Abstract

To investigate early effects of beta-amyloid (Aβ) on neuronal function, elderly normal controls (NCs, age range 58–97) were scanned with Pittsburgh Compound-B (PIB) positron emission tomography (a measure of Aβ) as well as functional magnetic resonance imaging (a measure of brain activation) while performing an episodic memory–encoding task of natural scenes (also performed by young NCs; age range 18–30). Relationships between Aβ and activation were assessed across task-positive (regions that activate for subsequently remembered vs. forgotten scenes) and task-negative regions (regions that deactivate for subsequently remembered vs. forgotten scenes). Significant task-related activation was present in a distributed network spanning ventrolateral prefrontal, lateral occipital, lateral parietal, posterior inferior temporal cortices, and the right parahippocampal/hippocampus, whereas deactivation was present in many default mode network regions (posteromedial, medial prefrontal, and lateral temporoparietal cortices). Task-positive activation was higher in PIB+ compared with PIB− subjects, and this activation was positively correlated with memory measures in PIB+ subjects. Although task deactivation was not impaired in PIB+ NCs, deactivation was reduced in old versus young subjects and was correlated with worse task memory performance among old subjects. Overall, these results suggest that heightened activation during episodic memory encoding is present in NC elderly subjects with high Aβ.

Introduction

Although beta-amyloid (Aβ) plaques may play a role in the etiology of Alzheimer’s disease (AD) (Hardy and Selkoe 2002; Walsh and Selkoe 2007), this pathology is commonly observed in the brains of elderly normal control (NC) individuals at postmortem examination (Bennett et al. 2006; Kok et al. 2009; Savva et al. 2009). Similar findings have been revealed with in vivo imaging using positron emission tomography (PET) and Pittsburgh Compound-B (PIB), a ligand capable of quantifying fibrillar Aβ plaque burden (Morris et al. 2010; Rowe et al. 2010). This may reflect the earliest stage in AD development, with a possible 10-year delay between initial Aβ deposition and dementia onset (Sperling et al. 2011). This idea is supported by subtle “AD-like” changes in NCs with high Aβ burden, such as brain atrophy (Dickerson et al. 2009; Mormino et al. 2009; Storandt et al. 2009; Chetelat et al. 2010; Oh et al. 2010; Becker et al. 2011) and cognitive decline (Morris et al. 2009; Villemagne et al. 2011). However, some evidence suggests that cognitive reserve processes may modify the relationship between Aβ and detrimental downstream effects (Fotenos et al. 2008; Kemppainen et al. 2008; Rentz et al. 2010). These factors might enable individuals to cope with this pathology (Stern 2006) or at least prolong the delay period between pathological insult and cognitive impairment (Jack et al. 2010).

Many functional magnetic resonance imaging (fMRI) studies have revealed that increased activation in aging (Cabeza et al. 2002; Rosen et al. 2002; Park and Reuter-Lorenz 2009; de Chastelaine et al. 2011) as well as in populations at risk for developing AD (Bookheimer et al. 2000; Dickerson et al. 2005; Trivedi et al. 2008). These findings are often interpreted as a compensatory response to pathology, suggesting a neural basis of cognitive reserve processes. It is also possible that activation increases are detrimental, reflecting reduced neural efficiency and/or dedifferientation (Logan et al. 2002; Li et al. 2006). Distinction between beneficial versus detrimental processes is often addressed by isolating activation during successful events as well as by examining relationships with measures of overall memory performance.

To investigate the role of these processes during early Aβ deposition, elderly subjects underwent PIB-PET imaging as well as fMRI during incidental encoding of outdoor scenes. A group of younger control subjects also participated in the fMRI experiment. The event-related fMRI design enabled isolation of “subsequent memory effects” or activation during successfully encoded scenes. Additionally, deactivations (greater deactivation for subsequently remembered vs. forgotten scenes) are typically present in the “default mode network” (DMN), specifically posteromedial, medial prefrontal, and lateral temporoparietal cortices (Daselaar et al. 2004). Although previous work has not emphasized relationships between Aβ and task-positive activation, disruption of the DMN in NCs with high Aβ burden has been reported both at rest (Hedden et al. 2009; Sheline et al. 2009; Mormino et al. 2011) and during episodic memory encoding (Sperling et al. 2009; Vannini et al. 2011). To this end, the aim of the current study was to explore relationships between Aβ burden, activation, deactivation, and cognition in elderly NC subjects.

Materials and Methods

Subject Recruitment

Fifty elderly NC subjects underwent PIB-PET imaging and fMRI for this study, and 17 young NC subjects underwent fMRI. Eligibility requirements for all NC subjects were no MRI contradictions, living independently in the community, Mini-Mental State Examination (MMSE) ≥ 26, normal performance on cognitive tests, absence of neurological or psychiatric illness, and lack of major medical illnesses and medications that affect cognition. Five elderly subjects and 2 young were excluded due to insufficient trials of high-confidence hits (<20, N = 3), problems with data acquisition (N = 1), excessive motion (N = 2), and memory performance at chance levels (N = 1), resulting in a total number of 45 old and 15 young NC subjects for data analysis.

Neuropsychological Testing

Thirty-nine old NC subjects underwent a detailed cognitive testing battery. To reduce this data, a maximum-likelihood factor analysis with varimax rotation was completed using data from a larger cohort of 349 cognitively normal subjects (age range = 20–96, mean age = 58.2 (22.8), mean MMSE = 29.0 (1.5), mean education = 16.8 (2.1), 133 males) who underwent the same test protocol. The following measures were entered into the factor analysis: mental control, verbal paired associates, logical memory, and visual reproductions I and II from the Wechsler Memory Scale-Revised (Wechsler 1987), digits forward/backward and digit symbol tests from the Wechsler Adult Intelligence Scale-Revised, recall sum across learning trials in the California verbal learning test (Delis et al. 2000), Boston Naming Test (Kaplan et al. 1983), Trails B minus A (Reitan 1958), FAS phonemic fluency (Lezak 1995), and total correct in 60 s from the Stroop Interference test (Zec 1986). This factor analysis revealed 5 factors fulfilling Kaiser’s criterion (eigenvalues > 1) and were labeled according to the neuropsychological tests with the highest factor loadings (episodic memory, executive function, working memory, visual memory, and semantic memory; Table 1). Thompson factor scores were estimated for each subject (Burt and Thomson 1947) and used to compare cognition to PIB uptake and fMRI activation within old subjects. For subjects that had undergone multiple testing sessions, cognitive test scores closest to the PET scan date were used in the factor analysis (mean delay between PET and closest testing session was 4.4[3.0] months). The remaining 6 old subjects underwent a different neuropsychological battery and were excluded from the factor analysis due to intersession incompatibilities (however, neuropsychological testing for these subjects was used to ensure normality in these subjects).

Table 1

Factor analysis factor loadings

 Factor1/EM Factor2/Exe Factor3/WM Factor4/VisMem Factor5/SM 
WMS-R mental control 0.14 0.65 0.20 0.15 0.30 
WMS-R verbal paired associates 0.77 0.16 NA 0.17 0.17 
WMS-R logical memory 0.60 0.24 NA 0.12 0.28 
WMS-R visual reproduction I 0.42 0.31 0.12 0.83 0.11 
WMS-R visual reproduction II 0.57 0.33 NA 0.54 NA 
WAIS-R digit span forward 0.10 0.16 0.97 NA 0.11 
WAIS-R digit span backward NA 0.35 0.43 NA 0.33 
WAIS-R digit symbol 0.47 0.71 NA 0.25 −0.13 
CVLT 1–5 free recall 0.67 0.28 0.13 0.21 NA 
Boston naming test NA NA NA NA 0.40 
Trails B minus A −0.21 −0.44 −0.13 −0.19 −0.18 
FAS phonemic fluency 0.14 0.39 0.15 NA 0.43 
Stroop 0.39 0.62 0.13 0.19 NA 
 Factor1/EM Factor2/Exe Factor3/WM Factor4/VisMem Factor5/SM 
WMS-R mental control 0.14 0.65 0.20 0.15 0.30 
WMS-R verbal paired associates 0.77 0.16 NA 0.17 0.17 
WMS-R logical memory 0.60 0.24 NA 0.12 0.28 
WMS-R visual reproduction I 0.42 0.31 0.12 0.83 0.11 
WMS-R visual reproduction II 0.57 0.33 NA 0.54 NA 
WAIS-R digit span forward 0.10 0.16 0.97 NA 0.11 
WAIS-R digit span backward NA 0.35 0.43 NA 0.33 
WAIS-R digit symbol 0.47 0.71 NA 0.25 −0.13 
CVLT 1–5 free recall 0.67 0.28 0.13 0.21 NA 
Boston naming test NA NA NA NA 0.40 
Trails B minus A −0.21 −0.44 −0.13 −0.19 −0.18 
FAS phonemic fluency 0.14 0.39 0.15 NA 0.43 
Stroop 0.39 0.62 0.13 0.19 NA 

Note: Factor loadings above 0.40 are bolded. NAs are indicated for loadings between −0.1 and +0.1. Based on these loadings, factors were label as episodic memory (EM), executive function (Exe), working memory (WM), visuospatial memory (VisMem), and semantic memory (SM). WMS-R, Wechsler Memory Scale-Revised; WAIS-R, Wechsler Adult Intelligence Scale-Revised; CVLT, California verbal learning test; and NA, not applicable.

Apolipoprotein E Genotyping

DNA from blood samples for old NC subjects was analyzed for apolipoprotein E (APOE) polymorphisms using a standard protocol. For statistical comparison between groups, subjects were dichotomized into carriers and noncarriers of the E4 allele. Genotyping was unavailable for 3 old NC subjects.

PIB-PET Acquisition

PIB was synthesized at the Lawrence Berkeley National Laboratory’s (LBNL) Biomedical Isotope Facility using a published protocol and described in detail previously (Mathis et al. 2003; Mormino et al. 2009). PIB-PET imaging was performed at LBNL using an ECAT EXACT HR PET scanner (Siemens Medical Systems, Erlangen, Germany) in 3D acquisition mode. Ten to fifteen millicurie of PIB was injected into an antecubital vein. Dynamic acquisition frames were obtained as follows: 4 × 15, 8 × 30, 9 × 60, 2 × 180, 8 × 300, and 3 × 600 s (90 min total). Ten-minute transmission scans for attenuation correction were obtained for each PIB scan. Filtered backprojected reconstructions were performed on the transmission and emission data to judge transmission alignment with each frame of emission data. In the case of misalignment, the transmission image was coregistered to that individual emission frame and then forward projected to create an attenuation correction file specific to that head position. PET data were reconstructed using an ordered subset expectation maximization algorithm with weighted attenuation. Images were smoothed with a 4 mm Gaussian kernel with scatter correction.

MRI Acquisition

All subjects underwent MRI scanning at LBNL on a 1.5T Magnetom Avanto System (Siemens Medical Systems) with a 12 channel head coil run in triple mode. A high-resolution structural T1-weighted volumetric magnetization prepared rapid gradient echo scan (MP-RAGE, axially acquired, time repetition [TR]/time echo [TE]/time to inversion [TI] = 2110/3.58/1100 ms, flip angle = 15°, 1.00 × 1.00 mm2 in plane resolution, 1.00 mm thickness with 50% gap) and a low-resolution structural T1-weighted in plane to the fMRI scans were collected (axially acquired, TR/TE = 591/10 ms, flip angle = 150°, 0.90 × 0.90 mm2 in plane resolution, 3.40 mm thickness with 15% gap). For fMRI scanning, 4 T2*-weighted gradient-echo echo planar images (EPI) were collected (28 axially acquired slices, TR/TE = 2200/50 ms, flip angle = 90°, 3.40 × 3.40 mm2 in plane resolution, 3.40 mm thickness with 15% gap).

Episodic Memory Paradigm

Two hundred outdoor images of natural outdoor scenes were presented for 4.4 s each, and subjects were instructed to indicate whether water was present at any point during the image’s presentation (collected over 4 EPI scans, with 50 scenes and 185 TRs per EPI). Zero to 5 TRs of fixation (green crosshair on black background) were randomly intermixed between scenes to allow separation of individual trials (average interstimulus interval = 3.46 s (3.01); Dale 1999). A postscan surprise recognition task with all stimuli presented during encoding as well as 100 foils was used to assess performance and sort fMRI data (there was a 15-min delay between the last stimulus encoded and the start of the postscan memory test). For each target and foil, subjects were asked “Have you seen this image,” and allowed to respond with 1 of 4 responses: high-confidence yes, low-confidence yes, high-confidence no, and low-confidence no. This recognition task was self-paced, and subjects were encouraged to be as accurate as possible. Overall task memory performance was assessed with the discriminability index d-prime, which uses signal detection theory to compare recognition accuracy and false-positive rates. All low-confidence responses were discarded to calculate d-prime values.

PIB-PET Processing

PIB-PET data were preprocessed using the SPM8 software package (http://www.fil.ion.ucl.ac.uk/spm, last accessed 15 August 2011). Realigned PIB frames corresponding to the first 20 min of acquisition were averaged and used to guide coregistration to the subject’s structural MRI scan. Distribution volume ratios (DVRs) for PIB images were created using Logan graphical analysis with frames corresponding to 35–90 min postinjection and a gray matter masked cerebellum reference region defined using FreeSurfer software (Logan et al. 1996; Lopresti et al. 2005).

Structural MRI Processing

MP-RAGE scans were processed as described previously (Mormino et al. 2009) using FreeSurfer version 4.5.0 (http://surfer.nmr.mgh.harvard.edu/, last accessed 15 August 2011) to derive regions of interest (ROIs) in each subject’s native space (Dale et al. 1999; Fischl et al. 2001, 2002; Segonne et al. 2004). PIB index values were derived by averaging PIB DVR values from prefrontal, cingulate, lateral temporal, and parietal ROIs.

Dichotomization into PIB+ and PIB− Groups

Older NC subjects were divided into PIB+ and PIB− groups based on cutoff values defined using 11 young NC subjects who were scanned with PIB-PET without evidence of PIB uptake and were therefore presumably Aβ-free (Kok et al. 2009; Braak and Del Tredici 2011). Seven young subjects scanned with fMRI were part of this young group (5 males, mean age = 24.5 (3.4), mean education = 16.2 (1.9)). The PIB positive cutoff was defined as 2 standard deviations (SDs) above the mean PIB index value across young subjects (mean = 1.04, SD = 0.02), resulting in a value of 1.08 (Mormino et al. 2011).

fMRI Processing

fMRI data were processed using FSL version 4.1.6 (http://www.fmrib.ox.ac.uk/fsl, last accessed 15 August 2011). Images were motion corrected, highpass filtered (100 s), and smoothed with a 5 mm Gaussian kernel. To define the spatial transformation from fMRI space to Montreal Neurological Institute (MNI) template space, a multistep registration procedure was employed. First, the mean fMRI image from each encoding run was linearly registered to the subject’s in-plane T1 structural image using 7 degrees of freedom (df). The in-plane T1 image was then registered to the high-resolution structural scan using 6 df. Finally, the high-resolution structural scan was nonlinearly aligned to the standard MNI 152 brain using FNIRT, and resulting parameters were used to transform fMRI data.

fMRI Modeling and Higher Level Analyses

fMRI trials were classified into 4 types (high-confidence hits, low-confidence hits, high/low-confidence misses, and nonwater response trials), modeled by a box function with duration of 4.4 s, and convolved with a standard gamma hemodynamic response function. These covariates, as well as corresponding temporal derivatives and the 6 rigid body realignment motion parameters, were entered in a general linear model predicting fMRI signal intensity. This lower level analysis was completed separately for each encoding run (N = 4) for each subject, then combined across runs using a fixed effects analysis within each subject. Resulting contrast and variance maps (corresponding to high-confidence hits vs. misses) for all young and old subjects were carried forward into a one-sample t-test random effects model to determine areas showing significant differences between conditions, covarying for memory performance (task positive = higher activation in hits vs. misses; task negative = higher deactivation in hits vs. misses; thresholded at z >1.64 with a cluster significance of P = 0.01, corrected for multiple comparisons).

Peak activations and deactivations were selected from this one-sample t-test and used to create spherical ROIs (6 mm centered around selected peak coordinates). Additionally, a hippocampal ROI was defined by masking the activation map with an anatomically defined ROI from the Harvard–Oxford subcortical atlas. Contrast values were extracted from spherical peak ROIs as well as from the hippocampal ROI and used to test effects of age (young vs. old) and PIB group (within the old group, controlling for age).

Statistical Analyses

All statistical analyses and plots were completed using R version 2.11 (http://www.r-project.org/, last accessed 15 August 2011). Group differences in demographic variables were determined with t-tests for continuous variables and chi-squared tests for dichotomous variables. Within the old NC group, multiple regression was used for relationships between PIB group and cognitive measures as well as fMRI activation (controlling for age). Within group correlations between cognitive measures and fMRI contrast values were computed with Spearman rank correlations (young, all old, PIB− old, and PIB+ old group separately; Spearman rank was used to account for the small number of subjects within groups). P < 0.05 was considered significant, and trends of P < 0.10 were noted.

Results

Group Characteristics

Group characteristics are listed in Table 2. Old subjects were significantly more educated than young subjects who were mainly still students (t = 2.78, P = 0.007); there was no difference in gender between old and young groups. Young subjects were more accurate than old subjects on the water/no water judgment (91% vs. 85%; t = −4.18, P < 0.0001), and there was a trend for better memory recognition accuracy (t = 1.38, P = 0.17) and better task memory performance in young versus old subjects (assessed with d-prime; t = 1.68, P = 0.10). There was no difference in recognition false-positive rate between young and old subjects. Furthermore, young subjects had faster reaction times than old subjects for the water/no water judgment (1388 vs. 1863 ms; t = −5.56, df = 58, P < 0.0001) as well as during the postscan recognition judgment (1878 vs. 3174 ms; t = −5.88, df = 58, P < 0.0001).

Table 2

Subject characteristics

 yNC PIB− oNC PIB+ oNC 
N 15 30 15 
Age 23.2 (3.6) 74.9 (6.6) 76.3 (7.9) 
Gender 9M 11M 6M 
Education+ 15.6 (1.7) 17.3 (2.0) 17.0 (2.0) 
APOE4 carriers/noncarriers NA 9/18 8/7 
Recognition accuracy 0.76 (0.25) 0.68 (0.19) 0.66 (0.27) 
Recognition false-positive rate 0.38 (0.25) 0.36 (0.19) 0.42 (0.27) 
d-prime 1.24 (0.79) 0.98 (0.51) 0.81 (0.66) 
PIB index NA 1.03 (0.04) 1.27 (0.18) 
MMSE NA 29.2 (1.1) 29.1 (1.1) 
Factor score 1 (EM) NA −0.02 (0.78) −0.04 (0.63) 
Factor score 2 (Exe) NA −0.25 (0.79) −0.29 (0.48) 
Factor score 3 (WM) NA −0.14 (0.62) −0.24 (0.74) 
Factor score 4 (VisMem) NA −0.13 (1.08) −0.17 (1.16) 
Factor score 5 (SM)* NA 0.01 (0.57) 0.46 (0.49) 
 yNC PIB− oNC PIB+ oNC 
N 15 30 15 
Age 23.2 (3.6) 74.9 (6.6) 76.3 (7.9) 
Gender 9M 11M 6M 
Education+ 15.6 (1.7) 17.3 (2.0) 17.0 (2.0) 
APOE4 carriers/noncarriers NA 9/18 8/7 
Recognition accuracy 0.76 (0.25) 0.68 (0.19) 0.66 (0.27) 
Recognition false-positive rate 0.38 (0.25) 0.36 (0.19) 0.42 (0.27) 
d-prime 1.24 (0.79) 0.98 (0.51) 0.81 (0.66) 
PIB index NA 1.03 (0.04) 1.27 (0.18) 
MMSE NA 29.2 (1.1) 29.1 (1.1) 
Factor score 1 (EM) NA −0.02 (0.78) −0.04 (0.63) 
Factor score 2 (Exe) NA −0.25 (0.79) −0.29 (0.48) 
Factor score 3 (WM) NA −0.14 (0.62) −0.24 (0.74) 
Factor score 4 (VisMem) NA −0.13 (1.08) −0.17 (1.16) 
Factor score 5 (SM)* NA 0.01 (0.57) 0.46 (0.49) 

Note: Means and SDs are reported for continuous variables. Neuropsychological factor scores were unavailable for 6 old normal controls (oNCs) (2 PIB+ and 4 PIB−) and 3 PIB− oNC were missing APOE genotyping. Significant differences between young and old subjects are denoted with a + (P < 0.05). Significant differences between PIB− and PIB+ oNC are denoted with a * (P < 0.05). yNC, young normal controls; EM, episodic memory; Exe, executive function; WM, working memory; VisMem, visuospatial memory; SM, semantic memory; and NA, not applicable.

Based on the PIB index cutoff value of 1.08, 15 old NCs were classified as PIB+ and 30 old NCs were PIB−. There were no significant differences for age, education, or gender between PIB+ and PIB− groups. Fifty-three percent of PIB+ subjects were APOE4 carriers, whereas 33% of PIB− subjects were APOE4 carriers (this difference was not significant). Controlling for age, there were no significant relationships between PIB status and MMSE or any factor score other than semantic memory (PIB− were higher than PIB+; P = 0.014). There were no differences in water/no water accuracy, recognition accuracy or false-positive rate, overall memory performance assessed with d-prime, or reaction time measures between PIB+ and PIB− subjects.

Task Activation and Deactivation Patterns in Young and Old Subjects

One-sample t-tests of young and old subjects combined are shown in Figure 1. Task activations (hits > misses) are found bilaterally in ventrolateral prefrontal (bordering dorsolateral prefrontal), lateral occipital/parietal, posterior inferior temporal, and right parahippocampal/hippocampus. Deactivations (hits < misses) were present in bilateral medial occipital, posteromedial, angular, medial prefrontal, superior/dorsolateral prefrontal, and left central/postcentral cortices. To create task-positive ROIs, the highest peak coordinate for each cluster was selected (Table 3). Furthermore, an additional task-positive ROI was created in the hippocampus by masking cluster 3 by the Harvard–Oxford subcortical atlas anatomically defined hippocampus ROI, resulting in a right hippocampus ROI 92 voxels in size (736 mm3, “RHip”). To create task-negative ROIs (Table 4), the highest peak for clusters 3 and 4 were selected. For clusters 5 and 6, local maximum were selected in addition to or rather than the highest peak since the peak coordinate for these clusters did not coincide with regions typically considered part of the DMN. We choose to focus on DMN-typical regions due to previous publications stressing the concordance between this network and Aβ deposition. However, follow-up voxelwise analyses enabled examination of all regions irrespective of criterion used to select ROIs. Specifically, for cluster 5, a local maximum in medial prefrontal cortex (mPFC) was selected (an area typically associated with the DMN) rather than the highest peak for that cluster, which was in the frontal pole. For cluster 6, the peak coordinate was in medial occipital cortex, whereas a local maximum was in precuneus (consequently, 2 ROIs were created for this cluster, “LOcc” and “Precun”). ROIs from clusters 1 and 2 were not selected since these regions are outside the DMN. Overall, this resulted in 5 task-positive and 5 task-negative ROIs for subsequent analyses.

Table 3

Activation clusters and peak coordinates

Cluster index Cluster size Z x y z Location 
9189 6.17 −30 −78 26 L lateral superior occipital (LOcc) 
6.15 −28 −90 26 L occipital pole/lateral occipital 
5.87 −24 −68 50 L lateral superior occipital 
5.68 −48 −56 −16 L inferior temporal 
5.56 −24 −78 40 L lateral superior occipital 
5.53 −34 −92 12 L lateral superior occipital 
8781 6.03 38 −86 18 R lateral superior occipital (ROcc) 
5.96 24 −38 −18 R fusiform/posterior parahippocampal gyrus 
5.86 26 −66 52 R lateral superior occipital 
5.83 42 −80 18 R lateral superior occipital 
5.72 38 −82 26 R lateral superior occipital 
5.69 26 −68 46 R lateral superior occipital 
1893 4.96 50 34 12 R inferior frontal gyrus/frontal pole (RIFG) 
4.65 44 28 R precentral gyrus/inferior frontal gyrus 
4.5 40 14 26 R inferior/middle frontal gyrus 
4.49 44 36 10 R frontal pole/inferior frontal gyrus 
4.4 36 16 26 R inferior/middle frontal gyrus 
4.35 40 32 R precentral/middle frontal gyrus 
1616 4.55 −40 8 26 L inferior frontal gyrus/precentral gyrus (LIFG) 
4.12 −52 14 38 L middle/inferior frontal gyrus 
3.74 −44 38 L frontal pole/inferior frontal gyrus 
3.37 −38 30 16 L inferior/middle frontal gyrus 
3.14 −56 32 18 L middle/inferior frontal gyrus 
Cluster index Cluster size Z x y z Location 
9189 6.17 −30 −78 26 L lateral superior occipital (LOcc) 
6.15 −28 −90 26 L occipital pole/lateral occipital 
5.87 −24 −68 50 L lateral superior occipital 
5.68 −48 −56 −16 L inferior temporal 
5.56 −24 −78 40 L lateral superior occipital 
5.53 −34 −92 12 L lateral superior occipital 
8781 6.03 38 −86 18 R lateral superior occipital (ROcc) 
5.96 24 −38 −18 R fusiform/posterior parahippocampal gyrus 
5.86 26 −66 52 R lateral superior occipital 
5.83 42 −80 18 R lateral superior occipital 
5.72 38 −82 26 R lateral superior occipital 
5.69 26 −68 46 R lateral superior occipital 
1893 4.96 50 34 12 R inferior frontal gyrus/frontal pole (RIFG) 
4.65 44 28 R precentral gyrus/inferior frontal gyrus 
4.5 40 14 26 R inferior/middle frontal gyrus 
4.49 44 36 10 R frontal pole/inferior frontal gyrus 
4.4 36 16 26 R inferior/middle frontal gyrus 
4.35 40 32 R precentral/middle frontal gyrus 
1616 4.55 −40 8 26 L inferior frontal gyrus/precentral gyrus (LIFG) 
4.12 −52 14 38 L middle/inferior frontal gyrus 
3.74 −44 38 L frontal pole/inferior frontal gyrus 
3.37 −38 30 16 L inferior/middle frontal gyrus 
3.14 −56 32 18 L middle/inferior frontal gyrus 

Note: Significant task-positive clusters and corresponding local maxima from the one-sample t-test of young and old subjects. The highest peak within each cluster was selected for subsequent ROI analyses (bolded and italicized). Additionally, cluster 3 was masked with a hippocampus anatomical mask to define an additional task-positive ROI in the right hippocampus (“RHip”). L, left; R, right.

Table 4

Deactivation clusters and peak coordinates

Cluster index Cluster size Z x y z Location 
6157 5.55 −2 −90 26 L medial occipital pole/cuneus (LOcc) 
4.98 −2 −88 30 R cuneus/occipital pole 
4.69 −72 26 L cuneus/precuneus 
4.51 −12 −90 28 L occipital pole 
4.28 0 −80 42 Precuneus (Precun) 
4.27 −82 40 L cuneus/precuneus 
3448 4.43 −20 50 26 L frontal pole 
4.12 −26 54 20 L frontal pole 
4.03 2 52 10 Paracingulate/frontal pole (mPFC) 
3.87 −22 42 22 L frontal pole 
3.65 −34 34 30 L middle frontal gyrus/frontal pole 
3.56 −30 36 28 L middle frontal gyrus/frontal pole 
3307 4.55 60 −46 32 R angular/supramarginal gyrus (RPar) 
4.25 50 −26 −12 R middle temporal gyrus 
4.2 66 −46 32 R supramarginal/angular 
3.87 58 −30 26 R parietal operculum/supramarginal 
3.73 56 −48 44 R angular/supramarginal gyrus 
3.62 58 −40 R middle temporal gyrus 
2831 3.47 −62 −46 14 L supramarginal/angular (LPar) 
3.42 −62 −40 38 L supramarginal 
3.39 −66 −36 28 L supramarginal 
3.38 −62 −26 L supramarginal 
3.36 −54 12 −4 L temporal pole 
3.31 −60 −36 24 L parietal operculum 
1840 3.91 −42 −22 50 L postcentral 
3.51 −32 −28 62 L postcentral 
3.35 −42 −38 62 L postcentral 
3.26 −46 −32 56 L postcentral 
3.15 −34 −36 68 L postcentral 
3.15 −28 −44 52 L superior parietal 
1777 4.06 26 44 30 R frontal pole 
3.65 24 56 28 R frontal pole 
3.63 30 48 26 R frontal pole 
3.39 22 58 22 R frontal pole 
3.33 36 46 28 R frontal pole 
3.29 28 56 20 R frontal pole 
Cluster index Cluster size Z x y z Location 
6157 5.55 −2 −90 26 L medial occipital pole/cuneus (LOcc) 
4.98 −2 −88 30 R cuneus/occipital pole 
4.69 −72 26 L cuneus/precuneus 
4.51 −12 −90 28 L occipital pole 
4.28 0 −80 42 Precuneus (Precun) 
4.27 −82 40 L cuneus/precuneus 
3448 4.43 −20 50 26 L frontal pole 
4.12 −26 54 20 L frontal pole 
4.03 2 52 10 Paracingulate/frontal pole (mPFC) 
3.87 −22 42 22 L frontal pole 
3.65 −34 34 30 L middle frontal gyrus/frontal pole 
3.56 −30 36 28 L middle frontal gyrus/frontal pole 
3307 4.55 60 −46 32 R angular/supramarginal gyrus (RPar) 
4.25 50 −26 −12 R middle temporal gyrus 
4.2 66 −46 32 R supramarginal/angular 
3.87 58 −30 26 R parietal operculum/supramarginal 
3.73 56 −48 44 R angular/supramarginal gyrus 
3.62 58 −40 R middle temporal gyrus 
2831 3.47 −62 −46 14 L supramarginal/angular (LPar) 
3.42 −62 −40 38 L supramarginal 
3.39 −66 −36 28 L supramarginal 
3.38 −62 −26 L supramarginal 
3.36 −54 12 −4 L temporal pole 
3.31 −60 −36 24 L parietal operculum 
1840 3.91 −42 −22 50 L postcentral 
3.51 −32 −28 62 L postcentral 
3.35 −42 −38 62 L postcentral 
3.26 −46 −32 56 L postcentral 
3.15 −34 −36 68 L postcentral 
3.15 −28 −44 52 L superior parietal 
1777 4.06 26 44 30 R frontal pole 
3.65 24 56 28 R frontal pole 
3.63 30 48 26 R frontal pole 
3.39 22 58 22 R frontal pole 
3.33 36 46 28 R frontal pole 
3.29 28 56 20 R frontal pole 

Note: Significant task-negative clusters from the one-sample t-test of young and old subjects are listed. Peaks selected for subsequent ROI analyses are bolded and italicized. L, left, R; right.

Figure 1.

Activation and deactivation statistical maps. Statistical maps are defined by one-sample t-tests across young and old subjects, covarying for task memory performance. Warm colors are activations (hits > misses), while cool colors are deactivations (hits < misses).

Figure 1.

Activation and deactivation statistical maps. Statistical maps are defined by one-sample t-tests across young and old subjects, covarying for task memory performance. Warm colors are activations (hits > misses), while cool colors are deactivations (hits < misses).

Relationships between Subsequent Memory Effects and Aβ in Old Subjects

Older subjects showed reduced activation compared with young subjects across all task-positive ROIs (P < 0.01) except the RHip ROI (P = 0.63). Controlling for age, PIB+ NCs showed significantly increased activation compared with PIB− NCs in RHip (P = 0.011) and LOcc (P = 0.019), while a trend was present in ROcc (P = 0.058; Fig. 2). An average activation value was additionally computed across all 5 task-positive ROIs and was significantly different between PIB+ and PIB− groups (Average Act; P = 0.012, Fig. 2). Results were similar after controlling for APOE status (RHip: P = 0.026; LOcc: P = 0.073; ROcc: P = 0.082; Average Act: P = 0.044).

Figure 2.

Task activation across groups. Mean contrast values (hits vs. misses) for each task-positive ROI as well as an average across the 5 ROIs. Error bars reflect the standard error estimates of each mean. *Indicates significant difference between PIB+ and PIB− groups (P < 0.05, controlling for age). LOcc, left occipital; LIFG, left inferior frontal gyrus; ROcc, right occipital; RIFG, right inferior frontal gyrus; RHip, right hippocampus; and Average Act, average activation across 5 ROIs.

Figure 2.

Task activation across groups. Mean contrast values (hits vs. misses) for each task-positive ROI as well as an average across the 5 ROIs. Error bars reflect the standard error estimates of each mean. *Indicates significant difference between PIB+ and PIB− groups (P < 0.05, controlling for age). LOcc, left occipital; LIFG, left inferior frontal gyrus; ROcc, right occipital; RIFG, right inferior frontal gyrus; RHip, right hippocampus; and Average Act, average activation across 5 ROIs.

Relationships between Task Activation and Behavioral Measures

To explore whether heightened activation is associated with cognitive processes, average task activation (across the 5 task-positive ROIs) was related to fMRI task memory performance assessed with d-prime as well as the 5 factor scores derived from neuropsychological data collected on a separate day from fMRI scanning (episodic memory, executive function, working memory, visual memory, and semantic memory). Due to the significant association with PIB+ status within old subjects, contrast values from the LOcc and RHip ROIs were also related to cognitive measures within PIB+ subjects.

Across the entire old group, within PIB− subjects, and within young subjects, there were no relationships between average task activation and task memory performance or any factor score. Within PIB+ subjects, there was a significant relationship between average task activation and the visual memory factor score (rho = 0.59, P = 0.035; Fig. 3). Furthermore, activation in LOcc was significantly related to task memory performance (rho = 0.66, P = 0.007), while activation in RHip was associated with visual memory (rho = 0.60; P = 0.031) in PIB+ subjects (Table 5 and Fig. 3). Relationships between average/regional activation and the remaining factor scores were not significant within PIB+ subjects.

Table 5

Correlations between memory, activation, and deactivation within PIB+ subjects

ROI Task memory performance Visual memory factor score 
Average activation rho = 0.29, P = 0.302 rho = 0.59, P = 0.035 
LOcc rho = 0.66, P = 0.007 rho = 0.24, P = 0.426 
RHip rho = −0.22, P = 0.428 rho = 0.60, P = 0.031 
Average deactivation rho = −0.36, P = 0.182 rho = 0.24, P = 0.426 
ROI Task memory performance Visual memory factor score 
Average activation rho = 0.29, P = 0.302 rho = 0.59, P = 0.035 
LOcc rho = 0.66, P = 0.007 rho = 0.24, P = 0.426 
RHip rho = −0.22, P = 0.428 rho = 0.60, P = 0.031 
Average deactivation rho = −0.36, P = 0.182 rho = 0.24, P = 0.426 

Note: Summary of correlations with memory measures in PIB+ subjects. Relationships were not identified with the other factor scores (episodic memory, executive function, working memory, and semantic memory). Trends and significant relationships are bolded (P < 0.1).

Figure 3.

Relationships between activation and cognitive measures. Positive relationships between fMRI contrast values and measures of memory were identified within PIB+ subjects. Specifically, activation in LOcc was correlated with memory performance during the fMRI experiment (measured with the d-prime of high-confidence recognition responses, left; P = 0.007), whereas RHip and the average activation across the 5 task-positive ROIs were correlated with the visual memory factor score (middle, right; P = 0.031 and 0.035, respectively). Values are ranked due to the small number of subjects. There were no relationships with the other factor scores or within young or PIB− old subjects.

Figure 3.

Relationships between activation and cognitive measures. Positive relationships between fMRI contrast values and measures of memory were identified within PIB+ subjects. Specifically, activation in LOcc was correlated with memory performance during the fMRI experiment (measured with the d-prime of high-confidence recognition responses, left; P = 0.007), whereas RHip and the average activation across the 5 task-positive ROIs were correlated with the visual memory factor score (middle, right; P = 0.031 and 0.035, respectively). Values are ranked due to the small number of subjects. There were no relationships with the other factor scores or within young or PIB− old subjects.

Relationships with Task Deactivation

Examination of the parameter estimates across trial types for areas showing significantly lower activation for hits than misses revealed that selected task-negative regions show greater deactivation during hits than misses relative to fixation (Fig. 4). To examine differences in parameter estimates across young and old subjects, a repeated measures analysis of variance with ROI and trial type (hits vs. misses) as within subject factors and group as a between subject factor as well as the interaction between group and trial type was conducted. This analysis revealed main effects for both ROI (F = 35.24, P < 0.001) and trial type (F = 22.29, P < 0.001) as well as a trend for the interaction between group and trial type (F = 3.6149, P = 0.058). Paired t-tests contrasting parameter estimates for average deactivation across the 5 selected ROIs in young and old groups separately revealed greater deactivation during hits relative to misses in both young (P < 0.001) and old groups (P < 0.001). Between group t-tests revealed greater deactivation in young versus old subjects during hits (P = 0.028), whereas no difference was observed for misses between old and young subjects (P = 0.42). Thus, the reduced deactivation between hits and misses observed in our old group was driven by a reduced magnitude of deactivation during hits.

Figure 4.

Trial parameter estimates for task-negative regions. Average trial parameter estimates for each task-negative ROI is plotted and reveals greater deactivation during hits than during misses relative to fixation across all subjects. Error bars reflect the standard error estimates of each mean. Although old subjects show significantly greater deactivation during hits than misses, the magnitude of deactivation during hits is reduced compared with young subjects. t-tests were performed within and between young and old subjects for the average deactivation measure, and significant differences are denoted with an asterisk. O = old, Y = young.

Figure 4.

Trial parameter estimates for task-negative regions. Average trial parameter estimates for each task-negative ROI is plotted and reveals greater deactivation during hits than during misses relative to fixation across all subjects. Error bars reflect the standard error estimates of each mean. Although old subjects show significantly greater deactivation during hits than misses, the magnitude of deactivation during hits is reduced compared with young subjects. t-tests were performed within and between young and old subjects for the average deactivation measure, and significant differences are denoted with an asterisk. O = old, Y = young.

Likewise, direct comparison of hits versus misses contrast values revealed reduced deactivations in old compared with young subjects in 2/5 examined ROIs (LOcc: P < 0.001; mPFC: P = 0.022) and in the average value across all 5 ROIs (Average deact: P = 0.041). There was no effect of PIB status on the contrast between hits versus misses in individual task-negative ROIs or on average deactivation (Fig. 5).

Figure 5.

Task deactivation across groups. Average contrast values for hits versus misses are plotted for each task-negative ROI. Error bars reflect the standard error estimates of each mean. Between group t-tests revealed reduced deactivation in old compared with young across multiple ROIs (LOcc and mPFC), but an effect of PIB status was not identified.

Figure 5.

Task deactivation across groups. Average contrast values for hits versus misses are plotted for each task-negative ROI. Error bars reflect the standard error estimates of each mean. Between group t-tests revealed reduced deactivation in old compared with young across multiple ROIs (LOcc and mPFC), but an effect of PIB status was not identified.

To ensure the null result between PIB status and task deactivations was not a consequence of ROI selection, an exploratory voxelwise analysis was conducted using permutation testing with FSL’s Randomize (with 5000 permutations, http://www.fmrib.ox.ac.uk/fsl/randomise/, last accessed 15 August 2011 restricted to the voxels that were significant in the one-sample t-test of hits < misses). Results were considered significant at a liberal threshold of P < 0.05, k = 50 (uncorrected). This analysis failed to reveal strong evidence for an effect of amyloid on deactivations in this cohort. At this liberal threshold, there were a few clusters showing reduced deactivation in PIB+ compared with PIB− subjects, as well as 2 clusters showing reduced deactivation in PIB− compared with PIB+ subjects (Table 6).

Table 6

Significant clusters and peak coordinates from exploratory analysis in task-negative regions

Cluster size Max P value x y z Location 
Less deactivation in PIB+ versus PIB− 
    163 0.0002 −56 −50 26 L supramarginal gyrus 
    91 0.001 −92 28 R medial occipital 
    73 0.003 −40 30 Posterior cingulate 
    53 0.002 −30 24 30 L middle frontal gyrus 
    50 0.001 60 −60 36 R lateral occipital 
Less deactivation in PIB− versus PIB+ 
    158 0.001 −92 Medial occipital 
    55 0.001 56 −44 26 R angular gyrus 
Cluster size Max P value x y z Location 
Less deactivation in PIB+ versus PIB− 
    163 0.0002 −56 −50 26 L supramarginal gyrus 
    91 0.001 −92 28 R medial occipital 
    73 0.003 −40 30 Posterior cingulate 
    53 0.002 −30 24 30 L middle frontal gyrus 
    50 0.001 60 −60 36 R lateral occipital 
Less deactivation in PIB− versus PIB+ 
    158 0.001 −92 Medial occipital 
    55 0.001 56 −44 26 R angular gyrus 

Note: Exploratory analysis within task-negative regions failed to reveal strong evidence for an effect of amyloid within old subjects. L, left; R, right.

Relationships between Task Deactivation, Cognition, and Task Activation

Although a strong effect of Aβ was not found in task-negative regions, we nevertheless sought to determine whether deactivations relate to memory performance in our cohort. An average task deactivation measure was calculated by averaging across the 5 selected task-negative ROIs and was related to task performance as well as the 5 neuropsychological factor scores. Across all old subjects, there was a negative relationship with task memory performance (i e., more deactivation, better performance; rho = −0.34, P = 0.021), however, this relationship did not reach significance when old groups were examined separately (PIB+ only: rho = −0.36, P = 0.182; PIB− only: rho = −0.31, P = 0.126). There were no significant relationships between average deactivation and any factor score or between task memory performance and deactivation within young subjects.

To assess independent contributions to fMRI task memory performance (de Chastelaine et al. 2011), multiple regression models were conducted with average task-negative and -positive contrast values as simultaneous predictors of task memory performance (conducted in all old subjects as well as within PIB+ and PIB− groups separately). These models revealed an independent effect of deactivation on performance across all old subjects (P = 0.038) as well as trends for dissociable effects of deactivation and activation on task memory performance within PIB+ subjects (P = 0.091 and P = 0.099, respectively; such that there was a trend for reduced deactivation in task-negative regions to be independently associated with worse performance, while there was a trend for increased task-positive activation to be independently associated with better performance, Table 7).

Table 7

Multiple regression model predicting memory performance

Model: memory performance ∼ average activation + average deactivation 
 Parameter estimate Standard error P value 
All old, N = 45 
    Average dectivation −0.0123 0.0057 0.038 
    Average activation 0.0084 0.0071 0.244 
PIB+, N = 15 
    Average dectivation −0.0162 0.0088 0.091 
    Average activation 0.0190 0.0106 0.099 
PIB−, N = 30 
    Average dectivation −0.0064 0.0081 0.441 
    Average activation –0.0002 0.0112 0.999 
Model: memory performance ∼ average activation + average deactivation 
 Parameter estimate Standard error P value 
All old, N = 45 
    Average dectivation −0.0123 0.0057 0.038 
    Average activation 0.0084 0.0071 0.244 
PIB+, N = 15 
    Average dectivation −0.0162 0.0088 0.091 
    Average activation 0.0190 0.0106 0.099 
PIB−, N = 30 
    Average dectivation −0.0064 0.0081 0.441 
    Average activation –0.0002 0.0112 0.999 

Note: Multiple regression models predicting task memory performance reveal an independent effect of task deactivation on performance across all old subjects as well as a trend for dissociable effects of deactivation and activation on performance within PIB+ subjects (reduced deactivation in task-negative regions was associated with worse performance; increased task-positive activation was associated with better performance). Significant relationships and trends are bolded (P < 0.1).

Discussion

In this study, we scanned young and old subjects with fMRI while performing incidental scene encoding, allowing isolation of brain regions associated with successful memory formation. Consistent with previous studies, we identified regions showing greater activation for successfully remembered versus subsequently forgotten scenes bilaterally in ventrolateral prefrontal, lateral occipital/parietal, posterior inferior temporal, and the right parahippocampal/hippocampus (task-positive regions). Furthermore, significant deactivations (hits < misses) were present in DMN regions (posteromedial, medial prefrontal, and lateral temporoparietal cortices) as well as in areas outside the DMN (bilateral medial occipital cortex, superior/dorsolateral prefrontal, and left central/postcentral gyri). Young subjects showed more activation in task-positive regions and greater deactivation in task-negative regions compared with old subjects. Within the old group, PIB+ subjects showed increased activation in task-positive regions compared with PIB− old NCs. Furthermore, positive relationships between higher task activation and better memory ability within PIB+ subjects suggest that these heightened activations are beneficial. Although we did not identify strong evidence for impaired deactivation in PIB+ subjects, a multiple regression approach revealed trends for dissociable effects of activation and deactivation on performance within this group, demonstrating that these networks may have independent contributions to performance among PIB+ subjects (such that greater activation was associated with better memory performance and impaired deactivation was associated with worse memory performance). Overall, our data suggest that heightened task-positive activation in PIB+ old subjects is beneficial to memory performance and that these increases are not directly related to deficits in DMN deactivation.

Heightened Task Activation in PIB+ NCs May Reflect Compensatory Processes

The pattern of elevated activation in PIB+ NC subjects is consistent with previous work showing elevated activation in elderly NCs destined for episodic memory decline (Persson et al. 2006; O'Brien et al. 2010) as well as in subjects at risk for AD (Bookheimer et al. 2000; Dickerson et al. 2005; Trivedi et al. 2008). Our results provide direct support for previous hypotheses that this hyperactivation may be a compensatory response to underlying AD pathology.

The design of this experiment allowed isolation of activation specific to successful memory. Therefore, the task-positive increases we identified in PIB+ NCs specifically occur during successful encoding. Additionally, we identified positive relationships between memory ability and heightened task activation within PIB+ subjects. Taken together, this pattern implies that heightened activation within PIB+ subjects during encoding promotes better overall memory performance that extends beyond the specific scanning session since relationships were identified with an independent measure of visual memory performed on a separate day.

Although the mechanism underlying this effect is unclear, a number of potential explanations exist. One possibility is that hyperactivity may reflect increased neuronal demand that counteracts detrimental effects of Aβ that occur at the level of synaptic function (Selkoe 2002). Another possibility is that PIB+ NC implement alternative processing strategies in the context of Aβ-related functional decline. It has been shown that heightened activation within task-positive regions is related to deep encoding (Fletcher et al. 2003) as well as encoding effort (Reber et al. 2002), which may represent mechanisms implemented by PIB+ subjects during this task. While the incidental encoding nature of our design makes it unlikely that different encoding strategies were implemented by PIB+ subjects, it is possible that heightened activation reflects these mechanisms at work during online task demands. The online demand of the current experiment was a basic visual search task, where subjects indicated whether or not water was present. A possibility is that PIB+ subjects engaged in deeper processing during visual search of presented scenes and that this deeper processing was beneficial for successful memory, whereas a similar level of processing was unnecessary for successful memory in PIB− subjects. Interestingly, the occipital cortex and hippocampus remain relatively free of Aβ deposition throughout AD development (Braak H and Braak E 1991; Thal et al. 2002), suggesting that these regions may compensate for amyloid-induced dysfunction in multimodal association areas that are highly vulnerable to Aβ deposition.

Heightened Task Activation May Precede Aβ Deposition

It is also possible that increased activation within this network predates Aβ deposition. This direction of causality is supported by the observation that Aβ release is activity dependent (Cirrito et al. 2005) and that Aβ deposition occurs in the most metabolically active areas of the brain (Buckner et al. 2005; Vlassenko et al. 2011). Furthermore, recent evidence from a mouse model of AD suggests that neuronal activity is predictive of the brain areas that are subsequently vulnerable to Aβ deposition (Bero et al. 2011). Given evidence that Aβ is released from the presynaptic terminal through exocytosis during fusion of synaptic vesicles, it is possible that hyperactivation in task-positive regions directly leads to Aβ deposition in connected multimodal association regions (Cirrito et al. 2005). The positive relationships we identified between heightened activation and behavioral measures suggest that this hyperactivation is nevertheless beneficial for these individuals who perhaps have preexisting limitations that require higher brain activity for normal cognitive function. Thus, it is possible that this activation pattern may confer early life advantages, although detrimental in the long term by promoting Aβ deposition. This is consistent with the observation that carriers of the APOE4 allelle show increased hippocampal activation during episodic memory (EM) processing in their late 20s (Filippini et al. 2009), which is likely before Aβ deposition has begun (Kok et al. 2009). Although we did not identify a relationship between APOE status and activation, the older age of our subjects makes the current study suboptimal for examining effects preceding Aβ deposition. Studies that have implicated a temporal sequence of events by investigating young APOE carriers (presumably before Aβ deposition; Mondadori et al. 2007; Filippini et al. 2009) or longitudinal brain activation preceding cognitive decline (presumably concomitant with Aβ deposition; Persson et al. 2006; O'Brien et al. 2010) have not converged to reveal a consistent mechanism underlying activation increases. The combination of amyloid imaging with fMRI will help disentangle the temporal order of these events by directly measuring underlying pathology rather than relying on proxy markers of AD risk.

Deactivations Are Impaired in Aging and Relate to Memory Performance But Not Aβ

In addition to isolating task-positive regions, we also investigated task-negative regions that show more deactivation during hits than misses. Many of the regions showing this behavior belong to the DMN, which consistently deactivates across a variety of externally driven cognitive tasks (Buckner et al. 2008). Task-negative regions were also identified outside the DMN, namely bilateral medial occipital, dorsolateral PFC, and left central gyrus. These non-DMN deactivations may be a specific response to the employed cognitive task. For instance, previous studies employing a subsequent memory paradigm have identified deactivations in dorsolateral prefrontal cortex (Daselaar et al. 2004), which may reflect reallocation of resources from this area when organization and manipulation are not involved in online task demands (Blumenfeld and Ranganath 2007). Consistent with previous studies, we found reduced deactivation in old compared with young subjects as well as a relationship between task memory performance and deactivation within old subjects (Morcom et al. 2003; Gutchess et al. 2005; Duverne et al. 2009; Kukolja et al. 2009; de Chastelaine et al. 2011).

Although we identified an effect of aging on deactivations during successful encoding, we did not find a strong relationship between deactivation failure and PIB status. This result is inconsistent with a recent study investigating Aβ in aging (Sperling et al. 2009). There are many factors that may contribute to this inconsistency. One possibility is that additional factors associated with aging contribute to deactivations, obscuring relationships between Aβ and deactivations in older subjects. For instance, precuneus deactivation has been shown to relate to levels of dopamine synthesis in elderly subjects (Braskie et al. 2010), and it is likely that age-related changes in the dopamine system are unrelated to levels of Aβ. Future studies that simultaneously measure multiple age-related brain changes may reveal distinct contributors of impaired deactivation in aging. Differences in experimental design may also contribute to the null finding in our study. For instance, Sperling et al. used face-name pairs as stimuli, and subjects were specifically instructed to encode task stimuli. It is unclear how stimuli differences and encoding intent affects relationships between deactivations and Aβ burden in aging.

Despite the lack of relationship between PIB status and deactivation in our elderly group, a multiple regression approach revealed trends for dissociable effects of task positive and negative networks on performance within PIB+ subjects. Although the ability to address this multivariate relationship in the current analysis was greatly limited by sample size, future studies with more participants will be able to determine whether these networks have independent contributions to performance amongst high PIB subjects.

Conclusions

In this study, we show that increased activation among task-positive regions is related to Aβ burden in cognitively normal elderly controls. Although the mechanisms underlying these increased activations remain unclear, occurrence during successful memory encoding and positive relationships with overall measures of memory ability suggest they are beneficial to individuals with high Aβ burden. The ability to elicit compensatory neuronal activation may allow cognitively normal elderly individuals to cope with underlying pathology and delay the onset of cognitive decline. It is also possible that heightened activation has a direct causal role in Aβ deposition in aged individuals.

Funding

National Institutes of Health (AG034570, AG032814) and Alzheimer’s Association (ZEN-08-87090).

Conflict of Interest : None declared.

References

Becker
JA
Hedden
T
Carmasin
J
Maye
J
Rentz
DM
Putcha
D
Fischl
B
Greve
DN
Marshall
GA
Salloway
S
, et al.  . 
Amyloid-beta associated cortical thinning in clinically normal elderly
Ann Neurol
 , 
2011
, vol. 
69
 (pg. 
1032
-
1042
)
Bennett
DA
Schneider
JA
Arvanitakis
Z
Kelly
JF
Aggarwal
NT
Shah
RC
Wilson
RS
Neuropathology of older persons without cognitive impairment from two community-based studies
Neurology
 , 
2006
, vol. 
66
 (pg. 
1837
-
1844
)
Bero
AW
Yan
P
Roh
JH
Cirrito
JR
Stewart
FR
Raichle
ME
Lee
JM
Holtzman
DM
Neuronal activity regulates the regional vulnerability to amyloid-beta deposition
Nat Neurosci
 , 
2011
, vol. 
14
 (pg. 
750
-
756
)
Blumenfeld
RS
Ranganath
C
Prefrontal cortex and long-term memory encoding: an integrative review of findings from neuropsychology and neuroimaging
Neuroscientist
 , 
2007
, vol. 
13
 (pg. 
280
-
291
)
Bookheimer
SY
Strojwas
MH
Cohen
MS
Saunders
AM
Pericak-Vance
MA
Mazziotta
JC
Small
GW
Patterns of brain activation in people at risk for Alzheimer's disease
N Engl J Med
 , 
2000
, vol. 
343
 (pg. 
450
-
456
)
Braak
H
Braak
E
Neuropathological stageing of Alzheimer-related changes
Acta Neuropathol
 , 
1991
, vol. 
82
 (pg. 
239
-
259
)
Braak
H
Del Tredici
K
The pathological process underlying Alzheimer's disease in individuals under thirty
Acta Neuropathol
 , 
2011
, vol. 
121
 (pg. 
171
-
181
)
Braskie
MN
Landau
SM
Wilcox
CE
Taylor
SD
O'Neil
JP
Baker
SL
Madison
CM
Jagust
WJ
Correlations of striatal dopamine synthesis with default network deactivations during working memory in younger adults
Hum Brain Mapp
 , 
2010
, vol. 
32
 (pg. 
947
-
961
)
Buckner
RL
Andrews-Hanna
JR
Schacter
DL
The brain's default network: anatomy, function, and relevance to disease
Ann N Y Acad Sci
 , 
2008
, vol. 
1124
 (pg. 
1
-
38
)
Buckner
RL
Snyder
AZ
Shannon
BJ
LaRossa
G
Sachs
R
Fotenos
AF
Sheline
YI
Klunk
WE
Mathis
CA
Morris
JC
, et al.  . 
Molecular, structural, and functional characterization of Alzheimer's disease: evidence for a relationship between default activity, amyloid, and memory
J Neurosci
 , 
2005
, vol. 
25
 (pg. 
7709
-
7717
)
Burt
C
Thomson
GH
The factorial analysis of human ability [Bk Rev.]
Br J Educ Psychol
 , 
1947
, vol. 
17
 (pg. 
40
-
48
)
Cabeza
R
Anderson
ND
Locantore
JK
McIntosh
AR
Aging gracefully: compensatory brain activity in high-performing older adults
Neuroimage
 , 
2002
, vol. 
17
 (pg. 
1394
-
1402
)
Chetelat
G
Villemagne
VL
Bourgeat
P
Pike
KE
Jones
G
Ames
D
Ellis
KA
Szoeke
C
Martins
RN
O'Keefe
GJ
, et al.  . 
Relationship between atrophy and beta-amyloid deposition in Alzheimer disease
Ann Neurol
 , 
2010
, vol. 
67
 (pg. 
317
-
324
)
Cirrito
JR
Yamada
KA
Finn
MB
Sloviter
RS
Bales
KR
May
PC
Schoepp
DD
Paul
SM
Mennerick
S
Holtzman
DM
Synaptic activity regulates interstitial fluid amyloid-beta levels in vivo
Neuron
 , 
2005
, vol. 
48
 (pg. 
913
-
922
)
Dale
AM
Optimal experimental design for event-related fMRI
Hum Brain Mapp
 , 
1999
, vol. 
8
 (pg. 
109
-
114
)
Dale
AM
Fischl
B
Sereno
MI
Cortical surface-based analysis. I. Segmentation and surface reconstruction
Neuroimage
 , 
1999
, vol. 
9
 (pg. 
179
-
194
)
Daselaar
SM
Prince
SE
Cabeza
R
When less means more: deactivations during encoding that predict subsequent memory
Neuroimage
 , 
2004
, vol. 
23
 (pg. 
921
-
927
)
de Chastelaine
M
Wang
TH
Minton
B
Muftuler
LT
Rugg
MD
The effects of age, memory performance, and callosal integrity on the neural correlates of successful associative encoding
Cereb Cortex
 , 
2011
, vol. 
21
 (pg. 
2166
-
2176
)
Delis
D
Kramer
J
Kaplan
E
Ober
B
California verbal learning test
 , 
2000
San Antonio (TX)
The Psychological Corporation
Dickerson
BC
Bakkour
A
Salat
DH
Feczko
E
Pacheco
J
Greve
DN
Grodstein
F
Wright
CI
Blacker
D
Rosas
HD
, et al.  . 
The cortical signature of Alzheimer's disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals
Cereb Cortex
 , 
2009
, vol. 
19
 (pg. 
497
-
510
)
Dickerson
BC
Salat
DH
Greve
DN
Chua
EF
Rand-Giovannetti
E
Rentz
DM
Bertram
L
Mullin
K
Tanzi
RE
Blacker
D
, et al.  . 
Increased hippocampal activation in mild cognitive impairment compared to normal aging and AD
Neurology
 , 
2005
, vol. 
65
 (pg. 
404
-
411
)
Duverne
S
Motamedinia
S
Rugg
MD
The relationship between aging, performance, and the neural correlates of successful memory encoding
Cereb Cortex
 , 
2009
, vol. 
19
 (pg. 
733
-
744
)
Filippini
N
MacIntosh
BJ
Hough
MG
Goodwin
GM
Frisoni
GB
Smith
SM
Matthews
PM
Beckmann
CF
Mackay
CE
Distinct patterns of brain activity in young carriers of the APOE-epsilon4 allele
Proc Natl Acad Sci U S A
 , 
2009
, vol. 
106
 (pg. 
7209
-
7214
)
Fischl
B
Liu
A
Dale
AM
Automated manifold surgery: constructing geometrically accurate and topologically correct models of the human cerebral cortex
IEEE Trans Med Imaging
 , 
2001
, vol. 
20
 (pg. 
70
-
80
)
Fischl
B
Salat
DH
Busa
E
Albert
M
Dieterich
M
Haselgrove
C
van der Kouwe
A
Killiany
R
Kennedy
D
Klaveness
S
, et al.  . 
Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain
Neuron
 , 
2002
, vol. 
33
 (pg. 
341
-
355
)
Fletcher
PC
Stephenson
CM
Carpenter
TA
Donovan
T
Bullmorel
ET
Regional brain activations predicting subsequent memory success: an event-related fMRI study of the influence of encoding tasks
Cortex
 , 
2003
, vol. 
39
 (pg. 
1009
-
1026
)
Fotenos
AF
Mintun
MA
Snyder
AZ
Morris
JC
Buckner
RL
Brain volume decline in aging: evidence for a relation between socioeconomic status, preclinical Alzheimer disease, and reserve
Arch Neurol
 , 
2008
, vol. 
65
 (pg. 
113
-
120
)
Gutchess
AH
Welsh
RC
Hedden
T
Bangert
A
Minear
M
Liu
LL
Park
DC
Aging and the neural correlates of successful picture encoding: frontal activations compensate for decreased medial-temporal activity
J Cogn Neurosci
 , 
2005
, vol. 
17
 (pg. 
84
-
96
)
Hardy
J
Selkoe
DJ
The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics
Science
 , 
2002
, vol. 
297
 (pg. 
353
-
356
)
Hedden
T
Van Dijk
KR
Becker
JA
Mehta
A
Sperling
RA
Johnson
KA
Buckner
RL
Disruption of functional connectivity in clinically normal older adults harboring amyloid burden
J Neurosci
 , 
2009
, vol. 
29
 (pg. 
12686
-
12694
)
Jack
CR
Jr.
Knopman
DS
Jagust
WJ
Shaw
LM
Aisen
PS
Weiner
MW
Petersen
RC
Trojanowski
JQ
Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade
Lancet Neurol
 , 
2010
, vol. 
9
 (pg. 
119
-
128
)
Kaplan
E
Goodglass
H
Weintraub
S
Boston naming test
 , 
1983
Philadelphia
Lea and Febiger
Kemppainen
NM
Aalto
S
Karrasch
M
Nagren
K
Savisto
N
Oikonen
V
Viitanen
M
Parkkola
R
Rinne
JO
Cognitive reserve hypothesis: Pittsburgh Compound B and fluorodeoxyglucose positron emission tomography in relation to education in mild Alzheimer's disease
Ann Neurol
 , 
2008
, vol. 
63
 (pg. 
112
-
118
)
Kok
E
Haikonen
S
Luoto
T
Huhtala
H
Goebeler
S
Haapasalo
H
Karhunen
PJ
Apolipoprotein E-dependent accumulation of Alzheimer disease-related lesions begins in middle age
Ann Neurol
 , 
2009
, vol. 
65
 (pg. 
650
-
657
)
Kukolja
J
Thiel
CM
Wilms
M
Mirzazade
S
Fink
GR
Ageing-related changes of neural activity associated with spatial contextual memory
Neurobiol Aging
 , 
2009
, vol. 
30
 (pg. 
630
-
645
)
Lezak
M
Neuropsychological assessment
 , 
1995
2nd ed
New York
Oxford University Press, Inc
Li
SC
Brehmer
Y
Shing
YL
Werkle-Bergner
M
Lindenberger
U
Neuromodulation of associative and organizational plasticity across the life span: empirical evidence and neurocomputational modeling
Neurosci Biobehav Rev
 , 
2006
, vol. 
30
 (pg. 
775
-
790
)
Logan
J
Fowler
JS
Volkow
ND
Wang
GJ
Ding
YS
Alexoff
DL
Distribution volume ratios without blood sampling from graphical analysis of PET data
J Cereb Blood Flow Metab
 , 
1996
, vol. 
16
 (pg. 
834
-
840
)
Logan
JM
Sanders
AL
Snyder
AZ
Morris
JC
Buckner
RL
Under-recruitment and nonselective recruitment: dissociable neural mechanisms associated with aging
Neuron
 , 
2002
, vol. 
33
 (pg. 
827
-
840
)
Lopresti
BJ
Klunk
WE
Mathis
CA
Hoge
JA
Ziolko
SK
Lu
X
Meltzer
CC
Schimmel
K
Tsopelas
ND
DeKosky
ST
, et al.  . 
Simplified quantification of Pittsburgh Compound B amyloid imaging PET studies: a comparative analysis
J Nucl Med
 , 
2005
, vol. 
46
 (pg. 
1959
-
1972
)
Mathis
CA
Wang
Y
Holt
DP
Huang
GF
Debnath
ML
Klunk
WE
Synthesis and evaluation of 11C-labeled 6-substituted 2-arylbenzothiazoles as amyloid imaging agents
J Med Chem
 , 
2003
, vol. 
46
 (pg. 
2740
-
2754
)
Mondadori
CR
de Quervain
DJ
Buchmann
A
Mustovic
H
Wollmer
MA
Schmidt
CF
Boesiger
P
Hock
C
Nitsch
RM
Papassotiropoulos
A
, et al.  . 
Better memory and neural efficiency in young apolipoprotein E epsilon4 carriers
Cereb Cortex
 , 
2007
, vol. 
17
 (pg. 
1934
-
1947
)
Morcom
AM
Good
CD
Frackowiak
RS
Rugg
MD
Age effects on the neural correlates of successful memory encoding
Brain
 , 
2003
, vol. 
126
 (pg. 
213
-
229
)
Mormino
EC
Brandel
MG
Madison
CM
Rabinovici
GD
Marks
S
Baker
SL
Jagust
W
Not quite PIB-positive, not quite PIB-negative: slight PIB elevations in elderly normal control subjects are biologically relevant
Neuroimage
 , 
2011
 
doi: 10.1016/j.neuroimage.2011.07.098.
Mormino
EC
Kluth
JT
Madison
CM
Rabinovici
GD
Baker
SL
Miller
BL
Koeppe
RA
Mathis
CA
Weiner
MW
Jagust
WJ
Episodic memory loss is related to hippocampal-mediated beta-amyloid deposition in elderly subjects
Brain
 , 
2009
, vol. 
132
 (pg. 
1310
-
1323
)
Mormino
EC
Smiljic
A
Hayenga
AO
Onami
S
Greicius
MD
Rabinovici
GD
Janabi
M
Baker
SL
Yen
IV
Madison
CM
, et al.  . 
Relationships between beta-amyloid and functional connectivity in different components of the default mode network in aging
Cereb Cortex
 , 
2011
 
doi: 10.1093/cercor/bhr025.
Morris
JC
Roe
CM
Grant
EA
Head
D
Storandt
M
Goate
AM
Fagan
AM
Holtzman
DM
Mintun
MA
Pittsburgh compound B imaging and prediction of progression from cognitive normality to symptomatic Alzheimer disease
Arch Neurol
 , 
2009
, vol. 
66
 (pg. 
1469
-
1475
)
Morris
JC
Roe
CM
Xiong
C
Fagan
AM
Goate
AM
Holtzman
DM
Mintun
MA
APOE predicts amyloid-beta but not tau Alzheimer pathology in cognitively normal aging
Ann Neurol
 , 
2010
, vol. 
67
 (pg. 
122
-
131
)
O'Brien
JL
O'Keefe
KM
LaViolette
PS
DeLuca
AN
Blacker
D
Dickerson
BC
Sperling
RA
Longitudinal fMRI in elderly reveals loss of hippocampal activation with clinical decline
Neurology
 , 
2010
, vol. 
74
 (pg. 
1969
-
1976
)
Oh
H
Mormino
EC
Madison
C
Hayenga
A
Smiljic
A
Jagust
WJ
beta-Amyloid affects frontal and posterior brain networks in normal aging
Neuroimage
 , 
2010
, vol. 
54
 (pg. 
1887
-
1895
)
Park
DC
Reuter-Lorenz
P
The adaptive brain: aging and neurocognitive scaffolding
Annu Rev Psychol
 , 
2009
, vol. 
60
 (pg. 
173
-
196
)
Persson
J
Nyberg
L
Lind
J
Larsson
A
Nilsson
LG
Ingvar
M
Buckner
RL
Structure-function correlates of cognitive decline in aging
Cereb Cortex
 , 
2006
, vol. 
16
 (pg. 
907
-
915
)
Reber
PJ
Siwiec
RM
Gitelman
DR
Parrish
TB
Mesulam
MM
Paller
KA
Neural correlates of successful encoding identified using functional magnetic resonance imaging
J Neurosci
 , 
2002
, vol. 
22
 (pg. 
9541
-
9548
)
Reitan
R
Validity of the Trail Making Test as an indication of organic brain damage
Percept Mot Skills
 , 
1958
, vol. 
8
 (pg. 
271
-
276
)
Rentz
DM
Locascio
JJ
Becker
JA
Moran
EK
Eng
E
Buckner
RL
Sperling
RA
Johnson
KA
Cognition, reserve, and amyloid deposition in normal aging
Ann Neurol
 , 
2010
, vol. 
67
 (pg. 
353
-
364
)
Rosen
AC
Prull
MW
O'Hara
R
Race
EA
Desmond
JE
Glover
GH
Yesavage
JA
Gabrieli
JD
Variable effects of aging on frontal lobe contributions to memory
Neuroreport
 , 
2002
, vol. 
13
 (pg. 
2425
-
2428
)
Rowe
CC
Ellis
KA
Rimajova
M
Bourgeat
P
Pike
KE
Jones
G
Fripp
J
Tochon-Danguy
H
Morandeau
L
O'Keefe
G
, et al.  . 
Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging
Neurobiol Aging
 , 
2010
, vol. 
31
 (pg. 
1275
-
1283
)
Savva
GM
Wharton
SB
Ince
PG
Forster
G
Matthews
FE
Brayne
C
Age, neuropathology, and dementia
N Engl J Med
 , 
2009
, vol. 
360
 (pg. 
2302
-
2309
)
Segonne
F
Dale
AM
Busa
E
Glessner
M
Salat
D
Hahn
HK
Fischl
B
A hybrid approach to the skull stripping problem in MRI
Neuroimage
 , 
2004
, vol. 
22
 (pg. 
1060
-
1075
)
Selkoe
DJ
Alzheimer's disease is a synaptic failure
Science
 , 
2002
, vol. 
298
 (pg. 
789
-
791
)
Sheline
YI
Raichle
ME
Snyder
AZ
Morris
JC
Head
D
Wang
S
Mintun
MA
Amyloid plaques disrupt resting state default mode network connectivity in cognitively normal elderly
Biol Psychiatry
 , 
2009
, vol. 
67
 (pg. 
584
-
587
)
Sperling
RA
Aisen
PS
Beckett
LA
Bennett
DA
Craft
S
Fagan
AM
Iwatsubo
T
Jack
CR
Kaye
J
Montine
TJ
, et al.  . 
Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging and the Alzheimer's Association workgroup
Alzheimers Dement
 , 
2011
, vol. 
7
 (pg. 
280
-
292
)
Sperling
RA
Laviolette
PS
O'Keefe
K
O'Brien
J
Rentz
DM
Pihlajamaki
M
Marshall
G
Hyman
BT
Selkoe
DJ
Hedden
T
, et al.  . 
Amyloid deposition is associated with impaired default network function in older persons without dementia
Neuron
 , 
2009
, vol. 
63
 (pg. 
178
-
188
)
Stern
Y
Cognitive reserve and Alzheimer disease
Alzheimer Dis Assoc Disord
 , 
2006
, vol. 
20
 (pg. 
S69
-
S74
)
Storandt
M
Mintun
MA
Head
D
Morris
JC
Cognitive decline and brain volume loss as signatures of cerebral amyloid-beta peptide deposition identified with Pittsburgh compound B: cognitive decline associated with Abeta deposition
Arch Neurol
 , 
2009
, vol. 
66
 (pg. 
1476
-
1481
)
Thal
DR
Rub
U
Orantes
M
Braak
H
Phases of A beta-deposition in the human brain and its relevance for the development of AD
Neurology
 , 
2002
, vol. 
58
 (pg. 
1791
-
1800
)
Trivedi
MA
Murphy
CM
Goetz
C
Shah
RC
Gabrieli
JD
Whitfield-Gabrieli
S
Turner
DA
Stebbins
GT
fMRI activation changes during successful episodic memory encoding and recognition in amnestic mild cognitive impairment relative to cognitively healthy older adults
Dement Geriatr Cogn Disord
 , 
2008
, vol. 
26
 (pg. 
123
-
137
)
Vannini
P
Hedden
T
Becker
JA
Sullivan
C
Putcha
D
Rentz
D
Johnson
KA
Sperling
RA
Age and amyloid-related alterations in default network habituation to stimulus repetition
Neurobiol Aging
 , 
2011
 
doi: 10.1016/j.neurobiolaging.2011.01.003.
Villemagne
VL
Pike
KE
Chetelat
G
Ellis
KA
Mulligan
RS
Bourgeat
P
Ackermann
U
Jones
G
Szoeke
C
Salvado
O
, et al.  . 
Longitudinal assessment of Abeta and cognition in aging and Alzheimer disease
Ann Neurol
 , 
2011
, vol. 
69
 (pg. 
181
-
192
)
Vlassenko
AG
Vaishnavi
SN
Couture
L
Sacco
D
Shannon
BJ
Mach
RH
Morris
JC
Raichle
ME
Mintun
MA
Spatial correlation between brain aerobic glycolysis and amyloid-beta (Abeta) deposition
Proc Natl Acad Sci U S A
 , 
2011
, vol. 
107
 (pg. 
17763
-
17767
)
Walsh
DM
Selkoe
DJ
A beta oligomers—a decade of discovery
J Neurochem
 , 
2007
, vol. 
101
 (pg. 
1172
-
1184
)
Wechsler
D
Wechsler memory scale-revised
 , 
1987
San Antonio (TX)
The Psychological Corporation
Zec
R
The stroop color-word test: a paradigm for procedural learning
Arch Clin Neuropsychol
 , 
1986
, vol. 
1
 (pg. 
274
-
275
)