Asymmetric Loss of Parietal Activity Causes Spatial Bias in Prodromal and Mild Alzheimer's Disease

av

bt

dfih

Asymmetric Loss of Parietal Activity Causes SpatialBias in Prodromal and Mild Alzheimer’s DiseaseChristian Sorg, Nicholas Myers, Petra Redel, Peter Bublak, Valentin Riedl, Andrei Manoliu,Robert Perneczky, Timo Grimmer, Alexander Kurz, Hans Förstl, Alexander Drzezga, Hermann J. Müller,Afra M. Wohlschläger, and Kathrin Finke

Background: In Alzheimer’s disease (AD), loss of effective neuronal activity is reflected by cortical glucose hypometabolism. Hypometab-olism in the posterior parietal cortex (PPC) is among the first in vivo signs of AD; however, its functional impact on large-scale brainmechanisms and behavior is poorly understood. The lateral PPC contributes to spatial attention constituting a basic function of the humanbrain. We hypothesized 1) that lateral PPC hypometabolism is associated with impaired spatial attention in very early AD and 2) that impairedcompetition of effective neuronal activity across hemispheres might underlie this deficit in terms of brain mechanisms.

Methods: A model-based imaging approach was applied to assess patients with prodromal (n � 28) and mild (n � 7) AD. Quantitativettention parameters, derived from performance on simple psychophysical tasks and analyzed by Bundesen’s computational theory ofisual attention, were related to brain metabolism, measured by 18F-fluorodeoxyglucose positron emission tomography.

Results: Patients’ left and right lateral PPC metabolism was reduced. Nine patients had significant spatial attentional bias on the left sideand two patients on the right. Direction and degree of spatial bias was correlated with direction and degree of an interhemisphericmetabolism bias in the inferior parietal lobe and temporoparietal junction.

Conclusions: Our data provide evidence that in very early AD, asymmetric hypometabolism of the lateral PPC causes spatial attentionalbias. Results are broadly consistent with the model that asymmetrically impaired effective neuronal PPC activity in AD biases the competi-
tion of visual objects for cortical representation and access to awareness to one side.
wbobat(aatiftffidtajesacSca

wtopnb

Key Words: Bundesen’s theory of visual attention, FDG-PET, model-ased imaging, parietal cortex, prodromal Alzheimer’s disease, spa-

ial attention

A lzheimer’s disease (AD) is the most frequent cause of age-related dementia and is characterized by progressive cogni-tive and behavioural deficits (1). AD is a neurodegenerative

isease neuropathologically characterized by amyloid-plaques andbrillary tangles (2). One of the first in vivo signs of AD is glucoseypometabolism in the posterior parietal cortex (PPC) (3,4). Be-

cause cerebral glucose metabolism is critically triggered by synap-tic activity (5– 8), PPC hypometabolism reflects at the cellular levelimpaired effective neuronal activity in AD (4). However, at both thebrain system and behavioral level, the functional relevance of AD’sPPC hypometabolism is poorly understood. The lateral PPC isstrongly involved in spatial attention (9,10). Spatial attention repre-sents a basic cognitive function of the human brain (9) and is im-paired in very early AD (11). Here we asked whether lateral PPChypometabolism is associated with impaired spatial attention invery early AD and whether impaired competition of effective neu-ronal activity across hemispheres might underlie this deficit at thebrain system level.

From the Departments of Psychiatry (CS, AM, RP, TG, AK, HF), Neuroradiol-ogy (CS, NM, VR, AM, AMW), Nuclear Medicine (CS, AD), and Neurology(VR, AMW), Klinikum rechts der Isar, Technische Universität München,Munich; Graduate School of Systemic Neurosciences (NM, VR), LudwigMaximilian University Munich, Martinsried; General and ExperimentalPsychology/Neuro-Cognitive Psychology (PR, HJM, KF), Ludwig Maximil-ian University, Munich; and Neuropsychology Unit (PB), NeurologyClinic, Friedrich Schiller University Jena, Jena, Germany.

Authors CS and NM contributed equally to this work.Address correspondence to Christian Sorg, M.D., Department of Psychiatry,

Klinikum rechts der Isar, Ismaningerstrasse 22, 81675 München, Ger-many; E-mail: [email protected].

wReceived May 9, 2011; revised Sep 19, 2011; accepted Sep 20, 2011.

0006-3223/$36.00doi:10.1016/j.biopsych.2011.09.027

To link patients’ PPC metabolism with attentional functions in aay sensitive for basic attentional subprocesses, we used a model-ased neuroimaging approach relying on the computational the-ry of visual attention (TVA) (12). TVA is based on the principle ofiased competition (13). This principle describes visual attention asmodulating (or biasing) factor on the visual processing stream

hat selects neural populations representing certain visual featuressuch as a particular part of visual space); processing in these visualreas will be more likely to reach higher stages of processing (suchs consciously perceived). Anatomically, this functional architec-ure divides the visual system into a basic visual processing streamn the occipital cortex and a control system in the parietal androntal lobes (13). Mechanistically, attentional selection is charac-erized as a set of mechanisms responsible for selecting objectsrom the visual environment for conscious perception. TVA speci-es these mechanisms mathematically, finally resulting in four in-ependent parameters (12): visual processing speed (C) and short-

erm memory capacity (K) represent general attentional capacity;ttentional weights of distractors relative to targets (�) and of ob-

ects relative to their place in the visual field (w�) represent thefficiency and spatial distribution of selective attention. In a precur-or study using the TVA framework, we found impaired spatialttentional weighting w� in patients with mild AD or amnestic mildognitive impairment (aMCI), which is a high-risk state for AD (11).patial bias was more frequently directed to the left side and in-reased during progression to AD. Here we ask whether this spatialttentional bias is related to metabolism in the lateral PPC.

We applied simple partial- and whole-report paradigms (Figure 1),hich permit the TVA-based estimation of attentional functioning

hat is not confounded by cognitive or motor slowing. We focusedn the earliest stages of AD by investigating only patients withrodromal and mild AD (i.e., very early AD [pAD], n � 28; mild AD,� 7; Table 1). Prodromal AD is defined as aMCI with at least oneiological sign for AD (14). To quantify brain metabolism in patients,
e used resting-state 18F-fluorodeoxyglucose positron emission
BIOL PSYCHIATRY 2012;71:798–804© 2012 Society of Biological Psychiatry

mailto:[email protected]

w

1DococnsFabhir(

P

tvdf�tacf

wIlapftipd

TC

NSECMDw�CK

stdgtmb

pbltrfi(

C. Sorg et al. BIOL PSYCHIATRY 2012;71:798–804 799

tomography (FDG-PET) (3). First, we expected to find reduced PPCmetabolism (compared with healthy control subjects) and spatialbias in patients as well as a correlation of this metabolic reductionwith the degree of spatial bias independent of other attentionalsubprocesses. Second, we hypothesized that an interhemisphericmetabolism bias (M�) in the PPC would correlate with spatial bias

�. This hypothesis was motivated by the observation that unilat-eral parietal lesions caused by stroke typically lead to ipsilesionalspatial bias that increases with lesion volume (15).

Methods and Materials

SubjectsThirty-five patients (20 women, age range 47–79 years) diag-

nosed with prodromal (n � 28) or mild AD (n � 7) participated in thestudy (Table 1). All participants provided informed consent in accor-dance with the Human Research Committee guidelines of the Klini-kum Rechts der Isar, Technische Universität, München. Patientswere recruited from the Memory Clinic of the Department of Psy-chiatry. Examination of every participant included medical history,neurological examination, informant interview (Clinical DementiaRating; CDR) (16), neuropsychological assessment (Consortium toEstablish a Registry for Alzheimer’s Disease battery; CERAD) (17),structural magnetic resonance imaging (MRI; within 18 months),FDG-PET, and blood tests. Patients with prodromal AD met criteriafor aMCI and demonstrated temporoparietal hypometabolism asone of six supportive biological features in the definition of prodro-mal AD (14). Amnestic MCI criteria include reported and neuropsy-chologically assessed memory impairments, largely intact activitiesof daily living, and excluded dementia (18). Patients with mild AD

Figure 1. Event train of a single trial of the partial-report paradigm. Afterresentation of a fixation cross (as a warning signal), one or two letters wereriefly presented followed by a mask (prespecified exposure duration of

etters for each individual). Patients were asked to report only the red letterargets (T) and to ignore the green letter distractors (D). From the accuracy ofeports, subsequent analysis was based on the theory of visual attention,nally resulting in task- and space-related weights of attentional selection

see Methods and Materials).

fulfilled criteria for dementia (Clinical Dementia Rating global score � p

) and the National Institute of Neurological and Communicativeisorders and Stroke and the Alzheimer’s Disease and Related Dis-rders Association (NINCIDS-ADRDA) criteria for AD (19). Exclusionriteria for entry into the study were other neurological, psychiatric,r systemic diseases (e.g., stroke, depression, alcoholism) or clini-ally remarkable MRI (e.g., stroke lesions) potentially related to cog-itive impairment. None of the patients exhibited predominant orevere visual impairments typical of posterior cortical atrophy.ourteen patients were treated for hypertension (beta-blockers,ngiotensin converting enzyme inhibitors, and calcium channellockers) and eight for hypercholesterolemia (statins). One patientad diabetes mellitus, eight patients received antidepressive med-

cation (mirtazapine, escitalopram), and four patients with mild ADeceived cholinesterase inhibitors. Behavioral data of 19 patients15 aMCI and 4 AD) have been presented previously (11).

artial-Report Task and Theory of Visual AttentionWe used TVA and partial- and whole-report (PR, WR) paradigms

o assess individuals’ attentional performance (Figure 1). TVA pro-ides a computational framework to analyze PR and WR behavioralata (11,20) (see also Supplement 1). Data analysis results in the

ollowing parameters: spatial attentional bias w�, top-down control, visual short-term memory storage capacity K, and visual percep-

ual processing speed C. W� represented the parameter of interest,nd �, K, and C were used to control for other attentional subpro-esses than spatial attentional selection (21). All participants per-ormed the PR and WR tasks within 60 days after FDG-PET.

PR and WR tasks consist of briefly presented arrays of letters,hich are to be reported verbally (11, 20) (see also Supplement 1).

n the WR task, columns of five letters were presented either on theeft or the right side with varying presentation time. By analyzingccuracy scores as a function of effective presentation time, TVAermits close estimation of K and C by fitting an exponential growth

unction to accuracy scores of each individual participant (20). Inhe PR task, stimuli were presented at corner positions of an imag-nary square (Figure 1). On each trial, a single red target, a red targetlus a green distractor, or two red targets were presented for pre-etermined exposure duration. The individual performance accu-

able 1. Demographic, Clinical, and Behavioral Data of Patients andontrol Subjects

Very Early AD HC-PET HC-TVA

35 23 36ex (female/male) 20/15 12/11 20/16ducation (years) 10.5 (1.8) 10.0 (2.1) 10.5 (2.1)DR Sum Score 2.5 (1.2) 0 0MSE 26.2 (2.3)a/a 29.6 (1.1) 29.0 (1.0)elayed Recall 3.0 (1.8)a 6.4 (2.0) —

� .54 (.12)a — .48 (.06).56 (.52)a — .32 (.17)

15.3 (6.3)a — 19.1 (5.7)2.86 (.47) — 2.95 (.34)

SD in parentheses.C, processing speed; CDR, Clinical Dementia Rating; Delayed Recall:

core on the verbal delayed recall subtest of the battery of the Consortiumo Establish a Registry for Alzheimer’s Disease (CERAD); FDG-PET, 18F-fluoro-eoxyglucose positron emission tomography; HC-PET, healthy controlroup that provided the 18FDG-PET data; HC-TVA, healthy control group

hat participated in the theory of visual attention paradigm; K, short-termemory capacity; MMSE, Mini-Mental State Examination; w� and �: distri-

ution of attentional weights across the hemifields and task, respectively.a
Scores that differ between patient and control group, two-sample t test
� .05.

www.sobp.org/journal

wt

avPa

ndistAtTafiLSto

rlboa

B

tpcMrccbpcrmtma(

tvmt

R

H

pIpcpi

S

bhiat9pl.3Tsp

SV

mcriitrSl�g

800 BIOL PSYCHIATRY 2012;71:798–804 C. Sorg et al.

racy across conditions (one or two targets, or one target and onedistractor, separately for each visual hemifield) was modeled by amaximum-likelihood algorithm, yielding estimates of attentionalweights w and � (11). The relative spatial bias w� was computed as

� � wleft/(wleft � wright), with w� ranging from 0 (total neglect forhe left hemifield) to 1 (total neglect for the right hemifield).

To evaluate patients’ spatial bias, the distribution of w� wasestimated for an age, sex, and education matched control group ofhealthy participants (11). Patients with w� scores greater than 2standard deviations from the mean of the control group were cat-egorized as significantly biased. Additionally, we examined signifi-cant deviations from the unbiased value of .5 using a one-sample ttest, and group differences by means of a two-sample t test on bothw� and on its deviation from .5.

PET MethodologyFor all patients, PET images were acquired on a Siemens 951

R/31 PET scanner, 30 min after injection of an intravenous bolus of370 MBq of 18FDG. During acquisition, patients were at rest withtheir eyes closed and positioned with the head parallel to the can-thomeatal line within the gantry. Three frames of 10-min durationwere acquired and averaged to a single frame. Image data wereacquired in two-dimensional mode with a total axial field of view of10.5 cm and no interplane gap space. Attenuation correction wasperformed with a standard ellipse-fitting method (3). Images werenormalized to a standard template in Montreal Neurological Insti-tute space using a nonlinear algorithm implemented in SPM5(http://www.fil.ion.ucl.ac.uk/spm) and smoothed using a 12 � 12 �12 mm3 Gaussian kernel. Individual global counts were proportion-

lly scaled to a mean value of 50 mg/100 mL/min. Global effects ofery early AD on glucose metabolism were explored by comparingET images from the patient group with images of 23 healthyge-matched control subjects (two-sample t test, p � .01, false

discovery rate [FDR] corrected, minimum of 50 contiguous voxels;for description of control subjects, see Drzezga et al. [3]).

Multiple Regression and Metabolic Bias-Index (M�) AnalysesTo test the hypothesis that patients’ PPC metabolism is related

to spatial bias, we used a multiple-regression model with patients’regional metabolism as the dependent variable and TVA-parame-ter w� as explanatory variable. Because we were interested in the

eural correlates of altered spatial attentional weighting, we intro-uced age and TVA parameters �, K, and C as variables of no interest

nto the model, controlling for processes that are not involved inpatial attentional selection (21). Initially, we restricted our analysiso subareas of the lateral PPC (i.e., the angular-supramarginal gyrus;ng-SMG), the posterior part of the superior temporal gyrus (pSTG),

he inferior parietal lobe (IPL), and the superior parietal lobe (SPL).his parcellation reflects the functional segregation of the PPC intoreas distinctively contributing to spatial attention (10,22). We de-ned regions of interest (ROIs) using the Automated Anatomicalabeling (AAL) tool (23) to extract the mean of all voxels in the ROI.ubsequently, we used a whole-brain analysis to look for correla-ions outside of these a priori ROIs (p � .001, uncorrected, minimumf 25 contiguous voxels).

To test whether the relative hypometabolism of functionallyelevant PPC subareas compared with their contralateral homo-ogues corresponds with spatial bias, an index of laterality of meta-olic bias (M�) was created by dividing the averaged metabolism Mf the right ROI by the sum of the averaged metabolism of the leftnd right ROI: M� � Mright/(Mright � Mleft). Relations between M�

and w were studied by means of partial correlations (p � .05,
�
onferroni-corrected for four ROI pairs), using the Mini-Mental Sta- s


us Examination (MMSE) score and mean metabolism in bilateralosterior cingulate (defined by AAL) as control variables (24). Appli-ation of control variables for general cognitive status (usingMSE) and for metabolism in the posterior cingulate cortex, a

egion representative for AD severity (3,4), ensures that potentialorrelations between M� and w� cannot be explained by globalognitive and metabolic decline in patients. To ensure that spatialias is not related to biased metabolism in the visual system, weerformed the same procedure for primary visual areas along thealcarine sulcus (defined by AAL). Frontal areas are known to beelevant for spatial attention (13,22). Because of the patients’ hypo-

etabolism in the superior frontal gyrus (SFG, see Results), weested the possibility that a metabolic imbalance in this region

ight also be related to their spatial attention bias. Using the samepproach as for the parietal regions, we calculated M� for the SFGas defined by AAL).

To examine M� relationships between IPL and ventral PPC, weested a multiple regression model with M�-IPL as the dependentariable and M�-Ang-SMG, M�-pSTG, and mean posterior cingulateetabolism as independent measures; cingulate metabolism con-

rols for disease severity (3,4,24).

esults

ypometabolism in the Lateral PPC in Very Early ADPatients demonstrated bilaterally reduced metabolism com-

ared with healthy elderly in the lateral PPC—namely, in the SPL,PL, and Ang-SMG, and pSTG. Furthermore, we found bilateral hy-ometabolism in the superior and middle frontal gyri, the anterioringulate cortex, the middle and posterior cingulate cortex, therecuneus, as well as in the cuneus and superior and middle occip-

tal gyri (Figure 2; Table S1 in Supplement 1).

patial Bias of Attentional Weighting in Very Early ADPatients’ w� deviated significantly from the optimal spatially

alanced value of .5 [t (34) � 2.2, p � .05] and was significantlyigher compared with control subjects [t (51.3) � 2.63, p � .05],

ndicating that patients’ distribution of spatial attention was later-lized toward the left hemifield (Figure 2, Table 1). Compared withhe control group of 36 healthy elderly people from Redel et al. (11),

patients demonstrated a strong spatial bias to the left, and 2atients were biased toward the right. In addition, patients had

arger deviations from .5 than control subjects [t(49.9) � 3.43, p �005], which was also manifest as an increase in variance [F (34,35) �.58, p � .001]. Spatial bias w� did not correlate with any of the otherVA parameters (highest r � �.27, lowest p � .11). Because thistudy focused on spatial attentional weighting in PPC, other TVAarameters are not discussed here.

patial Attentional Bias Is Related to PPC Hypometabolism inery Early AD

Spatial bias of attentional weighting w� significantly predictedetabolism in the left pSTG [t (30) � 3.1, p � .017, Bonferroni-

orrected for multiple comparisons, partial correlation coefficient� �.439], indicating increasing spatial bias to the left side with

ncreasing left-sided hypometabolism. Because we were interestedn whether additional, particularly frontal dysfunctions might con-ribute to spatial bias in very early AD, we repeated the multipleegression analysis across all voxels in the brain (Figure 3; Table S2 inupplement 1). Negative correlations with w� were restricted to the

eft ventral PPC, with local maxima in the pSTG [peak x � �54, y �46, z � 18, t (30)�4.58] and the nearby posterior middle temporal

yrus [peak x � �50, y � �66, z � 0, t (30) � 4.28]. No further areas
howed any significant correlations.
http://www.fil.ion.ucl.ac.uk/spm

acF

v

.

tm

tl.

D

csedsbcpa

S

p


Direction and Degree of Spatial Attentional Bias Is Related toDirection and Degree of the PPC’s InterhemisphericMetabolism Bias in Very Early AD

The metabolism bias M� correlated significantly with the spatialttention bias w� in the Ang-SMG (r � .44, p � .01 Bonferroni-orrected for multiple comparisons) and the IPL (r � .41, p � .016;igure 4). Neither indices of the pSTG nor the SPL correlated with w�

(r � .09 and r � .29, respectively). The metabolism bias M� of theisual cortex was not correlated with w� (r � �.04). We found a

significant positive relationship between w� and SFG (r � .363, p �.038; see Figure S1 in Supplement 1). When restricting the SFG-ROIto hypometabolic voxels (as defined by the t test between patientsand control subjects, thresholded at an false discovery rate– cor-rected p � .01), the correlation was reduced to a trend (r � .338, p �054).

Given reports of a strong interaction between areas of the ven-ral PPC and the IPL (25), we explored whether the interhemispheric

etabolism bias M� of the IPL and the Ang-SMG or pSTG wererelated. We found a highly significant correlation between metab-olism bias M of the IPL and the Ang-SMG [t (33) � 5.05, p � .001,
�
R2 � .539, p � .001, for the whole model], confirming our sugges- b

ion that these areas do functionally interact. The bias of metabo-ism M�-pSTG did not explain additional variance of M�-IPL [t (33) �51, p � .6].

iscussion

In this study, a model-based imaging approach based on theomputational theory of visual attention and FDG-PET was used totudy the relationship between metabolism in the posterior pari-tal cortex and spatial attention in very early AD. We found evi-ence that patients’ asymmetric lateral PPC metabolism corre-ponds with the direction and strength of their spatial attentionalias. This result is consistent with the hypothesis that asymmetri-ally impaired effective neuronal PPC activity in AD biases the com-etition of visual objects for cortical representation and access towareness to one side.

patial Attentional Weighting Is Impaired in Very Early ADWe found significant spatial bias in very early AD, in line with

revious results (11) (Figure 2). Left-sided bias dominated. A spatial

Figure 2. Hypometabolism and deficits in selective at-tention in very early Alzheimer’s disease (veAD). (Toppanel) Patients with veAD demonstrated hypometab-olism compared with healthy control (HC) subjects(two-sample t test, p � .01, false discovery rate– cor-rected, minimum cluster extent of 50 voxels). Spatialpattern of hypometabolism is rendered onto the lateralsurfaces of a population-average, landmark- and sur-face-based (PALS) atlas brain (47) and onto an axial slicein Montreal Neurological Institute space (Z � 34). Col-ors on the rendered brains range from 0 to 5 (right colorbar) and indicate t scores. Colors on the axial slice indi-cate t scores ranging from 0 to 8 (left color bar). (Bottompanel) Mean values for spatial attentional bias w� inthe partial report task of patients with veAD andhealthy control subjects. Error bars indicate the stan-dard error. Patients’ w� deviated significantly from theoptimal spatially balanced value of .5 [t(34) � 2.2, p �.05] and was significantly higher compared with con-trol subjects [t(51.3) � 2.63, p � .05], indicating thatpatients’ distribution of spatial attention was lateral-ized toward the left hemifield. Furthermore, nine pa-tients demonstrated a strong spatial bias to the left,and 2 patients were biased toward the right (i.e., thesepatients were far from the mean w� of healthy controlsubjects by more than two standard deviations). �, top-down control; w�, visual field.

ias toward the left hemifield has been demonstrated in patients


ilrrfiaoadas

r nd C bs �, spa

b

c


with mild or moderate AD (26 –28) and to become stronger duringthe progression from aMCI to mild AD (11).

Interhemispheric Metabolic Bias in Lateral PPC Correspondswith Spatial Attentional Bias in Very Early AD

We found bilateral hypometabolism in the lateral PPC in veryearly AD (Figure 2). Because AD progressively degrades neuronaland synaptic activity underlying the FDG-PET signal (2,4,6 – 8,29),hypometabolism indicates reduced effective neuronal and synap-tic activity. Decreasing metabolism in the left pSTG was associatedwith left-lateralized spatial bias (Figure 3); that is, when metabolism

Figure 3. Regional correlates of spatial bias w� in very early Alzheimer’s dise(middle temporal gyrus, posterior part, and superior temporal gyrus, shownof attentional weights distributed across the left and right visual hemifield.egressor of interest and further theory of visual attention parameters �, K, aize 25 voxels). Colors indicate t values ranging from 0 to 4 (see color bar). w

Figure 4. Metabolic biases M� correlate with spatial bias w�. In analogy to sparain region as Mright/(Mright � Mleft), where Mright/left represents the mean m

of the angular-supramarginal gyrus (Ang-SMG; lower right panel) and the inorrelation, p � .05, Bonferroni-corrected for four ROI pairs). M� of the super

(pSTG; lower left) do not correlate. The image in the center shows the ROsurface-based (PALS) atlas average brain using CARET software (47). The rig
(blue bar graphs) and Ang-SMG (green) in exemplary patients exhibiting either athe scatterplots is indicated by purple (HS) and orange (MM) circles. FDG-PET, 18F

n left pSTG is low, attentional weights are higher for objects in theeft hemifield, which are represented by perceptual neurons in theight hemisphere. This implies that objects in the left hemifieldeceive higher attentional weights than objects in the right hemi-eld, and consequently we expected the relationship between left-nd right-hemispheric metabolism to be a more sensitive predictorf spatial bias than left-hemisphere metabolism alone. Directionnd degree of spatial bias were indeed predicted by direction andegree of metabolic bias of the IPL and Ang-SMG (Figure 4, as wells the SFG, see Figure S1 in Supplement 1). Because across patientspatial bias was distributed either to the left (n � 9) or to the right

he spatial map represents regions of the patient group correlating with w�

e sagittal (X � �50) and one axial slice (Z � 18). w� reflects the spatial biasap is based on a voxel-wise multiple regression analysis with w� being theeing the regressors of no interest (p � .001, uncorrected, minimum clustertial bias.

ias w�, we defined an index for the lateral bias of metabolism (M�) of a givenlism in a right/left hemisphere region of interest (ROI). Metabolic biases M�

parietal lobe (IPL; upper right) correlate significantly with w� (x axes, partialrietal lobe (SPL; upper left) and the superior temporal gyrus, posterior parted for the analysis, rendered onto a population-average, landmark- andnel shows left (L) and right (R) hemisphere regional metabolism in the IPL

ase. Ton on

The m

tial betaboferiorior paIs us

ht pa
right- (HS) or left-lateralized (MM) spatial attentional bias. Their location on-fluorodeoxyglucose positron emission tomography.

Th(pmsp

tspaRmcgot

snOmtlAaicorsrb

IS

ipcsSpptic

tbtmticrltslu

rb(l

C

bepbtcshsdtaciifithmcwaeel

C

saccpme

aFcs8

c


(n � 2), this relationship most probably holds for both hemifields.he metabolic bias reflects a stable interhemispheric imbalance ofomologue areas known to be responsible for spatial attention

9,10). In contrast to the IPL and Ang-SMG, metabolic bias in theSTG was not a better predictor of spatial bias than left hemisphereetabolism alone. Possibly, this reflects an additional influence on

patial attention that does not depend on interhemispheric com-etition, which could be explicitly tested in future studies.

Because we found a relationship between SFG and spatial bias,his frontal region may also contribute to an impaired allocation ofpatial attention in patients. This finding is consistent with severalrevious findings on dorsolateral frontal involvement in spatialttention tasks in healthy humans (22,30) and in monkeys (31,32).egions in SFG have been shown to be involved in both eye move-ent plans and biases for activity in retinotopic areas of visual

ortex (33–36), that is, SFG has a prominent role in the top-downuidance of spatial attention. Asymmetric top-down bias signalsriginating from the SFG might therefore additionally contribute to

he asymmetry in the lateral parietal cortex.Metabolic bias in the pericalcarine cortex was not related to

patial bias, indicating that the correlations we found in the PPC areot the result of imbalances in an upstream visual processing area.ur findings also cannot be explained by global cognitive andetabolic decline, because correlations with spatial bias were con-

rolled for general cognitive status (using MMSE) and for metabo-ism in the posterior cingulate cortex, a region representative forD-severity (3,4,24). In support of our regional hypothesis, Megurond colleagues found that in patients with moderate AD, omissions

n a visual search task are related to the overall perfusion of theontralateral parietal lobe (37). In stroke patients, lesions in the leftr right PPC lead to an ipsi-lesional bias the strength of which is

elated to the lesion’s volume (10,15). In patients with right-hemi-phere stroke, the task-evoked activity imbalance between theight and left IPL is correlated with the degree of right-sided spatialias (38).

nterregional Competition as a Candidate Mechanism forpatial Attentional Bias in Very Early AD

Models of the functional anatomy of attention suggest strongnteractions between the IPL, as it computes expectancy-drivenriority maps, and the ventral PPC, which modulates these maps inases of bottom-up breaches of expectancy (10,22). We found atrong correlation between metabolic bias in the IPLs and the Ang-MGs of the ventral PPC, which has also been found in strokeatients with parietal lesions (25). The network degeneration hy-othesis suggests that AD propagates along transsynaptic connec-

ions (39,40), and thus it seems likely that because of the strongnterconnection of these areas, an imbalance in one area will in-rease the imbalance in the other.

When considering the possible mechanism of imbalanced at-entional weights in very early AD, we conclude that our results areroadly consistent with the biased-competition principle of atten-

ion (13), although noncompetitive mechanisms (in the left pSTG)ay also be in play. Because our measures of attentional parame-

ers were based on the assumptions made by TVA (12), we cannterpret the observed left-sided spatial bias as a permanently de-reased competitiveness of processing units responsible for theight hemifield (compared to left hemifield units). At the neuralevel, this may mean that competition between neuronal popula-ions encoding different hemifields is biased in favor of right-hemi-pheric neurons (encoding the left hemifield) because of a largeross of effective neurons in the homologous left-hemispheric pop-
lation. Because we found a loss of effective neurons in the PPC
ather than the primary visual cortex, we conclude that the spatialias was caused by a loss of neurons responsible for setting the bias

9,10), whereas neurons involved in primary visual processing wereikely spared.

linical-Behavioral ImplicationsThe patient group we described had a relatively small spatial

ias (compared to the hemispatial neglect that often follows pari-tal stroke) (41, 42), indicating that the parietal cortex in theseatients is still partly functional. Nonetheless, their spatial bias cane described as pathological in the sense that it indicates a subop-

imal distribution of attentional resources (compared with healthyontrol subjects). In addition, patients’ spatial attention deficiteems to be accompanied by further attention deficits. Recent be-avioral findings suggest that nonspatial forms of attention are alsoubtly impaired in very early AD and MCI (e.g., task-related top-own attention) (11,43,44). These small but broad deficits of selec-

ive attention including spatial attention in the earliest stages of ADre largely ignored in clinical and everyday contexts, particularlyompared with patients’ more obvious memory deficits. However,

mpairments in the efficient use of attention to select goal-relevantnformation from a complex and dynamic environment (such asnding the right train in a crowded station or driving a car) is likelyo affect the daily life of patients (45). Indeed, patients with MCIave prominent problems navigating such environments. Further-ore, attention and PPC processes critically influence memory pro-

esses that are associated with the medial temporal lobe system,hich in turn are strongly affected by AD (1,2,46); consequently,

ttentional impairments associated with PPC changes might exac-rbate memory deficits in early AD. Future studies are necessary toxplore the potential effects of spatial attention on patients’ daily

ife and memory impairments.

onclusionIn very early AD, asymmetric lateral PPC hypometabolism causes

patial attentional bias. Data are consistent with the model thatltered distributions of space-related attentional weights in the PPContribute to a spatial bias in the competition of visual objects forortical representation and access to awareness. Because PPC hy-ometabolism is one of the first in vivo signs of AD, the outlinedechanism might underlie spatial attentional impairments of very

arly AD.

AMW is supported by the German Federal Ministry of Educationnd Research (Grant No. 01EV0710), VR is supported by the Alzheimerorschung Initiative (Grant No. 08860, NM is supported by the Well-ome Trust, and CS is supported by the Kommission für Klinische For-chung of the University Hospital Klinikum Rechts der Isar (Grant No.765162).

All authors report no biomedical financial interests or potentialonflicts of interest.

Supplementary material cited in this article is available online.

1. Blennow K, de Leon MJ, Zetterberg H (2006): Alzheimer’s disease. Lancet368:387– 403.

2. Braak H, Braak E (1991): Neuropathological staging of Alzheimer-relatedchanges. Acta Neuropathol (Berl) 82:239 –259.

3. Drzezga A, Lautenschlager N, Siebner H, Riemenschneider M, Willoch F,Minoshima S, Schwaiger M, et al. (2003): Cerebral metabolic changesaccompanying conversion of mild cognitive impairment into Alzhei-
mer’s disease: a PET follow-up study. Eur J Nucl Med Mol Imaging 30:1104 –1113.

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4


4. Mosconi L, Pupi A, De Leon MJ (2008): Brain glucose hypometabolismand oxidative stress in preclinical Alzheimer’s disease. Ann N Y Acad Sci1147:180 –195.

5. Kasischke KA, Vishwasrao HD, Fisher PJ, Zipfel WR, Webb WW (2004):Neural activity triggers neuronal oxidative metabolism followed by as-trocytic glycolysis. Science 305:99 –103.

6. Pellerin L, Magistretti PJ (1994): Glutamate uptake into astrocytes stim-ulates aerobic glycolysis: A mechanism coupling neuronal activity toglucose utilization. Proc Natl Acad Sci U S A 91:10625–10629.

7. Raichle ME, Mintun MA (2006): Brain work and brain imaging. Annu RevNeurosci 29:449 – 476.

8. Rocher AB, Chapon F, Blaizot X, Baron JC, Chavoix C (2003): Resting-statebrain glucose utilization as measured by PET is directly related to re-gional synaptophysin levels: A study in baboons. Neuroimage 20:1894 –1898.

9. Sack AT (2009): Parietal cortex and spatial cognition. Behav Brain Res202:153–161.

10. Vandenberghe R, Gillebert CR (2009): Parcellation of parietal cortex:Convergence between lesion-symptom mapping and mapping of theintact functioning brain. Behav Brain Res 199:171–182.

11. Redel P, Bublak P, Sorg C, Kurz A, Forstl H, Muller HJ, Schneider WX et al.(2010): Deficits of spatial and task-related attentional selection in mildcognitive impairment and Alzheimer’s disease [published online aheadof print June 17]. Neurobiol Aging.

12. Bundesen C (1990): A theory of visual attention. Psychol Rev 97:523–547.13. Desimone R, Duncan J (1995): Neural mechanisms of selective visual

attention. Annu Rev Neurosci 18:193–222.14. Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger-Gateau P, Cum-

mings J, Delacourte A, et al. (2007): Research criteria for the diagnosis ofAlzheimer’s disease: Revising the NINCDS-ADRDA criteria. Lancet Neurol6:734 –746.

15. Peers PV, Ludwig CJ, Rorden C, Cusack R, Bonfiglioli C, Bundesen C,Driver J, et al. (2005): Attentional functions of parietal and frontal cortex.Cereb Cortex 15:1469 –1484.

16. Morris JC (1993): The Clinical Dementia Rating (CDR): Current versionand scoring rules. Neurology 43:2412–2414.

17. Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G,Mellits ED, et al. (1989): The Consortium to Establish a Registry for Alz-heimer’s Disease (CERAD). Part I. Clinical and neuropsychological as-sessment of Alzheimer’s disease. Neurology 39:1159 –1165.

18. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, Bel-leville S, et al. (2006): Mild cognitive impairment. Lancet 367:1262–1270.

19. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM(1984): Clinical diagnosis of Alzheimer’s disease: Report of theNINCDS-ADRDA Work Group under the auspices of Department ofHealth and Human Services Task Force on Alzheimer’s Disease. Neurol-ogy 34:939 –944.

20. Bublak P, Redel P, Sorg C, Kurz A, Forstl H, Muller HJ, Schneider WX et al.(2011): Staged decline of visual processing capacity in mild cognitiveimpairment and Alzheimer’s disease. Neurobiol Aging 32:1219 –1230.

21. Husain M, Rorden C (2003): Non-spatially lateralized mechanisms inhemispatial neglect. Nat Rev Neurosci 4:26 –36.

22. Corbetta M, Shulman GL (2002): Control of goal-directed and stimulus-driven attention in the brain. Nat Rev Neurosci 3:201–215.

23. Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O,Delcroix N, Mazoyer B, et al. (2002): Automated anatomical labeling ofactivations in SPM using a macroscopic anatomical parcellation of theMNI MRI single-subject brain. Neuroimage 15:273–289.

24. Vukovich R, Perneczky R, Drzezga A, Förstl H, Kurz A, Riemensch-neider M (2009): Brain metabolic correlates of cerebrospinal fluidbeta-amyloid 42 and tau in Alzheimer’s disease. Dement Geriatr CognDisord 27:474 – 480.

25. He BJ, Snyder AZ, Vincent JL, Epstein A, Shulman GL, Corbetta M (2007):Breakdown of functional connectivity in frontoparietal networks under-
lies behavioral deficits in spatial neglect. Neuron 53:905–918.

6. Drago V, Foster PS, Ferri R, Arico D, Lanuzza B, Heilman KM (2008):Distractibility and Alzheimer disease: The “neglected” phenomenon. JAlzheimers Dis 15:1–10.

7. Ishiai S, Koyama Y, Seki K, Orimo S, Sodeyama N, Ozawa E, Lee EY et al.(2000): Unilateral spatial neglect in AD: Significance of line bisectionperformance. Neurology 55:364 –370.

8. Mendez MF, Cherrier MM, Cymerman JS (1997): Hemispatial neglect onvisual search tasks in Alzheimer’s disease. Neuropsychiatry NeuropsycholBehav Neurol 10:203–208.

9. Selkoe DJ (2002): Alzheimer’s disease is a synaptic failure. Science 298:789 –791.

0. Mesulam MM (1999): Spatial attention and neglect: Parietal, frontal andcingulate contributions to the mental representation and attentionaltargeting of salient extrapersonal events. Phil Trans R Soc Lond B Biol Sci354:1325–1346.

1. Everling S, Tinsley CJ, Gaffan D, Duncan J (2002): Filtering of neuralsignals by focused attention in the monkey prefrontal cortex. Nat Neu-rosci 5:671– 676.

2. Moore T, Fallah M (2004): Microstimulation of the frontal eye field and itseffects on covert spatial attention. J Neurophysiol 91:152–162.

3. Paus T (1996): Location and function of the human frontal eye-field: Aselective review. Neuropsychologia 34:475– 483.

4. Hanes DP, Schall JD (1996): Neural control of voluntary movement initi-ation. Science 274:427– 430.

5. Moore T, Armstrong KM (2003): Selective gating of visual signals bymicrostimulation of frontal cortex. Nature 421:370 –373.

6. Gregoriou GG, Gotts SJ, Zhou H, Desimone R (2009): High-frequency,long-range coupling between prefrontal and visual cortex during atten-tion. Science 324:1207–1210.

7. Meguro K, Shimada M, Someya K, Horikawa A, Yamadori A (2001):Hemispatial visual-searching impairment correlated with decreasedcontralateral parietal blood flow in Alzheimer disease. NeuropsychiatryNeuropsychol Behav Neurol 14: 213–218.

8. Corbetta M, Kincade MJ, Lewis C, Snyder AZ, Sapir A (2005): Neural basisand recovery of spatial attention deficits in spatial neglect. Nat Neurosci8:1603–1610.

9. Palop JJ, Chin J, Mucke L (2006): A network dysfunction perspective onneurodegenerative diseases. Nature 443:768 –773.

0. Seeley WW, Crawford RK, Zhou J, Miller BL, Greicius MD (2009): Neuro-degenerative diseases target large-scale human brain networks. Neuron62:42–52.

1. Bublak P, Finke K, Krummenacher J, Preger R, Kyllingsbaek S, Müller HJ,Schneider WX (2005): Usability of a theory of visual attention (TVA) forparameter-based measurement of attention II: Evidence from two pa-tients with frontal or parietal damage. J Int Neuropsychol Soc 11:843–854.

2. Duncan J, Bundesen C, Olson A, Humphreys G, Chavda S, Shibuya H(1999): Systematic analysis of deficits in visual attention. J Exp PsycholGen 128:450 – 478.

3. Neufang S, Akhrif A, Riedl V, Förstl H, Kurz A, Zimmer C, Sorg C,Wohlschläger AM (2011): Disconnection of frontal and parietal areascontributes to impaired attention in very early Alzheimer’s disease. JAlzheimers Dis 25:309 –321.

4. Tales A, Snowden RJ, Phillips M, Haworth J, Porter G, Wilcock G, Bayer A(2011): Exogenous phasic alerting and spatial orienting in mild cogni-tive impairment compared to healthy ageing: Study outcome is relatedto target response. Cortex 47:180 –190.

5. Perneczky R, Pohl C, Sorg C, Hartmann J, Tosic N, Grimmer T, Heitele S,Kurz A (2006): Impairment of activities of daily living requiring memoryor complex reasoning as part of the MCI syndrome. Int J Geriatr Psychia-try 21:158 –162.

6. Cabeza R, Ciaramelli E, Olson IR, Moscovitch M (2008): The parietalcortex and episodic memory: An attentional account. Nat Rev Neurosci9:613– 625.

7. Van Essen DC, Drury HA, Dickson J, Harwell J, Hanlon D, Anderson CH
(2001): An integrated software suite for surface-based analyses of cere-bral cortex. J Am Med Inform Assoc 8:443– 459.

Asymmetric Loss of Parietal Activity Causes Spatial Bias in Prodromal and Mild Alzheimer's Disease

Documents