NeuroImage: Clinical 24 (2019) 102097 Contents lists available at ScienceDirect NeuroImage: Clinical journal homepage: www.elsevier.com/locate/ynicl Decreased meta-memory is associated with early tauopathy in cognitively T unimpaired older adults Patrizia Vanninia,c,d,⁎, Federico d'Oleire Uquillasc, Heidi I.L. Jacobse, Jorge Sepulcree, Jennifer Gatchelb,c, Rebecca E. Amariglioc,d, Bernard Hanseeuwa,e,f, Kathryn V. Pappa,c, Trey Heddeni, Dorene M. Rentzc,d, Alvaro Pascual-Leoneg,h, Keith A. Johnsonc,d,e, Reisa. A. Sperlinga,b,c,d a Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02114, USA b Department of Psychiatry, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02114, USA c Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02115, USA d Center for Alzheimer Research and Treatment, Department of Neurology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA e Department of Radiology, Division of Molecular Imaging and Nuclear Medicine, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02114, USA f Department of Neurology, Cliniques Universitaires Saint-Luc, Institute of Neurosciences, Université Catholique de Louvain, 10 Av. Hippocrate, 1200 Brussels, Belgium g Berenson-Allen Center for Noninvasive Brain Stimulation and Cognitive Neurology Unit, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02115, USA h Institut Guttmann, Universitat Autonoma, Barcelona, Spain i Icahn School of Medicine at Mount Sinai, New York, USA ARTICLE INFO ABSTRACT Keywords: The ability to accurately judge memory efficiency (meta-memory monitoring) for newly learned (episodic) in- Alzheimer's disease formation, is decreased in older adults and even worse in Alzheimer's disease (AD), whereas no differences have Amyloid been found for semantic meta-memory. The pathological substrates of this phenomenon are poorly understood. Feeling-of-knowing Here, we examine the association between meta-memory monitoring for episodic and semantic information to Meta-memory the two major proteinopathies in AD: amyloid (Aβ) and tau pathology in a group of cognitively unimpaired older Tau adults. All participants underwent multi-tracer PET and meta-memory monitoring was assessed using a feeling- of-knowing (FOK) task for non-famous (episodic) and famous (semantic) face-name pairs. Whole brain voxel- wise correlations between meta-memory and PET data were conducted (controlling for memory), as well as confirmatory region-of-interest analyses. Participants had reduced episodic FOK compared to semantic FOK. Decreased episodic FOK was related to tauopathy in the medial temporal lobe regions, including the entorhinal cortex and temporal pole, whereas decreased semantic FOK was related to increased tau in regions associated with the semantic knowledge network. No association was found with Aβ-pathology. Alterations in the ability to accurately judge memory efficiency (in the absence of memory decline) may be a sensitive clinical indicator of AD pathophysiology in the pre-symptomatic phase. 1. Introduction may involve engaging in compensatory or mnemonic strategies (con- trol) that can help them avoid forgetting. Past research suggests that our The ability to accurately monitor and control one's cognitive abil- meta-memory abilities (i.e. the processes involved in memory self- ities, known as meta-cognition (Nelson and Narens, 1990), guides much monitoring) are affected by the aging process (Souchay et al., 2007; of human behavior. Meta-cognition plays an important role throughout Morson et al., 2015) and deteriorate further along the prodromal life: in early childhood when we learn new skills, or in old age when (Perrotin et al., 2007) and AD dementia stages (Souchay et al., 2002a,b; cognitive abilities decline. In the early stages of Alzheimer's disease Rosen et al., 2014). In AD, poor meta-cognition also facilitates ano- (AD), a time when many individuals begin to experience difficulties sognosia (i.e. impaired awareness of their disease) (Babinski, 1914), a with memory, ongoing judgments (self-monitoring) of material in condition with far-reaching clinical implications for the individual. For memory can help individuals change and adapt their behavior. This example, anosognosia is associated with diminished treatment ⁎ Corresponding author at: Massachusetts General Hospital, 149 13th Street, Suite 10.024, Charlestown, MA 02129, USA. E-mail address: patrizia@bwh.harvard.edu (P. Vannini). https://doi.org/10.1016/j.nicl.2019.102097 Received 29 July 2019; Received in revised form 9 November 2019; Accepted 15 November 2019 Available online 18 November 2019 2213-1582/ © 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 adherence, increased engagement in high-risk situations, and increased cognitively unimpaired older adults, as well as by investigating the caregiver burden (Turro-Garriga et al., 2013; Turro-Garriga et al., association of FOK to the two major proteinopathies in AD (Aβ and tau) 2016). Although there has been extensive work on the behavioral using multi-tracer PET. characterization of meta-memory, our knowledge of the pathological substrates underlying meta-memory disturbance in aging and AD re- 2. Methods and materials mains very limited. Studies of meta-memory typically ask the person to predict their 2.1. Participants memory performance on a given task. Their predictions are then com- pared to their actual memory performance, thus forming an estimate of A total of 102 cognitively unimpaired older adults from the Harvard how well the person is aware of their own memory. One of the most Aging Brain Study (HABS) (Dagley et al., 2015) underwent 18F Flor- studied meta-memory processes in aging and AD is the feeling-of- taucipir imaging (FTP-PET) for tau and C11 Pittsburg Compound-B knowing, FOK (Hart, 1965; Nelson and Narens, 1990) in which pre- (PiB) PET imaging for Aβ, as well as a functional magnetic resonance dictions are made about the likelihood of subsequent recognition of imaging (fMRI) task to investigate meta-memory at the Massachusetts currently non-recallable information. FOK judgments can be assessed General Hospital (MGH) and Athinoula A. Martinos Center for Biome- using either semantic (Hart, 1965; Nelson and Narens, 1990) or re- dical Imaging, Boston, MA. Both PET procedures were performed cently learned episodic information (Schacter, 1983; Souchay et al., within 6 months of the FOK MRI task [PiB-PET: 0.2 years (S.D. = 0.29), 2000). A large number of studies have confirmed that subjective con- FTP-PET: 0.31 years (S.D. = 0.62)]. Participants completed an ex- fidence is predictive of objective memory performance (e.g. Hart, 1965; tensive neurocognitive testing battery (see (Dagley et al., 2015) for full Freedman and Landauer, 1966; Hart, 1967; Gardiner et al., 1973; list of tests), including the selective reminding test (SRT) (Grober et al., Gruneberg and Monks, 1974; Schacter, 1983; Nelson, 1984; 2000) to test episodic memory, and the category fluency test (CAT) Leonesio and Nelson, 1990; Kelemen et al., 2000; Nelson and (Benton, 1968) to test semantic fluency. Inclusion criteria to participate Narens, 1990). For instance, in two subsequent studies that tested FOK in the study were: a Mini-Mental State Examination (MMSE) using face-name association tasks, one study by Chua and colleagues (Folstein et al., 1975) score of 27–30 (inclusive with educational ad- demonstrated that young individuals’ FOK predictions on episodic or justment); and a global Clinical Dementia Rating (CDR) (Morris, 1993) non-famous faces and names were accurate (Chua et al., 2009), and the score of 0. Exclusionary criteria were: history of neurologic or major other by Hosey and colleagues further demonstrated that FOK predic- psychiatric disorder, contraindications for MRI; severe cardiovascular tions on semantic or famous faces and names are accurate in young disease; alcohol or substance abuse; or known cerebrovascular disease individuals (Hosey et al., 2009). However, these two meta-memory [as determined by a Hachinski Ischemic Score (Rosen et al., 1980) processes seem to be differentially affected in age and disease. While higher than 4 and/or presence of cortical infarct; multiple lacunar semantic FOK is preserved in aging (e.g. Bäckman and Karlsson, 1985; strokes; or extensive white matter hyperintensities on structural MRI]. Backman and Lipinska, 1993; Lipinska and Bäckman, 1996; Allen- Two participants were removed from the analysis due to having no Burge and Storandt, 2000; Marquie and Huet, 2000; Souchay et al., errors of omission, meaning that no meta-memory data could be ana- 2007), episodic FOK has been shown to decline with old age lyzed for these participants. Demographic information for remaining (Souchay et al., 2007; Morson et al., 2015) and further decrease in mild individuals is detailed in Table 1. This study was approved by and cognitive impairment (prodromal stage of AD) (Perrotin et al., 2007; conducted under the auspices of the Partners Human Research Com- Anderson and Scmitter-Edgecomb, 2010) and AD dementia mittee at the BWH and MGH (Boston, MA). Every participant provided (Souchay et al., 2003; Souchay and Moulin, 2013; Rosen et al., 2014), informed written consent prior to experimental procedures. although one study showed decreased semantic FOK but not episodic FOK in mild AD (Pappas et al., 1992). 2.2. Feeling-of-knowing task While the pathological basis of altered meta-memory processes re- mains elusive, recent neuroimaging findings reveal associations with The meta-memory task consisted of a modified version of the Recall- brain regions implicated in internally directed cognition or self-refer- Judgment-recognition (RJR) paradigm, originally proposed by Dr. ential processing (Northoff et al., 2006; van der Meer et al., 2010) and Joseph Hart (Hart, 1965). The task was created as an event-related self-awareness (Zamboni and Wilcock, 2011), including frontal, lateral fMRI task. Given that this study focused on the FOK behavioral re- temporal, and medial parietal regions. Studies investigating FOK found sponses in relation to PET data, only details of the FOK task itself will be that less accurate meta-memory in a group of cognitively diverse older adults was associated with reduced cortical thickness in the posterior Table 1 cingulate and medial prefrontal cortex (Bertrand et al., 2018) as well as Demographic, clinical and cognitive variables. in the right insula (Cosentino et al., 2015). Interestingly, these brain Mean and S.D. Range regions are vulnerable to the hallmark pathologies of AD, including extracellular amyloid-beta (Aβ) in the association cortex of temporal, Age (years) 76.1 (6.1) 64.8–92.3 parietal, and frontal lobes (Klunk et al., 2004; Engler et al., 2006; Gender (%) = F 63 NA Education (years) 16.2 (3.2) 6–20 Kemppainen et al., 2006; Mintun et al., 2006; Edison et al., 2007; CDR Sum-of-Boxes 0.2 (0.4) 0–2 Forsberg et al., 2008). Further, in initial stages of the AD pathophy- MMSE (unadjusted score) 29.3 (1.1) 25–30 siologic process (Braak I-III), intraneuronal neurofibrillary tangles (tau) Depression (GDS score) 3.8 (3.9) 0–21 are found concentrated in medial temporal lobe structures (Braak and Episodic memory (SRT delayed recall) 6.7 (3.4) 0–12 Semantic memory (CAT (n = 97)) 43.9 (11.6) 21–80 Braak, 1991, 1995). An important question that remains unsolved is APOE status (%) (n = 93) = e4+ 31 NA whether early evidence of tau and/or Aβ pathology may be related to Amyloid status (%) = positive 33 NA changes in meta-memory monitoring in older individuals. Moreover, as FR (%) Episodic 0.6 (0.3) 0–1.0 previous FOK studies primarily utilized episodic information to in- Semantic 0.5 (0.3) 0–1.0 vestigate FOK, it is unknown whether the pathological substrates un- FCR (%) Episodic 0.5 (0.2) 0–1.0 Semantic 0.4 (0.2) 0.1–0.8 derlying deficits in FOK are the same or different for episodic and se- mantic information. Mean and standard deviation (S.D.). CDR = clinical dementia rating; The present study aimed to answer these questions and extend GDS = Geriatric depression scale; MMSE = Mini-Mental State Examination. previous findings by investigating whether there is a behavioral dif- FR = Free recall, FCR = forced choice recognition; SRT = selective reminding ference in FOK for episodic and semantic information in a group of test; CAT = category fluency test. 2 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 presented here. The RJR paradigm was a face–name association task Rentz et al., 2010) with a Siemens/CTI ECAT HR scanner. Synthesis, executed in consecutive phases: encoding (for episodic stimuli) and preparation and administration of the PET tracers were done according recall (both episodic and semantic stimuli) was performed in the to MGH Radioactive Drug Research Committee-approved protocols scanner, and judgment and recognition (both episodic and semantic (Shoup et al., 2013). stimuli) was performed outside the scanner. For measuring Aβ deposition, acquisition of Pittsburgh Compund-B Episodic stimuli consisted of 75 face photographs of non-famous (PiB)-PET imaging data was done following a transmission scan, where individuals, with a fictional first name printed in white Times New a 10–15 mCi 11C-PiB bolus was injected intravenously, immediately Roman 36-point font underneath. During the encoding phase, partici- followed by a 60-min dynamic PET scan in 3-D mode (63 image planes, pants were explicitly told to try and remember the name associated 15.2 cm axial field of view, 5.6 mm transaxial resolution and 2.4 mm with each face and indicate with a button press whether they thought slice interval; 69 frames: 12 × 15 s, 57 × 60 s). the name was a good fit, not a good fit, or if they were uncertain. This For measuring tau, acquisition of Flortaucipir (FTP)-PET (also was a purely subjective decision. The rating of how well a name fits to a known as 18 F-labeled AV1451 or T807) was done using a 3D list mode face has been shown to enhance encoding processes (Sperling et al., dynamic protocol on the PET camera described above. Images were 2003). The experiment consisted of three encoding runs. In each run, acquired from 80 to 100 min in 4 × 5-min frames after a mean (SD) participants viewed 25 face-name pair stimuli, each shown for 2.75 s , bolus injection of 10.0 (1.0) mCi. and which were presented in a pseudorandom order in groups of five All PET data were reconstructed, attenuation-corrected, evaluated face-name pairs. Each of the 25 face-name pairs were repeated twice for head motion, and co-registered to the corresponding T1 image for (with a mean delay of 15 s between first and second presentations) each participant using 6 df rigid body registration. For PiB-PET and during the encoding task. Face-name pairs were randomly intermixed FTP-PET, cerebellar gray matter was used as the reference region from with fixation trials of a white crosshair (+) centered on a black back- the FreeSurfer Desikan-Killiany atlas as previously described ground. Participants were told to focus their attention on the fixation (Becker et al., 2011; Johnson et al., 2016). FTP-PET measures were cross during presentation of the crosshair. Each encoding run lasted for computed as the standardized uptake value ratios (SUVRs) and for PiB- 5.7 min. PET, a summary distribution volume ratio (DVR) was used. For PiB-PET After each encoding run, participants were asked to recall the names and FTP-PET surface analyses, we performed a partial volume correc- for these previously (n = 25) encoded items as well as recall the names tion by using the Müller-Gartner correction, estimating the signal in of (n = 25) famous people (recall phase). In total, there were 3 recall each grey matter voxel, as implemented in FreeSurfer software (version phases, and the total amount of semantic stimuli consisted of 75 pic- 6.0; https://surfer.nmr.mgh.harvard.edu/). For confirmatory Regions- tures of face photographs of famous people. Each face was presented for of-Interest (ROI) analyses, we performed partial volume correction on 4 s and the participant was instructed to try and recall the first name of PET data by using the geometric transform matrix method the person within the presentation time, and press a button as quickly (Rousset et al., 1998) as implemented in FreeSurfer software (version as possible to indicate whether they remembered or did not remember 6.0; https://surfer.nmr.mgh.harvard.edu/) and described in each stimuli. (Greve et al., 2016), providing one mean binding value per region. Outside the scanner, all the non-remembered face stimuli (both fa- Partial volume correction processing was performed assuming a uni- mous and non-famous) were presented again to the participants, and form 6-mm point spread function. they were then asked to judge the likelihood of recognizing the name Neocortical PiB retention was assessed using a large cortical ROI (feeling-of-knowing, FOK) on an upcoming test (judgment phase). aggregate that included frontal, lateral temporal and retrosplenial They ranked their degree of FOK using one of 3 levels from 1 (maximum cortices (FLR) and used as a variable when investigating the contribu- degree of FOK) to 3 (minimum degree of FOK). We tested the accuracy tion of amyloid to the association between meta-memory and tau pa- of participant's FOK using a forced-choice-recognition (FCR) test with thology. three-alternatives as well as a ‘don't know’ option for all the FOK trials (recognition phase). For the recalled trials, the participants had a free 2.5. Statistical analyses recall (FR) test of each face image. The behavioral outcome (FR and FCR) of the task are presented in Table 1. One sample t-tests were used to test whether the mean for episodic The paradigm was designed and generated on an external (Macbook respectively semantic FOK were significantly different from zero, in- pro®) computer using JAVA software. Responses were collected using dicating whether meta-memory monitoring was above chance accuracy an MR compatible fiber-optical key press device with 3 buttons held in or not. A paired sample t-tests was performed to test whether there was the right hand, and responses were recorded by a computer interfaced a behavioral difference for episodic and semantic FOK in the in- with the optical switch outside the scanner room. Participants received dividuals. Pearson correlation was used to investigate the relationship detailed oral instructions prior to each run and completed a practice between episodic and semantic FOK, as well as the association between session both inside and outside of the MR-scanner before the experi- FOK with demographic and clinical variables. ment began. Whole-brain vertex-vise correlations were used to determine meta- memory-associated tau and Aβ deposition separately for episodic or 2.3. Relative accuracy of judgments semantic FOK. The maps were created by taking the participants’ image-mapping native PET images to the fs average surface in To investigate whether participants could discriminate between FreeSurfer and smoothing with the equivalent of an 8 mm Gaussian their FOK responses, a Goodman–Kruskal gamma was calculated for kernel. An uncorrected threshold (p < 0.01) was applied to the maps, each participant, which indicates the strength of association between and maps were first unadjusted, and then adjusted for age, education, ordinal variables by examining item-by-item associations between and memory performance. Additional analyses were also performed (FOK) predictions and memory (FCR) performance. A score ap- adjusting for either meta-memory or amyloid pathology in the maps. proaching −1 indicates discordance between prediction and outcome, Analyses were conducted using MATLAB (generalized linear model whereas scores approaching +1 indicate concordance. scripts can be accessed at http://mrtools.mgh.harvard.edu/). To confirm the results found in the surface based analyses we per- 2.4. PET acquisition formed sensitivity analyses using an ROI approach: using the average FTP SUVR within individual corresponding Desikan–Killiany atlas re- All PET imaging took place at the PET facility at MGH and acquired gions (using FreeSurfer) that overlapped with regions identified in the as previously described (Johnson et al., 2007; Gomperts et al., 2008; previous exploratory whole-brain map approach. A series of linear 3 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 Fig. 1. Relative accuracy of FOK judgements for episodic and semantic information regressions were then conducted to determine the influence of tau on memory (using SRT delayed recall) (r = 0.01, p = 0.92), age meta-memory using these independent Desikan–Killiany regions. These (r = −0.02, p = 0.82), or MMSE scores (r = 0.19, p = 0.06). There models were both unadjusted and adjusted for age, education, and was no sex difference in meta-memory monitoring for semantic in- memory performance. Multiple comparisons for these models were formation (t98 = −1.51, p = 0.14). We found a significant association accounted for using a Bonferroni correction method. with education (r = 0.25, p = 0.01), such that individuals with higher education had better resolution of the semantic gamma score. 3. Results 3.3. Association between meta-memory and amyloid pathology 3.1. Relative accuracy of judgments Using a whole brain voxel-vise analysis we did not observe any We found that the mean for both episodic and semantic FOK showed significant association for either episodic or semantic FOK and PiB-PET overall concordance (Fig. 1). Both were significantly greater than zero binding. (episodic FOK, t99 = −4.7, p < 0.001 and semantic FOK t99 = −16.1, p < 0.001) indicating that meta-memory monitoring was above chance 3.4. Association between episodic meta-memory and tau pathology accuracy. However, a paired sample t-test demonstrated that partici- pants were significantly better at meta-memory monitoring for se- We found that lower meta-memory monitoring for episodic in- mantic relative to episodic information (t99 = 6.8, p < 0.001). formation was related to greater FTP binding in medial temporal lobe A significant association was found between episodic and semantic regions including the temporal pole and entorhinal cortex, in addition FOK (r = 0.38, p < 0.001), indicating that decreased resolution of the to the superior temporal cortex (Fig. 2A and B). These effects remained episodic FOK was related to decreased resolution of the semantic FOK. significant after controlling for age, education, accuracy on the FCR The mean of both episodic and semantic FOK showed overall con- task, and accuracy on an episodic memory task (SRT delayed recall; cordance. A one sample t-test demonstrated that both meta-memory Supplemental Figure 1 A). No significant positive associations be- processes were above chance. A paired sample t-test however demon- tween meta-memory monitoring for episodic information and FTP-PET strated that meta-memory was better for semantic as compared to binding were found. episodic information, *** = p < 0.001. Whole brain vertex-wise analyses visualizing the association be- tween greater FTP-PET binding and lower meta-memory for episodic 3.2. Association of FOK with demographics and clinical variables (a) and semantic (c) information. Maps are unadjusted and corrected at p = 0.01. The color bar represents t-values. Scatterplots showing the Meta-memory monitoring for episodic information was not sig- association between greater FTP-PET binding and lower meta-memory nificantly associated with education (r = 0.12, p = 0.12), episodic for episodic (b) and semantic (d) information in representative regions- memory (using the SRT delayed recall) (r = 0.16, p = 0.12) or se- of-interests. mantic memory (using CAT fluency task) (r = 0.15, p = 0.13). There was no sex difference in meta-memory for episodic information 3.5. Controlling for semantic meta-memory (t98 = 0.35, p = 0.73). However, we did find a significant association between meta-memory for episodic information and age (r = -0.23, To investigate whether association between episodic meta-memory p = 0.03) as well as MMSE score (r = 0.28, p = 0.004), such that older and tau was independent of semantic meta-memory, we ran additional age as well as lower MMSE scores were related to lower resolution of whole-brain vertex-wise analyses controlling for semantic meta- the episodic gamma score. Similarly, meta-memory monitoring for se- memory. Our findings remained the same (see Supplemental Fig. 1 B) mantic information was not significantly associated with semantic such that lower meta-memory monitoring for episodic information was memory (using the CAT fluency task) (r = 0.05, p = 0.61), episodic related to greater FTP binding in the medial temporal lobe regions, the 4 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 Fig. 2. The association between tau pathology and meta-memory accuracy. temporal pole and superior temporal cortex. medial orbitofrontal cortex demonstrated reduced spatial extent as compared to our original maps. 3.6. Investigating the contribution of amyloid pathology on the association between episodic meta-memory and tau pathology 3.9. Investigating the contribution of amyloid pathology on the association between semantic meta-memory and tau pathology To further investigate the contribution of amyloid pathology on our findings, we ran a whole-brain vertex-wise analyses controlling for To further investigate the contribution of amyloid pathology on our amyloid pathology from an aggregate region (see methods). Our find- findings, we ran a whole-brain vertex-wise analyses controlling for ings remained the same (see Supplemental Fig. 1 C) such that lower amyloid pathology from an aggregate region (see Methods). Again, our meta-memory monitoring for episodic information was related to findings remained largely the same (see Supplemental Fig. 1 C). In greater FTP binding in the medial temporal lobe regions, including the particular, lower meta-memory monitoring for episodic information temporal pole and superior temporal cortex. was related to greater FTP binding in the superior parietal lobe and inferior parietal lobe. 3.7. Association between semantic meta-memory and tau pathology 3.10. Regions-of-interest analysis We found that lower semantic FOK was related to greater FTP binding in several neocortical brain regions including the superior Confirmatory ROI analyses were performed by extracting the tau parietal lobe and inferior parietal lobe such as the angular gyrus signal within individual corresponding Desikan–Killiany atlas regions (Fig. 2C and D). The temporal and medial orbitofrontal cortex was also (using FreeSurfer) that overlapped with regions identified in the whole- significantly associated with greater FTP-PET binding. These effects brain map approach (see Fig. 3A and B). remained significant after controlling for age, education, accuracy on We could confirm the negative association found between meta- the FCR task, and accuracy on an semantic memory task (category memory for episodic information and FTP binding in regions in the fluency; Supplemental Fig. 1 A. No significant positive associations right hemisphere including the temporal pole, right entorhinal cortex, were found between meta-memory monitoring for semantic informa- and right superior temporal cortex (see example in Fig. 2B). An asso- tion and FTP-PET binding. ciation between FTP-PET binding in the right inferior temporal cortex and meta-memory for episodic information was just above the corrected 3.8. Controlling for episodic meta-memory p-value threshold (Table 2). Maps showing the Desikan–Killiany atlas regions (using FreeSurfer) To investigate whether the association between semantic meta- used in the regions-of-interests analyses for episodic (A) and semantic memory and tau was independent of episodic meta-memory, we ran (B) FOK. additional whole-brain vertex-wise analyses controlling for semantic The ROI analyses for meta-memory for semantic information re- meta-memory. Our findings remained largely the same (see vealed weak relationships between FOK and FTP-PET binding across 6 Supplemental Fig. 1 B) such that lower meta-memory monitoring for ROIs. While none survived adjusting for multiple comparisons, an as- semantic information was related to greater FTP binding in the superior sociation with FTP-PET binding in the right temporal pole was just parietal lobe and inferior parietal lobe. The findings in the temporal and above the significance threshold (Table 2). 5 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 Fig. 3. Regions-of-interests. Table 2 When we also conducted the episodic FOK models while correcting Association between FOK and tau pathology in regions-of-interest. for semantic FOK, the association with FTP binding in the temporal pole Region-of-interest B SE B ß t p remained significant (beta = -0.25, p =0.01. No semantic FOK model remained significant after adjusting for episodic FOK. Episodic FOK Temporal pole BH −0.83 0.25 −0.33 −3.28 0.001 4. Discussion RH −0.54 0.17 −0.32 −3.15 0.002 LH −0.59 0.26 −0.23 −2.27 0.025 Entorhinal BH −0.37 0.17 −0.24 −2.14 0.035 This work investigated the association between meta-memory RH −0.40 0.13 −0.32 −2.99 0.004 monitoring for episodic and semantic information to the two major LH −0.16 0.18 −0.12 −0.87 0.39 proteinopathies in AD: amyloid (Aβ) and tau pathology in a group of Superior temporal BH −0.71 0.43 −0.18 −1.65 0.10 RH −0.97 0.37 −0.27 −2.67 0.009 cognitively unimpaired older adults. We found that while older adults LH −0.16 0.38 −0.05 −0.43 0.67 were overall concordant in their FOK accuracy, they displayed reduced Inferior temporal BH −0.39 0.30 −0.15 −1.29 0.20 accuracy of FOK when predictions were made on an episodic task as RH −0.61 0.24 −0.28 −2.52 0.013 compared to when predictions were made on a semantic task. LH 0.05 0.29 0.02 0.19 0.85 Furthermore, lower feeling-of-knowing for both episodic and semantic Lingual −0.64 0.42 −0.17 −1.52 0.13 RH −0.25 0.39 −0.07 −0.65 0.52 information were related to tauopathy, but in different brain regions. LH −0.73 0.35 −0.22 −2.1 0.04 The different spatial patterns in the association between episodic and Semantic FOK semantic FOK with tau pathology may highlight the relative contribu- Temporal pole BH −0.40 0.18 −0.23 −2.23 0.025 tions of episodic and semantic processes to the genesis of FOK. We RH −0.31 0.12 −0.26 −2.67 0.009 LH −0.15 0.19 −0.08 −0.78 0.44 found no associations with Aβ pathology. Inferior parietal BH −0.43 0.27 −0.17 −1.58 0.12 Overall, these findings may help to better understand the mechan- RH −0.34 0.25 −0.15 −1.36 0.18 isms underlying meta-memory deficits in aging and AD. LH −0.38 0.26 −0.16 −1.47 0.14 Consistent with previous findings we found that older adults have Superior parietal BH −0.33 0.23 −0.15 −1.41 0.16 preserved overall accuracy of FOK, but that older adults have reduced RH −0.52 0.19 −0.26 −2.59 0.01 LH −0.01 0.22 −0.01 −0.06 0.95 FOK accuracy when predictions are made on an episodic task as com- Isthmus cingulate BH −0.45 0.27 −0.18 −1.65 0.10 pared to FOK accuracy when predictions are made on a semantic task RH −0.13 0.24 −0.06 −0.53 0.59 (Souchay et al., 2007). This dissociation has been proposed to be a LH −0.59 0.25 −0.25 −2.40 0.018 result from the fact that episodic and semantic FOK draw upon different Rostral anterior cingulate BH −0.43 0.19 −0.22 −2.16 0.033 RH −0.39 0.19 −0.21 −2.11 0.037 aspects of inferential processing (Souchay et al., 2007). That is, ac- LH −0.38 0.19 −0.21 −2.03 0.045 cording to Koriat (1993), when a participant is unable to answer a Medial orbitofrontal BH −0.45 0.22 −0.21 −2.02 0.046 question, the FOK judgment will be based on partial information ac- RH −0.49 0.20 −0.24 −2.43 0.017 cessed during the search for the right answer (Koriat, 1993). As a −0.29 0.21 −0.14 −1.37 0.17 consequence, the accuracy of the FOK depends on the overall accuracy Linear regression models are adjusted for age, education, and memory perfor- of the partial information (Koriat, 1993), providing a feeling of re- mance. Multiple comparisons were accounted for using Bonferroni corrected a collection. Previous studies investigating episodic memory have found of 0.01 (for episodic FOK) respectively 0.0083 (for semantic FOK). BH that older adults fail to retrieve contextual information (Craik and =Bilateral hemispheres, RH = Right hemisphere, LH = Left hemisphere. Salthouse, 1992), and in an elegant study by Souchay and colleagues (2007), these findings were extended to include episodic FOK 6 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 (Souchay et al., 2007). Specifically, using a task design where they correlates. In addition, the current study only investigated cognitively compared FOK in tasks which did (episodic FOK) and did not (semantic unimpaired older adults. Future studies should investigate whether FOK) rely on proficiency of recollection, they only found an age effect these results are similar or change in individuals with mild cognitively on FOK accuracy when predictions were made on an episodic memory impairment or AD, as meta-cognitive deficits may increase in magni- task but not on a semantic task (Souchay et al., 2007). In a second tude. experiment, they examined the relationship between FOK accuracy and In conclusion, we found that older adults had reduced accuracy of the availability of contextual cues and partial information available FOK when predictions were made on episodic information as compared during examination as indexed by ‘remembering’(Souchay et al., 2007). to when FOK predictions were made on semantic information. Our They found that participants who made more accurate FOK predictions findings may help define and distinguish pathological substrates un- prior to the test phase made more recognition judgments on the basis of derlying metacognitive monitoring of episodic and semantic memory. remembering. The authors concluded that this demonstrates ‘a failure Overall, these findings may help to better understand the mechanism of metacognitive processes at retrieval, not encoding, that characterizes underlying meta-memory deficits in aging and ultimately unawareness the aging deficits in meta-cognition.’ (anosognosia) in Alzheimer's disease. Our findings demonstrate that meta-memory for episodic respec- tively semantic information were related to tauopathy, but not Aβ-pa- Financial disclosures thology, in different brain regions. Specifically, we found that deficits in metacognitive episodic knowledge were related to increased tau burden Dr. Rentz has served as a consultant for Eli Lilly, Biogen, and in episodic memory-related brain regions, brain regions vulnerable to Janssen Pharmaceutical, and sits on the Scientific Advisory Board pathological changes in early AD. Results remained whilst controlling for Neurotrack. for age, memory performance (overall accuracy on the task & SRT dr), Dr Johnson has served as paid consultant for Bayer, GE Healthcare, and even semantic FOK. In contrast, we found that decreased semantic Janssen Alzheimer's Immunotherapy, Siemens Medical Solutions, meta-memory was related to increased tau burden in the temporal pole Genzyme, Novartis, Biogen, Roche, ISIS Pharma, AZTherapies, GEHC, (similar as the episodic FOK) but also tau in regions associated with the Lundberg, and AbbVie. He is a site co-investigator for Lilly/Avid, Pfizer, semantic knowledge network. This later finding, in contrast to episodic Janssen, and Navidea. meta-memory, suggests that the association between tau and meta- Dr. Sperling has served as a paid consultant for AbbVie, Biogen, memory for semantic information is expanding beyond Braak stages I- Bracket, Genentech, Lundbeck, Roche, and Sanofi. She has served as a II. However, the association between semantic FOK and tau burden in co-investigator for Avid, Eli Lilly, and Janssen Alzheimer these regions did not survive in the ROI analyses after correction for Immunotherapy clinical trials. She has spoken at symposia sponsored multiple comparisons. These findings can be interpreted in regard to by Eli Lilly, Biogen, and Janssen. She receives research support from Koriat's (1993) accessibility model, in which FOK judgments are based Janssen and Eli Lilly. on the amount of information produced during retrieval. In this view, awareness of memory requires the retrieval of memory either fully or CRediT authorship contribution statement partially, which gives rise to a subjective state of consciousness such as recalling. Moreover, retrieval of episodic information has been asso- Patrizia Vannini: Conceptualization, Formal analysis, Writing - ciated with increased activation in the MTL regions, including the original draft. Federico d'Oleire Uquillas: Conceptualization, Formal hippocampus e.g., (Yonelinas et al., 2005). Our findings may indicate analysis, Writing - original draft. Heidi I.L. Jacobs: Conceptualization, that tau pathology in the MTL regions, especially in the entorhinal Formal analysis, Writing - original draft. Jorge Sepulcre: Writing - cortex, might influence an individual's judgment and reduce their ac- review & editing. Jennifer Gatchel: Writing - review & editing. curacy for episodic FOK. That is, we speculate that older adults har- Rebecca E. Amariglio: Writing - review & editing. Bernard boring MTL tau pathology may have particular difficulties in accessing Hanseeuw: Writing - review & editing. Kathryn V. Papp: Writing - or utilizing cues from retrieval attempts to accurately predict future review & editing. Trey Hedden: Writing - review & editing. Dorene M. recognition. Furthermore, the magnitude of tau-pathology maybe also Rentz: Writing - review & editing. Alvaro Pascual-Leone: Writing - be an important contributor to the observed behavioral deficit, as in- review & editing. Keith A. Johnson: Writing - review & editing. creased tau pathology was observed to be related to decreased meta- Reisa. A. Sperling: Writing - review & editing. memory for episodic information (see Fig. 2B). In addition, the differ- ential age effects between episodic and semantic FOK reported pre- viously and also observed in the current study, might actually reflect Declaration of competing interest differential amounts of underlying tau pathology. The fact that this effect remained significant even after adjusting for episodic memory None. performance (overall accuracy on the task & SRT dr) suggests that this ‘state of consciousness’ is separate from the process of recalling. The Acknowledgements fact that we did not see the same pattern for semantic FOK supports the idea that FOK for semantic and episodic memory relies on different This research was carried out in whole or in part at the Athinoula A. partial information (Yonelinas and Jacoby, 1995; Souchay et al., Martinos Center for Biomedical Imaging at the Massachusetts General 2002b). An alternative interpretation of these findings is that it is Hospital, using resources provided by the Center for Functional conceivable that the functionality of the memory systems is differen- Neuroimaging Technologies, P41EB015896, a P41 Regional Resource tially impaired, such that neurodegenerative process have already af- supported by the National Institute of Biomedical Imaging and fected the different memory system. Bioengineering (NIBIB), National Institutes of Health. Dr. P. Vannini received funding from NIH-NIA grant K01AG048287. Additional 5. Limitations funding came from NIA P01 AG036694 to R.A.S. and K.A.J., K24 AG035007 to R.A.S as well as the Alzheimer's Association (IIRG-06- We acknowledge that there are several limitations to this study. The 27374 to R.A.S.). H.I.L.J. received funding from the European Union's focus of this study was to investigate the pathological correlates of FOK. Horizon 2020 Research and Innovation Programme under the Marie However, future research is needed to investigate whether different Sklodowska-Curie Grant agreement (IF-2015-GF, 706714). The content measurements of meta-cognition e.g. judgment-of-learning, show dif- is solely the responsibility of the authors and does not necessarily re- ferent underlying cognitive processes and consequently different brain present the official views of the National Institutes of Health. 7 P. Vannini, et al. NeuroImage: Clinical 24 (2019) 102097 Supplementary Materials Kelemen, W.L., Frost, P.J., CAr, W., 2000. Individual differences in metacognition: evi- dence against a general metacognitive ability. Mem. Cognit. 28, 92–107. Kemppainen, N.M.M.D.P., Aalto, S.M., Wilson, I.A.P., Nagren, K.P., Helin, S.M., Bruck, Supplementary material associated with this article can be found, in A.M.D.P., Oikonen, V.M., Kailajarvi, M.M.D.P., Scheinin, M.M.D.P., Viitanen, the online version, at 10.1016/j.nicl.2019.102097. M.M.D.P., Parkkola, R.M.D.P., Rinne, J.O.M.D.P., 2006. Voxel-based analysis of PET amyloid ligand [11C]PIB uptake in Alzheimer disease. Neurology 67, 1575–1580. Klunk, W.E., Engler, H., Nordberg, A., Wang, Y., Blomqvist, G., Holt, D.P., Bergström, M., References Savitcheva, I., Huang, G.-F., Estrada, S., Ausén, B., Debnath, M.L., Barletta, J., Price, J.C., Sandell, J., Lopresti, B.J., Wall, A., Koivisto, P., Antoni, G., Mathis, C.A., Långström, B., 2004. Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Allen-Burge, R., Storandt, M., 2000. Age equivalence in feeling-of-knowing experiences. Compound-B. Ann. Neurol. 55, 306–319. J. Gerontol. B Psychol. Sci. Soc. Sci. 55, 214–223. Koriat, A., 1993. How do we know that we know? The accessibility model of the feeling of Anderson, J.W., Edgecomb, S., 2010. Mild cognitive impairment and feeling-of-knowing knowing. Psychol. Rev. 100, 609–639. in episodic memory. J. Clin. Exp. Neuropsychol. 32, 505–514. Leonesio, R.J., Nelson, T.O., 1990. Do different metamemory judgments tap the same Babinski, M.J., 1914. Contibution a l’etudedes troubles mantaux dans l’hemiplegie or- underlying aspects of memory? J. Exp. Psychol. Learn. Mem. Cognit. 16, 464–467. ganique cerebrale (Anasognosie). Rev. Neurol. 12, 845–848. Lipinska, B., Bäckman, L., 1996. Feeling-of-knowing in fact retrieval: further evidence for Bäckman, L., Karlsson, T., 1985. The relation between level of general knowledge and preservation in early Alzheimer’s disease. J. Int. Neuropsychol. Soc. 2, 350–358. feeling-of-knowing: an adult age study. Scand. J. Psychol. 26, 249–258. Marquie, J.C., Huet, N., 2000. Age differences in feeling-of-knowing and confidence Backman, L., Lipinska, L., 1993. Monitoring of general knowledge: evidence for pre- judgements as a function of knowledge domain. Psychol. Aging 15, 451–460. servation in early Alzheimer's disease. Neuropsychologia 31, 335–345. Mintun, M.A., LaRossa, G.N., Sheline, Y.I., Dence, C.S., Lee, S.Y., Mach, R.H., Klunk, W.E., Becker, J.A., Hedden, T., Carmasin, J., Maye, J., Rentz, D.M., Putcha, D., Fischl, B., Greve, Mathis, C.A., DeKosky, S.T., Morris, J.C., 2006. [11C]PIB in a nondemented popu- D.N., Marshall, G.A., Salloway, S., Marks, D., Buckner, R.L., Sperling, R.A., Johnson, lation: potential antecedent marker of Alzheimer disease. Neurology 67, 446–452. K.A., 2011. Amyloid-β associated cortical thinning in clinically normal elderly. Ann. Morris, J.C., 1993. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurol. 69, 1032–1042. Neurology 43, 2412–2414. Benton, A.L., 1968. Differential behavioural effects in frontal lobe disease. Morson, S.M., Moulin, C.J.A., Souchay, C., 2015. Selective deficits in episodic feeling of Neuropsychologia 6, 53–60. knowing. In: Ageing: A Novel Use of the General Knowledge Task. 15. pp. 85–92. Bertrand, E., Azar, M., Rizvi, B., Brickman, A.M., Huey, E.D., Habeck, C., Landeira- Nelson, T.O., 1984. A comparison of current measures of the accuracy of feeling-of- Fernandez, J., Mograbi, D.C., Cosentino, S., 2018. Cortical thickness and metacog- knowing predictions. Psychol. Bull. 95, 109–113. nition in cognitively diverse older adults. Neuropsychology 32, 700–710. Nelson, T.O., Narens, L., 1990. Metamemory: A theoretical framework and new findings. Braak, H., Braak, E., 1991. Neuropathological stageing of Alzheimer-related changes. Psychol. Learn. Motiv. 26, 125–173. Acta Neuropathol. 82, 239–259. Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., Panksepp, J., 2006. Braak, H., Braak, E., 1995. Staging of Alzheimer´s disease-related neurofibrillary changes. Self-referential processing in our brain – a meta-analysis of imaging studies on the Neurobiol. Aging 16, 271–284. self. Neuroimage 15, 440–457. Chua, E.F., Schacter, D.L., Sperling, R.A., 2009. Neural correlates of metamemory: a Pappas, B.A., Sunderland, T., Weingartner, H.M., Vitiello, B., Martinson, H., Putnam, K., comparison of feeling-of-knowing and retrospective confidence judgments. J. Cognit. 1992. Alzheimer’s disease and feeling-of-knowing for knowledge and episodic Neurosci. 21, 1751–1765. memory. J. Gerontol. 47, 159–164. Cosentino, S., Brickman, A.M., Griffith, E., Habeck, C., Cines, S., Farrell, M., Shaked, D., Perrotin, A., Belleville, S., Isingrini, M., 2007. Metamemory monitoring in mild cognitive Huey, E.D., Briner, T., Stern, Y., 2015. The right insula contributes to memory impairment: evidence of a less accurate episodic feeling-of-knowing. awareness in cognitively diverse older adults. Neuropsychologia 75, 163–169. Neuropsychologia 45, 2811–2826. Craik, F.I.M., Salthouse, T.A., 1992. Handbook of Aging and cognition. Hillsdale, Rentz, D.M., Locascio, J.J., Becker, J.A., Moran, E.K., Eng, E., Buckner, R.L., Sperling, Erlbaum. R.A., Johnson, K.A., 2010. Cognition, reserve, and amyloid deposition in normal Dagley, A., LaPoint, M., Huijbers, W., Hedden, T., McLaren, D.G., Chatwal, J.P., Papp, aging. Ann. Neurol. 67, 353–364. K.V., Amariglio, R.E., Blacker, D., Rentz, D.M., Johnson, K.A., Sperling, R.A., Schultz, Rosen, H.J., Alcantar, O., Zakrzewski, J., Shimamura, A.P., Neuhaus, J., Miller, B.L., A.P., 2015. Harvard aging brain study: dataset and accessibility. NeuroImage 144, 2014. Metacognition in the behavioral variant of frontotemporal dementia and 255–258. Alzheimer's disease. Neuropsychology 28, 436–447. Edison, P.M., Archer, H.A.M., Hinz, R.P., Hammers, A.P., Pavese, N.M., Tai, Y.F.M., Rosen, W.G., Terry, R.D., Fuld, P.A., Katzman, R., Peck, A., 1980. Pathological verifica- Hotton, G.M., Cutler, D.B., Fox, N.P., Kennedy, A.M., Rossor, M.M.D.D., Brooks, tion of ischemic score in differentiation of dementias. Ann. Neurol. 7, 486–488. D.J.M.D.D., 2007. Amyloid, hypometabolism, and cognition in Alzheimer disease: an Rousset, O.G., Ma, Y., Evans, A.C., 1998. Correction for partial volume effects in PET: [11C]PIB and [18F]FDG PET study. Neurology 68, 501–508. principle and validation. J. Nucl. Med. 39, 904–911. Engler, H., Forsberg, A., Almkvist, O., Blomquist, G., Larsson, E., Savitcheva, I., Wall, A., Schacter, D.L., 1983. Feeling-of-Knowing in episodic memory. J. Exp. Psychol. 9, 39–54. Ringheim, A., Langstrom, B., Nordberg, A., 2006. Two-year follow-up of amyloid Shoup, T.M., Yokell, D.L., Rice, P.A., Jackson, R.N., Livni, E., Johnson, K.A., Brady, T.J., deposition in patients with Alzheimer's disease. Brain 129, 2856–2866. Vasdev, N., 2013. A concise radiosynthesis of the tau radiopharmaceutical, [(18) Folstein, M.F., Folstein, S.E., McHough, P.R., 1975. “Mini-Mental State”: a practical F]T807. J. Label. Comp. Radiopharm. 56, 736–740. method for grading cognit9ive state of patients for the clinician. J. Psychiatr. Res. 12, Souchay, C., Insigrini, M., Gil, R., 2002Aa. Alzheimer's disease and feeling-of-knowing in 189–198. episodic memory. Neuropsychologia 40, 2386–2396. Forsberg, A., Engler, H., Almkvist, O., Blomquist, G., Hagman, G., Wall, A., Ringheim, A., Souchay, C., Insigrini, M., Roger, G., 2002Ab. Alzheimer’s disease and feeling-of- Långström, B., Nordberg, A., 2008. PET imaging of amyloid deposition in patients knowing. Episodic Mem. Neuropsychol. 40, 2386–2396. with mild cognitive impairment. Neurobiol. Aging 29, 1456–1465. Souchay, C., Isingrini, M., Espagnet, L., 2000. Aging, episodic memory feeling-of- Freedman, J.L., Landauer, T.K., 1966. Retrieval of long-term memory: “tip-of-the-tongue” knowing, and frontal functionning. Neuropsychology 2, 299–309. phenomenon. Psychon. Sci. 4, 309–310. Souchay, C., Isingrini, M., Pillon, B., Gil, R., 2003. Metamemory accuracy in Alzheimer's Gardiner, F.M., Craik, F.I., Bleasdale, F.A., 1973. Retrieval difficulty and subsequent re- disease and frontotemporal lobe dementia. Neurocase 9, 482–492. call. Mem. Cognit. 1, 213–216. Souchay, C., Moulin, C.J.A., 2013. Eliciting the implicit: metacognition in Alzheimer’s Gomperts, S.N.M.D.P., Rentz, D.M.P., Moran, E.B., Becker, J.A.P., Locascio, J.J.P., Klunk, disease. Cognit. Neurosci. W.E.M.D.P., Mathis, C.A.P., Elmaleh, D.R.P., Shoup, T.P., Fischman, A.J.M., Hyman, Souchay, C., Moulin, C.J.A., Clarys, D., Taconnat, L., Isingrini, M., 2007. Diminished B.T.M.D.P., Growdon, J.H.M., Johnson, K.A.M., 2008. Imaging amyloid deposition in episodic memory awareness in older adults: evidence from feeling-of-knowing and Lewy body diseases SYMBOL. Neurology 71, 903–910. recollection. Conscious. Cognit. 16, 769–784. Greve, D.N., Salat, D.H., Bowen, S.L., Izquierdo-Garcia, D., Schultz, A.P., Catana, C., Sperling, R.A., Bates, J.F., Chua, E.F., Cocchiarella, A.J., Rentz, D.M., Rosen, B.R., Becker, J.A., Svarer, C., Knudsen, G.M., Sperling, R.A., Johnson, K.A., 2016. Different Schacter, D.L., Albert, M.S., 2003. fMRI studies of associative encoding in young and partial volume correction methods lead to different conclusions: an (18)F-FDG-PET elderly controls and mild Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry 74, study of aging. Neuroimage 132, 334–343. 44–50. Grober, E., Lipton, R.B., Hall, C., Crystal, H., 2000. Memory impairment on free and cued Turro-Garriga, O., Garre-Olmo, J., Vilalta-Franch, J., Conde-Sala, J.L., de Gracia Blanco, selective reminding predicts dementia. Neurology 54, 827–832. M., Lopez, P., 2013. Burden associated with the presence of anosognosia in Gruneberg, M.M., Monks, J., 1974. “Feeling of knowing” and cued recall. Acta Psychol. Alzheimer’s disease. Int. J. Geriatr. Psychatr. 28 (S I), 291–297. 38, 257–265. Turro-Garriga, O., Garre-Olmoa, J., Rene-Ramırez, R., Laia Calvo-Perxasa, L., Gascon- Hart, J., 1965. Memory and the feeling-of-knowing experience. J. Educ. Psychol. 56, Bayarri, J., Conde-Sala, J.-L., 2016. Consequences of anosognosia on the cost of 208–216. caregivers’ care in Alzheimer’s disease. J. Alzheimer’s Dis. 54, 1551–1560. Hart, J., 1967. Second-try recall, recognition, and the memory-monitoring process. J. van der Meer, L., Costafreda, S., Aleman, A., David, A.S., 2010. Self-reflection and the Educ. Psychol. 58, 193–197. brain: a theoretical review and meta-analysis of neuroimaging studies with im- Hosey, L.A., Peynircioğlu, Z.F., Rabinovitz, B.E., 2009. Feeling of knowing for names in plications for schizophrenia. Neurosci. Biobehav. Rev. 34, 935–946. response to faces. Acta Psychol. 130, 214–224. Yonelinas, A.P., Jacoby, L.L., 1995. Dissociating automatic and controlled processes in a Johnson, K.A., Gregas, M., Becker, J.A., Kinnecom, C., Salat, D.H., Moran, E.K., Smith, memory-search task: beyond implicit memory. Psychol. Res. 57, 156–165. E.E., Rosand, J., Rentz, D.M., Klunk, W.E., Mathis, C.A., Price, J.C., DeKosky, S.T., Yonelinas, A.P., Otten, L.J., Shaw, K.N., Rugg, M.D., 2005. Separating the brain regions Fischman, A.J., Greenberg, S.M., 2007. Imaging of amyloid burden and distribution involved in recollection and familiarity in recognition memory. J. Neurosci. 16, in cerebral amyloid angiopathy. Ann. Neurol. 62, 229–234. 3002–3008. Johnson, K.A., Schultz, A., Betensky, R.A., Becker, J.A., Sepulcre, J., Rentz, D., Mormino, Zamboni, G., Wilcock, G., 2011. Lack of awareness of symptoms in people with dementia: E., Chhatwal, J., Amariglio, R., Papp, K., Marshall, G., Albers, M., Mauro, S., Pepin, the structural and functional basis. Int. J. Geriatr. Psychiatry 26, 783–792. L., Alverio, J., Judge, K., Philiossaint, M., Shoup, T., Yokell, D., Dickerson, B., Gomez- Isla, T., Hyman, B., Vasdev, N., Sperling, R., 2016. Tau positron emission tomo- graphic imaging in aging and early Alzheimer disease. Ann. Neurol. 79, 110–119. 8