric explorations of mind. Hillsdale: Erlbaum.
Posner MI, Snyder CRR, Davidson BJ (
1980
) Attention and the detection of signals.
J Exp Psychiat Gen
 
109
:
160
–174.
Robbins TW, Everitt BJ (1995) Central norepinephrine neurons and behaviour. In: Psychopharmacology: the fourth generation of progress (Bloom FB and Kupfer DJ, eds). New York: Raven Press.
Robertson IH, Mattingley JB, Rorden C, Driver J (
1998
) Phasic alerting of neglect patients overcomes their spatial deficit in visual awareness.
Nature
 
395
:
169
–172.
Rosen AC, Rao SM, Caffarra P, Scaglioni A, Bobholz JA, Woodley SJ, Hammeke TA, Cunningham JM, Prieto TE, Binder JR (
1999
) Neural basis of endogenous and exogenous spatial orienting: a functional MRI study.
J Cogn Neurosci
 
11
:
135
–152.
Rushworth, MF, Nixon, PD, Renowden, S, Wade, DT, Passingham, RE (
1997
) The left parietal cortex and motor attention.
Neuropsychologia
 
35
:
1261
–1273.
Smith A and Nutt D (
1996
) Noradrenaline and attention lapses.
Nature
 
380
:
291
.
Sturm W, deSimone A, Krause BJ, Specht K, Hesselmann V, Radermacher I, Herzog H, Tellmann L, MullerGartner HW, Willmes K (
1999
) Functional anatomy of intrinsic alertness: evidence for a fronto-parietal-thalamic-brainstem network in the right hemisphere.
Neuropsychologia
 
37
:
797
–780.
Swanson JM, Posner MI, Potkin S, Bonforte S, Youpa D, Fiore C, Cantwell D, Crinella F (
1991
) Activating tasks for the study of visual-spatial attention in ADHD children: a cognitive anatomic approach.
J Child Neurol
 
6
:
S119
–S127.
Talairach J, Tournoux P (1988) A co-planar stereotaxic atlas of the human brain. Stuttgart: Thieme.
Weintraub S, Mesulam MM (
1987
) Right cerebral dominance in spatial attention. Further evidence based on ipsilateral neglect.
Arch Neurol
 
44
:
621
–625.
Witte EA, Marrocco RT (
1997
) Alteration of brain noradrenergic activity in rhesus monkeys affects the alerting component of covert orienting.
Psychopharmacology
 
132
:
315
–323.
Witte EA, Marrocco RT (
1997
) Effects of altering brain cholinergic activity on covert orienting of attention: comparison of monkey and human performance.
Psychopharmacology
 
132
:
324
–334.