INTRODUCTION:The locus coeruleus is among the earliest brain regions affected by tau pathology in Alzheimer's disease (AD), but the functional impact is difficult to measure in vivo. Task-evoked pupil dilation provides an index of locus coeruleus-norepinephrine function that might be altered in cognitively normal older adults with underlying AD pathology. METHODS:Cognitively normal older adults identified as AD biomarker positive (N = 25) or negative (N = 36) based on plasma phosphorylated tau (p-tau) levels completed a conjunctive visual search task that manipulated attentional load by varying set size. Pupil dilation responses during the task were analyzed using mixed-effects models and time-resolved regression. RESULTS:Despite comparable accuracy and reaction times, biomarker-positive adults showed reduced load-dependent modulation of pupil dilation during target-present trials. Weaker modulation was associated with higher p-tau levels and poorer executive and memory function. DISCUSSION:Attenuated task-evoked modulation of pupil dilation during visual search reveals locus coeruleus-norepinephrine dysfunction in preclinical AD.

INTRODUCTION:Relatively low rates of Latino participation in Alzheimer's disease and related dementias (ADRD) research makes it difficult to determine whether identified disease mechanisms, risk factors, or novel treatments generalize to this population. METHODS:We introduced ADRD research opportunities through a model cognitive screening program in primary care and neurology specialty care clinics in areas with high proportions of Latino residents. RESULTS:Out of 523 Latino adults (mean age = 72.1 years, mean education = 6.8 years, 62.1% female, 88.1% tested in Spanish), 520 allowed the use of screening data for research, and high percentages agreed to a research registry with contact about ADRD research opportunities (primary care: 91.9%, neurology: 86.6%). Registrants supported 368 referrals to 45 studies (21% successfully recruited). Thirty-one individuals participated in ≥1 study, producing 79 enrollments across 15 ADRD studies. DISCUSSION:Results demonstrate that combining a needed clinical service with recruitment efforts can enhance participation of older Latino adults in ADRD research.

BACKGROUND:Cognitive-functional composite measures are increasingly used as primary efficacy endpoints in early Alzheimer's disease (AD) trials, where greater sensitivity to decline can improve trial efficiency and reduce sample size requirements. OBJECTIVES:To compare sensitivity to decline of the Cognitive Functional Component 2 (CFC2), a novel cognitive-functional composite measure described by Raghavan et al. (2013), against the Clinical Dementia Rating - Sum of Boxes (CDR-SB) and other standard cognitive and functional outcomes including MMSE, FAQ, ADAS Cog 13 and ADCOMS using prospective randomized clinical trial data. DESIGN:The VIVA-MIND trial was a phase 2A/2B randomized controlled trial investigating the safety and efficacy of varoglutamstat in patients with mild cognitive impairment and mild dementia due to AD. SETTING:The VIVA-MIND trial was conducted between 2021-2024. It was prematurely terminated in mid-2024 by the study sponsor. PARTICIPANTS:This secondary analysis uses data from 98 participants in the modified intention-to-treat population from the VIVA-MIND trial with complete neuropsychological test data. MEASUREMENTS:Standard power calculations informed by parameters estimated from linear mixed-effects models were used to determine the relative efficiency of outcome measures. RESULTS:The CFC2 was more sensitive to decline than the CDR-SB in this population. Use of the CFC2 would yield a 15% reduction in required sample size relative to the CDR-SB. Application of an optimal weighting scheme further improved the sensitivity of the CFC2. CONCLUSIONS:Practically significant differences in the efficiency of clinical trials in early AD may be realized by the choice of clinical outcome measure and weighting scheme. Although further verification is needed, we replicate a previous finding that the CFC2 may outperform the CDR-SB in the early AD population.

Abstract INTRODUCTION Hispanic/Latino(a/x) (H/L) representation in Alzheimer's disease and related dementias (ADRD) research remains low, limiting generalizability of findings and interventions to this growing population. H/L community health workers (“Promotores”) can enhance awareness of brain health and help guide their communities toward ADRD research opportunities; however, the effectiveness of recruitment through Promotores compared to traditional clinic‐based approaches remains unknown. This study evaluates our Promotores training and compares the two approaches on recruitment success and characteristics of those recruited into a longitudinal study of ADRD. METHODS Nine Promotores completed Building Research Integrity and Capacity (BRIC) training on research concepts and ethics, and video training modules on ADRD‐related research procedures. Training effectiveness was evaluated via pre‐/post‐tests to evaluate content knowledge and program satisfaction was evaluated by a self‐report survey. Recruitment success was compared between Promotores‐based and clinic‐based approaches, assessing numbers referred and enrolled in ADRD research, and participants’ demographic and clinical characteristics and agreement to specific research procedures (i.e., lumbar puncture, brain donation). RESULTS Promotores showed significant gains in knowledge of research concepts (Wilcoxon test, p = 0.005; median increase = 52%, 95% confidence interval [CI] = 30%–70%) and ADRD content (p = 0.042; median increase = 10%, 95% CI = 5%–40%). Promotores strongly agreed (on a 5‐point Likert scale) that their training improved understanding of researcher responsibilities (x̄ = 4.88 ± 0.35), participant rights (x̄ = 4.25 ± 1.49), and ADRD research procedures (x̄ = 3.88 ± 1.36). Satisfaction with training materials (x̄ = 4.63 ± 0.74) and perceived community benefit (x̄ = 4.75 ± 0.46) were high. Promotores and clinic pathways had similar enrollment proportions (12.6% vs. 16.5%; Χ2 = 0.92, p = 0.34) and participants were comparable across demographic, clinical, social determinants of health, and willingness to undergo research procedures. DISCUSSION The Promotores training program was well received, improved research literacy, and was as successful as clinic‐based recruitment in enrolling older H/L adults into ADRD research. Results validate our Promotores training strategy and support a multipronged community‐oriented approach to H/L participation in ADRD research.

Cognitive assessments sensitive to the integrity of the medial temporal lobe, an area vulnerable to early tau deposition, may serve as low-cost adjunctive markers of underlying tau pathology in older adults. The Mnemonic Similarity Task (MST) is a fine memory discrimination task designed to assess hippocampal integrity. The current cross-sectional study utilized baseline data from two AD prevention trials (the Anti-Amyloid Treatment in Asymptomatic Alzheimer's (A4) study and the Exercise in Adults with Mild Memory Problems (EXERT) trial) to examine relationships between MST performance, amyloid-beta, tau, and hippocampal volume. We additionally explored relationships between performance on a traditional memory test, Logical Memory, and AD-related brain measures. Poorer fine memory discrimination was associated with higher tau as assessed by PET in A4 (N = 407, 59% female, mean age = 71.66, age range = 65-85) and CSF (p-tau181, total tau) in EXERT (N = 41, 61% female, mean age = 74.10, age range = 65-89). Poorer fine memory discrimination was also associated with higher amyloid PET in A4 and smaller hippocampal volume in EXERT. Poorer delayed recall on Logical Memory was associated with higher tau and amyloid burden in A4 and with lower hippocampal volume in EXERT. Poorer retention on Logical Memory was associated with higher tau in Braak I and amyloid in A4 and with CSF tau and lower hippocampal volume in EXERT. These results support the potential of fine memory discrimination as measured by the MST as an adjunctive, accessible screening measure associated with higher tau in cognitively normal, amyloid positive older adults and older adults with amnestic MCI.

Dementia risk factors often coexist and interact in ways that create distinct profiles, which may influence the efficacy of lifestyle-based interventions aimed at reducing dementia risk. This study aimed to: 1) characterize dementia risk profiles in participants of the CAN-THUMBS-UP Brain Health Support Program (BHSP), and 2) evaluate whether these profiles predict responses to a 45-week educational intervention. We analyzed data from 296 BHSP participants with complete baseline assessments. Principal component analysis identified risk profiles based on seven modifiable risk factors: physical activity, cognitive engagement, diet, sleep, social and psychological health, vascular health, and vision and hearing. All participants accessed Brain Health PRO, a web-based educational intervention targeting these modifiable risk factors. Risk factor changes were measured through lifestyle questionnaires administered every three months, and cognitive changes were measured via a neuropsychological test battery at baseline and 12 months. Linear mixed models with repeated measures, adjusted for age, sex, education, and cognitive status evaluated associations between risk profiles and changes in cognition and risk factors. Three distinct profiles were identified: 1) sleep and social/psychological health, 2) cognitive engagement, and 3) diet, physical activity and vascular heath. Across all profiles, lower baseline scores indicated a higher risk in the factors that characterized each profile. Participants with lower baseline scores in the first two profiles showed greater improvements in associated risk factor domains compared to those with better baseline scores. In the third profile, participants with lower baseline scores showed greater improvements in diet and physical activity, but no change was observed in vascular health. No association was observed between any of the risk profiles and cognitive change. A longer follow-up or a more intensive intervention beyond online education may be needed to impact vascular health and cognition. Understanding dementia risk factor profiles and their influence on intervention effects can help inform more personalized prevention strategies. Further research is needed to validate and refine these profiles.

Normative references for neuropsychological tests commonly adjust for years of education. However, quality of formal education varies widely, and additional education may be obtained outside of the classroom setting. Therefore, years of education may inadequately reflect full educational achievement. Grade-equivalency scores from tests of single-word reading provide additional information about learning experience and may alter interpretation of test results (Manly et al., 2002). Therefore, we explored the impact of adjusting cognitive test results for years of education vs. estimated grade equivalency in Spanish-speaking U.S. Latinos. 140 Spanish-speaking Latinos with subjective cognitive decline were referred by a community-based neurologist for objective cognitive testing. Participants were aged 47-88 (M=70.6, SD=8.0) with an average of 9.1 self-reported years of education (SD=4.5, range=0-20). Participants completed a cognitive screening battery using tests from the National Alzheimer's Coordinating Center Uniform Data Set Neuropsychological Battery (UDS3-NB). The Woodcock-Muñoz Letter-Word Identification Test was used to estimate grade-equivalency. Standardized scores (Z-scores) that adjusted for age, sex, language of testing, and either self-reported years of education or Woodcock-Muñoz grade-equivalency were obtained for each cognitive test using the UDS3-NB Latino Norms Calculator (Marquine et al.,2023). The two sets of Z-scores were compared using paired samples t-tests. Grade equivalency estimated by the Woodcock-Muñoz (M=13.5, SD=4.6; range=1.7-18) was significantly higher than self-reported years of education (M=9.1, SD=4.5; t [139]=10.9, p < .001). Figure 1 shows the distribution of the discrepancy between grade equivalency and self-reported educational attainment. For all neuropsychological tests, adjusted Z-scores derived using grade equivalency were significantly lower than those derived using years of education (see Table 1). The percentage of participants who fell below z-score cutoffs of -1, -1.5, or -2 was greater when adjusting for grade-equivalency than when adjusting for years of education (Table 2). Education-adjusted norms appeared to underestimate level of impairment in Spanish-speaking U.S. Latinos who presented to a neurologist with cognitive complaints. Self-reported years of education may not adequately reflect the educational experiences of this diverse group. Although additional validation is needed, these results indicate caution is warranted when interpreting education-adjusted test scores for Spanish-speaking U.S. Latinos with lower levels of education.

BACKGROUND:Dementia risk factors coexist and interact in meaningful ways. Identifying risk profiles can inform prevention strategies. METHODS:This study examined profiles using seven modifiable risk factors and their associations with cognition and sex among older adults at risk of dementia. Risk profiles were identified using principal component analysis, and linear regressions assessed their associations with cognition. RESULTS:Three profiles emerged: (1) better sleep and social/psychological health; (2) better diet, physical activity, and vascular health; and (3) better cognitive engagement. Similar profiles were observed in women, except that better cognitive engagement grouped with poorer vascular health. In men, better cognitive engagement grouped with better vascular health and vision/hearing, while lower cognitive engagement was grouped with better sleep and social/psychological health. The lower cognitive engagement profile in men was associated with poorer cognition. DISCUSSION:Prevention strategies should target meaningful combinations of risk factors to improve dementia prevention. Further research is needed to validate and refine these profiles. Highlights:This study evaluated risk profiles using a broad range of modifiable risk factors.This is the first study to analyze such profiles in a Canadian population, stratified by sex.Risk profiles differed by sex, especially for the cognitive engagement risk factor.In men, lower cognitive engagement profile was associated with poorer cognition.

Emerging evidence supports the diagnostic and prognostic utility of plasma biomarkers in Alzheimer's disease (AD), particularly in early disease stages. We sought to extend these findings by evaluating the prognostic value of plasma biomarkers in a clinical trial of mild-to-moderate AD. Post-hoc analyses investigated whether baseline concentrations of plasma biomarkers (Aβ42/Aβ40, T-tau, P-tau181, NfL, and GFAP) predicted change in ADAS-Cog11, CDR-SB, and volumetric MRI among participants in T2 Protect AD, a negative 48-week, phase-2, placebo-controlled trial of troriluzole in mild-to-moderate AD. All trial participants met diagnostic criteria for probable AD. Baseline concentrations of, and 48-week changes in, plasma biomarkers were assessed for association with 48-week change in outcomes using linear regression. Combinations of baseline biomarkers that best predicted change on the ADAS-Cog11 and CDR-SB were identified using least absolute shrinkage and selection operator (LASSO) regression. Biomarker-informed sample size calculations were modeled. Of 350 trial participants, 319 had all requisite biomarker and clinical outcome data for inclusion in these analyses (mean age 71.5, SD = 8.03; 58.6% female). Higher plasma NfL at baseline predicted worsening scores on the ADAS-Cog11 (effect size (ES) = 1.42, 95%CI = [0.43, 2.41], p = 0.026) and CDR-SB (ES = 0.42, 95%CI = [0.10, 0.73], p = 0.048). LASSO regression revealed that worsening on the ADAS-Cog11 was best predicted by the combination of baseline plasma NfL, T-tau, and Aβ42/40 ratio, whereas baseline NfL alone best predicted worsening on CDR-SB. Higher baseline NfL predicted increasing ventricular volume (ES = 1.30cm3, 95%CI = [0.43, 2.17], p = 0.018) and decreasing mid-temporal cortical volume (ES = -0.47, 95%CI = [-0.74, -0.20], p = 0.003). Increasing NfL over the 48-week trial was associated with worsening on CDR-SB but not ADAS-Cog11. Modeling of biomarker-informed power calculations revealed that including high NfL as a trial entry criterion could substantially reduce requisite trial sample size. Elevated baseline plasma NfL predicted more rapid clinical decline and MRI volume loss. Furthermore, increasing plasma NfL concentration over time was associated with worsening on the CDR-SB. Plasma NfL is an easily accessible biomarker that may enhance the design of clinical trials in mild-to-moderate AD. The T2 Protect AD trial was registered as NCT03605667 on clinicaltrials.gov on 2018-07-27.

Alzheimer’s disease (AD) pathology (e.g., beta-amyloid plaques, tau neurofibrillary tangles) accumulates before the emergence of cognitive deficits that lead to a diagnosis of mild cognitive impairment (MCI) or dementia. Early vulnerability of medial temporal lobe structures to AD suggests that subtle episodic memory decline should be among the first cognitive markers of the disease. Therefore, we examined the relationship between 2-year change in episodic memory and baseline cerebrospinal fluid (CSF) biomarkers of AD in cognitively normal (CN) older adults. Episodic memory was assessed with the California Verbal Learning Test, Second Edition (CVLT-II), a sensitive 16-item list learning test that measures multiple aspects of memory proficiency. We administered the CVLT-II at baseline, 1-year, and 2-year evaluations to older adults who were CN and CSF AD biomarker-negative (CN–; n = 94) or positive (CN+ or preclinical AD; n = 38; CSF tau/Aβ42 ratio>.609) at baseline. We compared groups on change over time on total recall on List A learning trials 1-5, list B free recall, short- and long-delay free and cued recall, intrusion errors, and recognition discriminability. We also examined target recognition (d’) in relation to distractor types that differed in degree of similarity to targets. Group X time repeated measures ANCOVAs controlling for age and sex showed significantly greater 2-year decline in CN+ than CN- on CVLT-II measures of long-delay free recall (interaction: F(2,256) = 4.39;p<.05) and recognition discriminability overall (interaction: F(2,256) = 3.08;p<.05) (Fig. 1) and against most distractor types (p’s<.05) (Fig. 2). Higher tau/Aβ42 ratio (i.e., more AD-like) in CN older adults was associated with greater decline on long-delay free recall ( r (130) = –.205; p <.02) and recognition discriminability ( r (130) = –.215; p <.02) (Fig. 3). However, magnitude of decline only marginally categorized individuals as CN- vs. CN+ in logistic regression analyses (95% specificity, but only 21% sensitivity). Subtle decline in episodic memory is evident during the preclinical phase of AD, affecting both delayed recall and recognition consistent with medial temporal lobe dysfunction. In addition, greater preclinical memory decline is associated with higher levels of an AD biomarker in CSF. Further research will determine if these memory changes can serve as a cognitive marker of preclinical AD.

Visual short-term memory binding (VSTMB) requires efficient functional connectivity between cortical regions and is a sensitive behavioral marker of Alzheimer's disease (AD). VSTMB impairments have been detected in individuals with subjective cognitive decline (SCD) who perform normally on standard neuropsychological tests. Research has linked SCD to reduced integrity of the rostral-middle locus coeruleus (LC), an area that accumulates tau in early preclinical AD. Since the LC plays a crucial role in maintaining cortical efficiency, VSTMB may be similarly associated with LC integrity, particularly among those with SCD. Data were from cognitively unimpaired men in the Vietnam Era Twin Study of Aging ( N  = 350; mean age=72.9, SD=2.4) who completed a test of VSTMB, an LC-sensitive MRI scan, and the 39-item Everyday Cognition (ECog) scale; 274 participants also had informant ECog ratings. The VSTMB task employs a change detection paradigm and compares performance in shape-color binding (SCB) versus shape-only (SO) conditions (see Figure 1). Rostral-middle and caudal LC integrity was calculated as a contrast-to-noise ratio (LC-CNR) using a pontine tegmentum reference region. Mixed models regressed VSTMB accuracy across condition (SO, SCB) and set size (2, 3) within person, while testing main effects and interactions with LC-CNR and ECog. Models adjusted for age 20 cognitive ability, current age, depressive symptoms, and state anxiety. Sensitivity analyses adjusted for global performance on standard neuropsychological measures. VSTMB accuracy was lower on the SCB than SO condition (b=-1.08, p  = .001), especially at higher set sizes (b=-1.90, p  < .001). Higher participant-rated, but not informant-rated, SCD was associated with decreasing accuracy on SCB relative to the SO condition (b=-.253, p  = .007). Lower rostral-middle, but not caudal, LC-CNR was associated with poorer SCB accuracy relative to SO condition accuracy (b=7.10, p  = .010). Results remained significant after adjusting for global neuropsychological performance. Subtle losses in cognitive efficiency detected in cognitively unimpaired older adults on a VSTMB task were associated with SCD and LC integrity, even after accounting for performance on traditional neuropsychological tests. Given its role in modulating cortical efficiency through cognitive effort, reduced LC integrity may be associated with SCD when capacity for increasing compensatory effort to perform tasks is exceeded.

Varoglutamstat (PQ912) is an oral small molecule inhibitor of glutaminyl cyclases which reduces the pyroglutamate formation of Aβ and CCL2. Preclinical and Ph1 trials support this Ph2 evaluation in early AD. Our objectives included selecting the highest safe and well-tolerated dose of varoglutamstat with futility analysis at 24 weeks (Ph2A) followed by seamless evaluation of longer-term safety and efficacy at 72 weeks at this dose (Ph2B). Participants with biomarker confirmed early AD were randomized to varoglutamstat or placebo. A sequential design tested 3 descending doses (600mg, 300mg, 150mg BID) using a continuous Pocock safety boundary. A Stage Gate futility analysis of the first 180 subjects at 24 weeks was designed to guide the decision from Ph2A-B (Figure 1). The Ph2B primary endpoint was CDR-sum-of-boxes (CDR-SB), with secondary endpoints CFC2, ABC score, spectral EEG, FAQ, ADAS-Cog-13, and NPI. Table 1 presents baseline characteristics of the 109 randomized and dosed participants. The first dose cohort of varoglutamstat 600mg or placebo did not cross the safety boundary in Ph2A, supporting this dose selection for all participants thereafter. An abbreviated futility analysis was conducted after the trial was terminated prematurely for administrative reasons, with the provisional result of ‘halt enrollment’(yellow) (Figure 1). There were 74% of participants who completed week 24 and 31% week 72 (Figure 2). There were no significant differences between varoglutamstat and placebo in LS mean squares from baseline to week 72 in CDR-SB (-0.05 [95% CI -1.03, 0.92]), CFC2 (0.97 [95% CI -3.34, 5.28]), ABC score (0.10 [95% CI -0.10, 0.30]), FAQ (-0.60 [95% CI -4.22, 3.01]), ADAS-Cog-13 (2.03 [95% CI -2.29, 6.36]); or NPI (-3.09 [95% CI -7.81, 1.62]). At least one TEAE during treatment was reported in 84.9% varoglutamstat and 76.8% placebo with treatment discontinuations due to AE 11.3% varoglutamstat and 3.4% placebo. Severe TEAEs occurred with similar frequencies across arms. Frequency of SAEs/AESIs was 18.9/1.9% varoglutamstat versus 8.9%/5.4% placebo. Varoglutamstat was safe and well-tolerated in this VIVA-MIND early AD RCT. Efficacy assessment did not demonstrate any significant treatment benefits however, the available sample was limited. Biomarker results and PK/PD are pending.

INTRODUCTION:EXERT was a multisite randomized controlled trial (RCT) examining the effects of moderate-high intensity aerobic training (AX) versus lower-intensity stretching/balance/range of motion (SBR) on cognitive trajectories in older adults with amnestic mild cognitive impairment (aMCI). METHODS:Preplanned post-hoc analyses were conducted to compare each arm to a propensity-matched usual care (no intervention) group from Alzheimer's Disease Neuroimaging Initiative 1 (ADNI-1) selected for similarity across key characteristics. Differences in 12-month trajectories in the primary endpoint (ADAS-Cog-Exec) and magnetic resonance imaging (MRI) volumes in prespecified brain regions were compared. RESULTS:AX and SBR showed significantly less 12-month cognitive decline than ADNI-1 (AX:n = 109, β = 0.169, 95% confidence interval [CI] 0.011-0.328; SBR:n = 105, β = 0.181, 95% CI 0.007-0.354). There were trends of less prefrontal cortex volume loss for both EXERT groups and less AD signature region volume loss for SBR relative to ADNI-1 over 12 months. DISCUSSION:Moderate-high intensity aerobic or low-intensity flexibility exercise for 12 months in participants with aMCI may provide protection against decline relative to usual care. CLINICAL TRIAL REGISTRATION:The EXERT clinical trial is registered at clinicaltrials.gov (NCT02814526). HIGHLIGHTS:EXERT was a randomized controlled trial in sedentary older adults with aMCI. EXERT arms were propensity-matched to a usual care (no intervention) group (Alzheimer's Disease Neuroimaging Initiative 1 [ADNI-1]). High and low-intensity exercise arms had less 12-mo cognitive decline than ADNI-1. There were trends of less prefrontal cortex volume loss for each arm versus ADNI-1.

Benfotiamine, a prodrug of thiamine, raises blood levels by 50-100 times to achieve pharmacologic effects. It provides a novel therapeutic direction addressing a well-characterized brain tissue thiamine deficiency and related changes in glucose metabolism in AD. BenfoTeam is a seamless phase 2A-2B “proof of concept” (POC), double-blind, placebo-controlled RCT investigating tolerability, safety, and efficacy of benfotiamine, as a first-in-class small molecule treatment for early AD. Participants (n = 406) with MCI or mild dementia due to AD with a positive plasma biomarker test result (C2N PrecivityAD2) will be enrolled at ∼50 sites in the U.S. Phase 2A will evaluate safety and tolerability to determine which dose (600 mg or 1200 mg) to carry forward into phase 2B. This dose decision will be determined at the first of either 21 total adverse tolerability events (TEs) or alternatively when 160 person-months of exposure have accumulated (see Figure 1). If 1200 mg/day does not have a significantly higher TE rate than placebo, then this dose will be taken forward to phase 2B and tested against placebo without stopping the trial, otherwise 600 mg/day will be carried forward. Phase 2B will assess efficacy and longer-term safety of benfotiamine in the full study sample through 72 weeks of treatment, at the selected dose. Our trial modeling is presented in Figure 2. We project that to achieve our total sample size of 406 (203/arm), ∼ 1290 participants will need to be screened. With a 20% dropout rate, and a 5% drop-in rate from the placebo arm to active treatment, the trial will be powered at 80% overall (90% on each endpoint separately) to detect a difference between arms on co-primary endpoints (CDR-SB and ADAS-Cog 13) at 72 weeks, corresponding to a Cohen’s D effect size of 0.38 (small to moderate effect size). BenfoTeam deploys an innovative seamless phase 2A-2B design to achieve its overall aim of clinical POC. With its early adaptive dose decision, exposure will be optimized to the highest tolerated dose. A positive phase 2 POC trial that is powered to achieve a clinically significant benefit will further validate our approach.

Background:Repeated cognitive testing can boost scores due to practice effects (PEs), yet it remains unclear whether PEs persist across multiple follow-ups and long durations. We examined PEs across multiple assessments from midlife to old age in a nonclinical sample. Method:Men (N=1,608) in the Vietnam Era Twin Study of Aging (VETSA) underwent neuropsychological assessment comprising 30 measures across 4 waves (~6-year testing intervals) spanning up to 20 years. We leveraged age-matched replacement participants to estimate PEs at each wave. We compared cognitive trajectories and MCI prevalence using unadjusted versus PE-adjusted scores. Results:Across follow-ups, a range of 7-12 tests (out of 30) demonstrated significant PEs, especially in episodic memory and visuospatial domains. Adjusting for PEs resulted in improved detection of cognitive decline and MCI, with up to 20% higher MCI prevalence. Conclusion:PEs persist across multiple assessments and decades underscoring the importance of accounting for PEs in longitudinal studies.

BACKGROUND:Normative references for neuropsychological tests commonly adjust for years of education. However, quality of formal education varies widely, and additional education may be obtained outside of the classroom setting. Therefore, years of education may inadequately reflect full educational achievement. Grade-equivalency scores from tests of single-word reading provide additional information about learning experience and may alter interpretation of test results (Manly et al., 2002). Therefore, we explored the impact of adjusting cognitive test results for years of education vs. estimated grade equivalency in Spanish-speaking U.S. Latinos. METHOD:140 Spanish-speaking Latinos with subjective cognitive decline were referred by a community-based neurologist for objective cognitive testing. Participants were aged 47-88 (M=70.6, SD=8.0) with an average of 9.1 self-reported years of education (SD=4.5, range=0-20). Participants completed a cognitive screening battery using tests from the National Alzheimer's Coordinating Center Uniform Data Set Neuropsychological Battery (UDS3-NB). The Woodcock-Muñoz Letter-Word Identification Test was used to estimate grade-equivalency. Standardized scores (Z-scores) that adjusted for age, sex, language of testing, and either self-reported years of education or Woodcock-Muñoz grade-equivalency were obtained for each cognitive test using the UDS3-NB Latino Norms Calculator (Marquine et al.,2023). The two sets of Z-scores were compared using paired samples t-tests. RESULTS:Grade equivalency estimated by the Woodcock-Muñoz (M=13.5, SD=4.6; range=1.7-18) was significantly higher than self-reported years of education (M=9.1, SD=4.5; t [139]=10.9, p < .001). Figure 1 shows the distribution of the discrepancy between grade equivalency and self-reported educational attainment. For all neuropsychological tests, adjusted Z-scores derived using grade equivalency were significantly lower than those derived using years of education (see Table 1). The percentage of participants who fell below z-score cutoffs of -1, -1.5, or -2 was greater when adjusting for grade-equivalency than when adjusting for years of education (Table 2). CONCLUSIONS:Education-adjusted norms appeared to underestimate level of impairment in Spanish-speaking U.S. Latinos who presented to a neurologist with cognitive complaints. Self-reported years of education may not adequately reflect the educational experiences of this diverse group. Although additional validation is needed, these results indicate caution is warranted when interpreting education-adjusted test scores for Spanish-speaking U.S. Latinos with lower levels of education.

Subjective memory decline (SMD) is a putative early indicator of Alzheimer's disease (AD). However, its association with changes in objective memory and AD-related brain structure is mixed, which may be partly due to reliance on single-item measures of subjective decline. The Everyday Cognition scale (ECog) was developed to improve detection of early AD risk by leveraging multiple items, focusing on behaviors rather than general perceptions, and including informant ratings. Data were collected from cognitively unimpaired men at waves 3 and 4 (mean ages 68 and 74) of the Vietnam Era Twin Study of Aging (VETSA). The ECog memory subscale, assessing SMD over 10 years, was completed by participants (wave 3: n = 904; wave 4: n = 516) and informants (wave 3: n = 811; wave 4: n = 473). Participants also completed objective memory testing, summarized as latent factor scores. For a subset completing MRI (wave 3: n = 388; wave 4: n = 191), AD brain signature scores were derived from hippocampal volume and cortical thickness in seven AD-related regions. Structural equation models estimated associations of ECog-SMD with objective memory and AD brain signature scores at wave 3, and predictive associations between ECog-SMD at wave 3 and changes in objective memory and AD brain signature scores by wave 4. Associations between changes in ECog-SMD and changes in objective memory and AD brain signature scores were also examined. Covariates included age 20 cognitive ability, depressive symptoms, and state anxiety. Shown in Figures 1 and 2, ECog-SMD at wave 3 was associated with worse objective memory (participant: b=-.59, p < .001; informant: b=10.99, p = .001) and lower AD brain signature scores (participant: b=-1.86, p = .007; informant: b=-1.71, p = .018). ECog-SMD also predicted declines in objective memory (participant: b=-.33, p < .001; informant: -5.50, p s < .001) and reductions in AD brain signature (participant :b=-1.10, p < .001) by wave 4. At wave 4, increases in ECog-SMD correlated with declines in objective memory (participant: b=-.08, p = .007; informant: b=-.12, p = .013) but were not associated with changes in AD brain signatures (participant and informant: p s > .05). Participant and informant ECog ratings of SMD reflected concurrent differences in objective memory decline and AD-related neurodegeneration. Participant-reported SMD were associated with future AD-related neurodegeneration, potentially capturing risk undetected by neuropsychological testing or informant reports.

The CAN-THUMBS-UP Brain Health Support Program (BHP; NCT05347966), a year-long web-based dementia risk reduction study, included two self-administered digital cognitive assessments as exploratory outcomes. Cogniciti's Brain Health Assessment (BHA), a web-based screening battery, was completed every 6-months. MyCogHealth, a smartphone-based ecological momentary assessment, was deployed in 7-day “bursts” every 3-months. Here, we compare these novel assessment approaches to a traditional neuropsychological battery (NTB) delivered remotely at study baseline and endpoint. Global composites from the NTB, BHA, and MyCogHealth were analyzed using mixed models for repeated measures (MMRM) covarying for age, sex, education, and cognitive status. Time was modeled categorically. A lifestyle composite, derived from seven questionnaires completed in three-month intervals, was modeled as a time-varying covariate. We used structural equation modeling of BHA and MyCogHealth composites to investigate coupled changes in cognition and lifestyle (dual-growth models). MMRM analyses revealed significant effects of age (β=-0.307 to -0.142, p  <0.001), education (β=0.066 to 0.095, p  <0.05), and cognitive status (β=-0.292 to -0.176, p  <0.001) on baseline cognition across measurement approaches. There were no consistent effects of time on cognition, or of covariates on change over time. The lifestyle composite was not associated with cognition. Dual growth models confirmed baseline associations of MyCogHealth and BHA with age, education, and cognitive status (ps<0.01). Longitudinally, the BHA remained stable, whereas MyCogHealth improved (β=0.342, p  <0.001) with quadratic attenuation (β=-0.049, p  <0.001). Education was marginally associated with improvement on MyCogHealth (β=0.029, p  = 0.047). Age negatively predicted MyCogHealth intercept (β=-0.350, p  <0.001), but also predicted more improvement (β=0.046, p  = 0.009) and faster attenuation over time (β=-0.009, p  = 0.007). Males trended less improvement (β=-0.047, p  = 0.09) and slower attenuation (β=0.014, p  <0.05). The lifestyle composite intercept was positively associated with the MyCogHealth intercept (β=0.042, p  = 0.01) and with improvement on MyCogHealth (β=0.011, p  <0.05). No coupled changes were observed between MyCogHealth and lifestyle slopes ( p  = 0.25). The BHSP digital cognitive outcomes were similar to the NTB in their baseline associations with participant characteristics. The increased measurement frequency permitted by digital assessments revealed subtle improvements and cognition-lifestyle associations not detectable by the NTB. These findings support digital assessments as flexible alternatives to traditional cognitive endpoints in lifestyle-based dementia risk reduction studies.

INTRODUCTION:The EXERT study (Exercise in Adults with Mild Memory Problems) was a Phase 3, multicenter, randomized controlled trial that examined effects of exercise on cognition and other measures of brain health in sedentary older adults with amnestic mild cognitive impairment (MCI). METHODS:Participants were randomized to moderate-high intensity aerobic training (AX) or low-intensity stretching/balance/range of motion (SBR) for 18 months. Exercise was supervised for the first 12 months. Assessments were administered at baseline and every 6 months. The primary outcome was a global cognitive composite. RESULTS:A total of 296 participants were enrolled, and intervention adherence was high (supervised session attendance: AX = 81%, SBR = 87%). Intervention effects on cognition did not differ for AX and SBR (regression = -0.078, standard error [SE] = 0.074; p = 0.3). Notably, there was no 12 month  cognition decline for either group, and mean 12 month  hippocampal volume loss for both groups was low at 0.51%. DISCUSSION:Exercise intensity did not differentially affect cognitive trajectory. Intervention delivery was successful (high adherence) and cognition remained stable over 12 months  for both MCI groups, an association that warrants further study. HIGHLIGHTS:Exercise in Adults with Mild Memory Problems (EXERT) was a large multisite randomized controlled trial of moderate-high intensity aerobic training versus lower-intensity flexibility and balance exercise in sedentary older adults with amnestic mild cognitive impairment (MCI). A sensitive and validated measure of global cognitive function, the Alzheimer's Disease Assessment Scale-Cognition supplemented with tests of executive function (ADAS-Cog-Exec), was used to assess intervention efficacy with 12 months of supervised exercise. There was no intervention group difference on the 12-month cognitive trajectory of the ADAS-Cog-Exec. Intervention delivery was successful (high adherence), and cognition remained stable over 12 months for both exercise groups. Regular supported moderate-high or lower-intensity exercise may stall decline in adults with amnestic MCI, but further investigation is needed.