Cognitive impairment in older adults is influenced by coexisting β-amyloid (Aβ), tau, and cerebral small vessel disease (CSVD). Cerebral microbleeds (CMBs) are associated with Aβ and CSVD, but their role on tau-related neurodegeneration remains unclear. We investigated whether the CMBs modify tau-related disease progression. A longitudinal, prospective cohort study was conducted involving participants with mild cognitive impairment, Alzheimer disease dementia from the memory disorder clinic of the single tertiary center, or cognitively unimpaired from the community. All participants underwent cognitive assessment, MRI, 18F-flutemetamol PET for Aβ, and 18F-MK-6240 PET for tau at baseline. Cognitive tests were performed annually and MRI at 2 years. Cognitive decline was defined by score changes over this period and cortical atrophy as annual cortical thickness change. Linear regression analyses were conducted after stratifying by total or lobar CMB presence. Among the 201 participants (mean age 71.3 ± 7.0 years, 66.7% female), 95 had CMBs and 106 did not. Baseline Aβ or tau burden did not significantly differ between the 2 groups while white matter hyperintensity volume and lacunes were greater in the CMB group. Cross-sectionally, greater tau burden correlated with worse cognition, as measured by the Clinical Dementia Rating Scale Sum of Boxes (CDR-SOB) or Mini-Mental State Examination (MMSE) in both groups. Longitudinally, baseline tau burden was associated with CDR-SOB progression in the non-CMB group (β = 1.558, SE = 0.249, p < 0.001), but not in the CMB group (β = -0.031, SE = 0.405, p = 0.940; p-for-interaction = 0.001). Similar group differences were found for MMSE changes (non-CMB: β = -2.365, SE = 0.566, p < 0.001; CMB: β = -0.816, SE = 0.653, p = 0.217; p-for-interaction = 0.073). Stratification by lobar CMBs confirmed significant interaction effects for both CDR-SOB (p-for-interaction = 0.007) and MMSE (p-for-interaction = 0.045) scores. Imaging analysis showed more extensive cortical atrophy in the CMB group, but tau-related cortical atrophy was widespread only in the non-CMB group and minimal in the CMB group. In the non-CMB group, tau burden was strongly associated with cognitive decline and cortical atrophy. By contrast, the CMB group exhibited greater CSVD burden and pronounced neurodegeneration not explained by tau, suggesting that additional mechanisms such as CSVD related to cerebral amyloid angiopathy or neuroinflammation may contribute to disease progression in this group.

BackgroundGait impairments in Alzheimer's disease (AD) and related dementias pose a major fall risk/contribute to morbidity/mortality. The Timed Up and Go (TUG) test is often used to assess mobility, gait changes, and dual‐task performance. The TUG‐Dual Task (TUG‐DT) version adds serial subtraction exercises to evaluate dual‐task cost (DTC). For cognitively unimpaired (CU) individuals or those with mild cognitive impairment (MCI), plasma biomarkers like pTau217, pTau181, and neurofilament light chain (NfL) can help assess the risk of AD. This study aimed to explore the relationship between TUG performance and plasma biomarkers in CU and MCI patients.MethodsParticipants included CU low‐risk (n = 75), CU high‐risk (n = 87), and CI (n = 32) older adults aged 55‐80, mean = .67.28 ± 6.062 years. Cognitive ability was assessed using the Clinical Dementia Rating Scale (CU = 0; CI = 0.5 or 1.0) and the Montreal Cognitive Assessment (CU ≥ 26; 18 ≤ CI ≤ 26). AD‐risk was determined by APOE genotyping and family history for CU groups. Plasma biomarkers pTau217, pTau181, and NfL were analyzed from fasting blood draws. Participants completed the TUG and TUG‐DT. ANOVAs, ANCOVAs, logistic regression, and generalized additive models (GAMs), were used to analyze the relationship between demographic factors, gait performance, and plasma biomarkers, with model comparisons guiding the final choice of GAMs for their flexibility in handling non‐linear relationships.ResultsStep count analysis on the TUG showed that the CU‐high‐risk and MCI groups performed similarly, while the CU‐low‐risk group completed significantly fewer steps than both. Plasma biomarkers, particularly pTau181 and NfL, interacted to predict gait performance only in the CU high‐risk group.ConclusionsThe TUG can predict plasma pTau217 levels with high specificity, distinguishing CU individuals not at risk for AD. Additionally, pTau181 and NfL interacted to predict performance on the TUG and TUG‐DT in the CU‐high‐risk group, suggesting subtle gait changes may signal early AD pathology. The CU‐low‐risk group's reduced step count compared to others indicates preclinical AD might manifest with subtle mobility impairments. These findings support using simple gait tasks like the TUG for AD risk assessment in older adults.

BackgroundSubjective Cognitive Decline (SCD) is considered to be the crucial predementia stage and serves as one of the early indicators of dementia. It refers to the self‐perceived decline in cognition without any evidence of objective decline. On the other hand, studies have found that poor sleep quality has a higher chance of developing dementia in the later stages. Thus, in the present study, we aim to look at the relationship between SCD and sleep and their effect on cognition.MethodThe present analysis was done using the baseline data of the Tata Longitudinal Study of Aging (TLSA) cohort. SCD was screened using a positive response to the memory question of the Geriatric Depression Scale (GDS) and scoring zero on the Clinical Dementia Rating (CDR) scale. An equivalent number of age, gender, and education‐matched healthy controls were selected for the study. The quality of sleep was assessed using the Pittsburgh Sleep Quality Index (PSQI) scale. Cognition was measured by the Hindi Mental Status Examination (HMSE) and Addenbrooke's Cognitive Examination (ACE‐III). Chi‐square test was done to check the association between SCD and sleep. The Generalized Linear Model (GLM) was used to assess the relationship between SCD and sleep with cognitive parameters.ResultThe study utilized 140 participants with a mean age of 58.37±9.71 years. Among them, 70 participants were screened as having SCD. Chi‐square test revealed that SCD participants have a higher proportion of poor sleep quality when compared to their healthy counterparts (68.8% vs 31.3%, p = 0.001). Unadjusted GLM revealed that the group with poor sleep and SCD scored lower in HMSE (β=‐0.414, p = 0.025). When adjusted for Generalized Anxiety Disorder scale and GDS scores, the group with poor sleep and SCD scored less in the ACE attention task (β=‐0.951, p = 0.018)ConclusionThe findings suggest that poor sleep quality can impact cognitive functioning in the early stage of SCD. Therefore, implementing strategies for improving sleep quality can aid in delaying the onset of objective cognitive impairment.

BackgroundHippocampal sclerosis (HS) is characterized by neuronal loss and gliosis in the cornu Ammonis (CA) region of the hippocampus and is associated with epilepsy, hypoxia, and neurodegenerative diseases. In dementia, HS is increasingly recognized as a potential biomarker for Limbic‐predominant Age‐related TDP‐43 Encephalopathy neuropathological change (LATE‐NC). However, LATE‐NC currently lacks a validated in vivo biomarker. Since HS is detectable by MRI, a key question is whether HS could serve as a reliable proxy for underlying LATE‐NC pathology. This study examines the prevalence and pathological associations of HS across neurodegenerative diseases in a large population‐based brain bank to assess its potential as a biomarker for LATE‐NC.MethodData were analyzed from the Biobank for Aging Study (BAS‐GEROLAB) in São Paulo, Brazil. Clinical and epidemiological information was obtained from next of kin using validated protocols, including the Clinical Dementia Rating (CDR) scale. Neuropathological assessments included morphological, vascular, and immunohistochemical analyses for beta‐ amyloid, tau, alpha‐synuclein, and TDP‐43. HS was defined as ≥70% neuronal loss in CA1 (unilateral). Individuals with a history of epilepsy were excluded.ResultAmong 1,307 individuals, 36.9% were non‐White and 49.7% were women. HS was identified in 41 participants (3.1%), with higher prevalence among older individuals, women, and those with lower education levels (p < 0.001). HS was also significantly associated with hyaline arteriolosclerosis (p = 0.001) and diabetes mellitus (p = 0.003) (Table 1). Multivariate analysis confirmed associations between HS and LATE‐NC (p < 0.001) as well as lacunar infarcts (p = 0.031) (Table 2). The positive predictive value (PPV) of HS for TDP‐ 43 pathology was 48.8%, increasing to 59.4% among cognitively impaired individuals (n = 432) (Table 3). Among individuals with HS but no TDP‐43 pathology (n = 21), 48.4% had Braak‐AD stage ≥3, and 64.5% exhibited arteriolosclerosis.ConclusionHS is relatively uncommon in this population‐based brain bank but is strongly associated with TDP‐43 pathology. However, vascular pathology also plays a significant role, particularly in populations with high cardiovascular risk. This limits its predictive value for LATE‐NC, even among cognitively impaired individuals. Given the absence of an in vivo biomarker for LATE‐NC, further research is needed to determine whether MRI‐detectable HS can serve as a reliable surrogate marker for LATE‐NC pathology, particularly in genetically diverse populations with multiple comorbidities.

EEG-based stratification in Alzheimer's disease: Cognitive progression, pathological marker associations, and therapeutic interventions.

BackgroundIn addition to the acute manifestations of COVID‐19, recent studies have shown that COVID‐19 may be associated with long‐term effects, including cognitive impairment (CI) and functional disability (FI). The aim of this study is to investigate the correlation between cognitive impairment and functional disability in elderly people after the diagnosis of COVID‐19.MethodCross‐sectional study nested in a prospective cohort, which recruited elderly people and their respective caregivers after the diagnosis of Covid‐19, confirmed in the laboratory, from 2021 to 2024. All participants signed the Free and Informed Consent Form (CAAE: 31799420.6.0000.5149). Cognition was measured by the Z‐score of cognitive performance on the Dementia Rating Scale and Functionality by the total Functional Independence Measure (FIM) and its subitems. Descriptive statistics, Chi‐square and Spearman correlation were used.ResultThe sample consisted of 80 participants. The mean time between the initial infection and the first evaluation was 10.09 ± 5.97 months (Min.: 2‐Max.: 24). 51.25% of the participants were female, while 57.50% were hospitalized, with a mean age of 73.91 ± 10.20 years; 8.03 ± 3.24 years of schooling, 6.54 ± 3.24 medications in use, Charlson Comorbidity Index of 1.79 ± 1.80 and 3.39 ± 1.63 doses of vaccine for Covid‐19 at the time of diagnosis. 70.87% reported memory difficulty after Covid‐19 and 54.43% of the sample had cognitive disability. The (FIM) indicated the presence of functional disability with a mean of 111.23 ± 11.62. Regarding the diagnosis, 15% had dementia; 62.5% mild cognitive impairment and 22.5% preserved cognition. A correlation was found between the presence of (CC) and (FI) after the diagnosis of COVID‐19 in 92% of the sample (p = 0.008). The total (FIM) (rho=‐0.374; p = 0.001) and the items related to memory (rho=‐0.374; p <0.001), social interaction (rho=‐0.372; p = 0.002) and problem‐solving (rho=‐0.366; p = 0.002) showed a weak correlation with cognitive impairment. Correlations were also found with the following items: bathing (rho=‐0.248; p = 0.038), urine control (rho=‐0.236; p = 0.049), transfers (rho=‐265; p = 0.026) and gait (rho=‐0.270; p = 0.024).ConclusionUnderstanding the relationship between (CC) and (FI) can help in the development of more effective prevention and rehabilitation strategies, contributing to improving the health of the elderly population and their caregivers.

BackgroundLecanemab, a monoclonal antibody targeting amyloid, has been approved as disease‐modifying treatment for Alzheimer’s disease (AD) and is gradually gaining regulatory approvals across the globe. However, its accessibility and applicability in low‐and middle‐income countries (LMICs) remains limited due to constraints in biomarker availability, cost effectiveness concerns and underrepresentation of diverse cohorts in clinical trials. This study aimed to evaluate the eligibility of patients with Mild cognitive impairment (MCI) and AD from Cognitive Disorders Clinic (CDC) for treatment with lecanemab, based on clinical, and imaging criteria used in clinical trials.MethodA retrospective review of patients diagnosed dementia was conducted using data from the CDC registry. Patients with MCI and AD dementia were included based on the Lecanemab drug trial’s inclusion criteria. Cognitive assessments including Mini Mental Status Examination (MMSE), Clinical Dementia Rating (CDR) scales and MRI were reviewed. Eligibility was assessed by applying the appropriate use criteria for Lecanemab to all patients with MCI/AD. Plasma and Brain amyloid biomarkers were not available for the patients.ResultOf the 1448 patients with dementia seen in CDC from May 2016 to May 2021, 643 (44.4%) were clinically diagnosed with MCI (n=252, 17.4%)/AD (n=350, 24.2%). Applying the CDR criteria excluded 43 MCI and 165 AD cases. An additional 20 MCI and 10 AD cases were excluded based on age criteria. Applying the MMSE criteria excluded 28 MCI and 132 AD cases. Additionally, 18 MCI cases were excluded due to the presence of cerebral microbleeds. The primary exclusion factors included, advanced disease stage (40%), and contraindications such as significant comorbidities (30%) and microbleeds 102 (20%). Thus, out of the total 1,448 dementia patients, 134 (9.25 %) and of 643 patients with clinical AD or MCI, 134 (20.8%) were potentially eligible treatment with Lecanemab.ConclusionA substantial proportion (80.75%) of dementia patients from LMICs may be ineligible for lecanemab due to diagnostic and resource limitations. However, 20.8% of clinically diagnosed AD could potentially be eligible with biomarker testing. Efforts to enhance biomarker availability, early detection, and cost‐effective strategies are crucial to enable equitable access to emerging AD therapies in LMICs.

BackgroundInformal family care partners are crucial in dementia care, serving as primary advocates and caregivers. A caregivers’ perception of their care recipient's dementia severity, whether they overestimate or underestimate the progression, can impact care satisfaction and applicability. Overestimation may lead to unnecessary interventions or distress, while underestimation can delay critical support. Caregivers offer valuable insights extending beyond clinical scores, capturing nuances essential for effective care planning. This study investigates the agreement between caregivers’ perception of dementia severity and clinical severity assessment, and how demographic and contextual factors influence accuracy.MethodData were collected from a convenience sample of 22 dyads. Caregivers were eligible if they were 18 years or older, spoke and read English, and provided unpaid care to a family member with dementia for at least 2 months. Caregivers completed surveys ranking their perception of their care recipient's dementia severity (e.g., mild, moderate, severe, very severe), along with various demographic and contextual factors. A collapsed version of the Clinical Dementia Rating (CDR) scale, administered by certified staff, objectively measured dementia severity. Agreement between caregivers’ perceived dementia severity and CDR severity was evaluated, and additional analyses identified predictors of accuracy.ResultForty‐five percent of caregivers (age = 62.8±14.7; 82% female) overestimated their care recipients’ (age = 81±9.6; 50% female) dementia severity (i.e., 70% of mild cases; 17% of moderate; 100% of severe), and eighteen percent underestimated severity (i.e., 17% of moderate cases; 75% of very severe). There was a significant positive association between caregivers’ perceived dementia severity and CDR severity (p = 0.041), indicating that perceived severity increased as CDR severity increased (OR=2.4). There was weak agreement between perceived and CDR severity (K = 0.1; p >0.05). No demographic or contextual factors influenced perception accuracy.ConclusionWhile caregivers’ perceptions of dementia severity often diverge from clinical assessments, they reflect the lived reality. Acknowledging the tendency to overestimate, improved communication and education between dyads and providers is vital. Dynamic knowledge sharing will help align expectations and ensure sufficient support is provided based on the needs of the care recipient, not solely their clinical score.

BackgroundCognitive reserve refers to the brain's capacity to maintain performance despite pathology and may influence the risk of cognitive decline. Therefore, studying the effect of cognitive reserve on older adults with memory complaints can provide insights into their cognitive trajectories and lifestyle factors that could impact on their cognitive deterioration.MethodsBRASCODE is an observational, prospective cohort study of cognitively unimpaired individuals with > 65 years‐old with cognitive complaints conducted at the Hospital de Clínicas de Porto Alegre, Brazil. The 1‐year follow‐up assessment began in March 2023 and is still ongoing. Sociodemographic data, Memory Complaint Scale (MCS), Mini‐Mental State Examination (MMSE), Clinical Dementia Rating (CDR) Scale and Cognitive Reserve Scale (CRS) were collected. In addition to the cognitive battery, patients were also assessed for behavioral and sleep symptoms using the Neuropsychiatric Inventory (NPI), the Mild Behavioral Impairment Checklist (MBI‐c), the Geriatric Anxiety Inventory (GAI), the Geriatric Depression Scale (GDS) and the Pittsburgh Sleep Quality Index (PSQI).ResultsTo date, 100 patients have completed their 1‐year assessment, with 75% being women. The median age was 71 years, 87% of participants were classified as having Subjective Cognitive Decline (SCD), and 28% as having low cognitive reserve (CR). The PSQI correlated positively with GAI scores (rho=0.254, p = 0.254). The subject MBI‐c correlated with global cognition (rho=0.225, p = 0.0273), and the informant MBI‐c correlated with the CDR‐sum of boxes (rho=0.282, p = 0.007). Compared to participants with high CR, those with low CR were more likely to be older, have fewer years of education, be male, and display significantly lower scores on global cognitive assessments (CDR and MMSE). In mixed linear regression, cognitive reserve did not significantly predict global cognition composite score (p = 0.419). However, education was a significant predictor (beta = ‐0.009; p = 0,0106).ConclusionOur results reinforce the impact of education and neuropsychiatric symptoms on cognitive performance in SCD. Follow‐up data and Alzheimer's Disease (AD) biomarkers analysis will help clarify the role of cognitive reserve on their cognitive trajectories.

BackgroundClinical Dementia Rating (CDR) and Hindi Mental State Examination (HMSE) are ubiquitously used by clinicians and researchers in India for quantifying and classifying severity of cognitive impairment. he clinical dementia rating (CDR) scale measures cognitive impairment in 6 domains, including memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care. It rates severity from 0 = normal to 3 = severe on the basis of functional and cognitive decline. The Hindi Mental Status Examination (HMSE), a 31‐point cognitive test designed for the Indian population, evaluates orientation, memory, attention, language, and visuospatial abilities. In this study, we examined the diagnostic concordance between CDR, HMSE and clinical evaluation (CE) by expert Geriatric Psychiatrists, in the NIH/NIA‐funded R01 project titled “India ENIGMA Initiative for Global Ageing and Mental Health” (R01AG060610).MethodData from 85 participants enrolled in the project were used for this study. The participants were clinically evaluated and diagnosed by expert psychiatrists; while CDR and HMSE were administered by trained and certified research fellows. The classification based on CE [ HE (n = 47), MCI (n = 24), and AD (n = 14)] by expert psychiatrists was considered as the ‘gold standard’. The overall agreement among CDR, HMSE, and CE was assessed using Fleiss' multi‐rater kappa, while Cohen's weighted kappa was used to evaluate agreement within individual categoriesResultClinical ratings of 85 participants using CA, CDR and HMSE showed “moderate” agreement (Fleiss kappa = 0.556; Z=11.885; p <0.001) with each other. The agreement between CA and CDR was “very good” weighted kappa=0.954; Z=10.994; p <0.001), and agreement with HMSE was “moderate” (weighted kappa=0.483; Z=6.041; p <0.001). The agreement between HMSE and CDR was “moderate” (weighted kappa=0.482; Z=5.99, p <0.001).Between all the three methods, the agreement on Healthy Elders was “very good” (weighted kappa=0.848; Z=9.311; p <0.001), agreement within the MCI category was “moderate” (weighted kappa=0.518; Z=6.171; p <0.001) and agreement within AD category was “good” (weighted kappa=0.784; Z=12.088; p <0.001)ConclusionThe results of the study highlight the excellent concordance between CDR‐derived diagnoses with CE‐based diagnoses, the gold standard, in this project that aims to examine the various factors contributing to the ‘brain age gap’ in AD and MCI in India.

BackgroundBPSD are common in persons living with dementia and are often associated with disease severity and caregivers’ distress. In Low‐ and Middle‐Income Countries, they are often the symptom complex that compel family to seek medical help. However, the association of BPSD with dementia severity and specific cognitive domains is underexplored in African populations. This study aims to evaluate BPSD among indigenous Africans and the association with specific cognitive domains.MethodThis cross‐sectional study analyzes data collected from nine African countries (Nigeria, Benin, Ghana, Cameroon, Ethiopia, Kenya, Uganda, Tanzania, and Mozambique) participating in the Recruitment and Retention for Alzheimer's Disease Diversity Cohorts in the Alzheimer's Disease Sequencing Project (READD‐ADSP) project, from June 2023 to December 2024. Data on neurocognitive factors were collected using the Uniform Data Set (UDS), Neuropsychiatric Inventory Questionnaire (NPI‐Q) for BPSD, and the Clinical Dementia Rating (CDR) scale for dementia severity. Data was managed on REDCap and analysis was performed using STATA (v16). Descriptive statistics were employed for symptom prevalence, and Spearman's rank correlation was used to analyze the relationship between NPI‐Q and CDR scores. Association between cognitive domains and behavioral symptoms was examined using hierarchical linear model adjusting for covariates such as age, gender, education, and illness duration.ResultA total of 641 Africans with dementia (61% female, mean age 76.7±10.1 years) were included. The mean (±SD) NPI and CDR scores were 4.91 (±6.25) and 1.84 (±1.25), respectively. The most prevalent BPSD symptoms were apathy (32%), agitation/aggression (28.7%), and depression (28.1%). Spearman's correlation revealed a weak but significant positive association between NPI‐Q severity scores and CDR scores (rho = 0.264, p < 0.001). Behavioral symptoms were significantly associated with dementia severity (β = 1.59, CI: 1.15–2.03) as well as specific cognitive domains, particularly memory (digit span backward: β = ‐0.38, CI: ‐0.64 to ‐0.13). However, the duration of illness was not a significant predictor of behavioral symptoms in this cohort.ConclusionThe findings suggest that BPSD are both an indicator and consequence of dementia severity, particularly in memory function. This study posits the need for targeted management of BPSD and improvement in overall patient care in LMICs

BackgroundFunctional impairment (FI) is a key component in diagnosing and monitoring the severity of neurodegenerative dementia (ND). It manifests in several dimensions, such as everyday cognition, mobility, social functioning, and activities of daily living (ADLs). However, traditional assessments have primarily focused on ADLs, often neglecting the other dimensions of functional phenotype.This study aims to assess the utility of the Chilean multidimensional functional assessment (Ch‐MFA, see Figure 1), including ADLs, social behavior, mobility, and everyday cognition in differentiating Alzheimer's disease (AD) and frontotemporal dementia (FTD), and to predict their severity measured in the Clinical Dementia Rating (CDR).Method308 patients were recruited in Peru, comprising 209 controls, 44 FTD, and 55 AD. Functional assessments include ADL (T‐ADLQ, DAD‐E), social behavior Blessed Dementia Rating Scale (BDR), mobility (WHODAS), everyday cognition (ECog), and CDR for severity. Data was normalized for each assessment, and groups were stratified regarding age and educational level, obtaining 72 controls, 24 FTD, and 45 AD for the predictive analysis. Random Forest (RF) models were trained using 70% of the data, while 30% was reserved for testing the prediction. Cross‐validation (20‐fold) and feature selection were used to reduce overfitting and identify a simpler set of questions.ResultThe Ch‐MFA indicates a 0.81 ± 0.03 average f1‐score when differentiating the group (control, FTD, and AD), which is 0.12 ± 0.10 better than the best‐performing unidimensional assessment (T‐ADLQ) with only 14 questions (4 T‐ADLQ, 2 BDR, 4 ECog and 3 WHODAS). As for the severity, Ch‐MFA shows a 0.65 ± 0.17 average f1‐score when differentiating severities of 0.5, 1, and 2, increasing to 0.88 ± 0.06 when it's only between 1 and 2 with a total of 4 questions (2 T‐ADLQ, 2 ECog). See Figure 2.ConclusionOur results show that the Ch‐MFA achieves better results than a unidimensional assessment. Interestingly, functional dimensions that distinguish types of dementia differed from those that predict their severity. Our results suggest that a multidimensional phenotype may better reflect the specificity of dementia subtypes.

BackgroundGlobal cognitive metrics are useful for tracking cognitive aging, yet there is no consensus on how to best define and measure global cognition. The aim of this study was to compare scoring methods for global cognition against validity indicators.MethodUsing harmonized factor scores of four cognitive domains (memory, executive function, language, visuospatial) in the United States and Korea (Table 1), we compared scoring methods for global cognition with validity indicators of cognition, functioning, and structural neuroimaging at baseline. Scoring approaches included combining cognitive domains with mean z‐scoring, congeneric confirmatory factor analysis (CFA; maximum likelihood robust estimation), principal component analysis (PCA), and weighting domains based on the Mini‐Mental Status Examination (MMSEComposite). Correlations (r) were used to compare composites against continuous validity indicators. The Clinical Dementia Rating Scale Sum of Boxes (CDR‐SB) and MMSE were used to measure cognition and functioning. Neuroimaging measures included brain volume relative to cerebrospinal fluid (BV/CSF) and intracranial volume‐normalized measures of brain volume (BV/ICV), ventricular volume (VV), frontal lobe volume (FL), and medial temporal lobe volume (MTL). Receiver operating characteristic (ROC) curves assessed the ability to detect cognitive impairment (CDR‐Global≥0.5).ResultCognitive composites were highly intercorrelated (Figure; r range:0.94,1.00), correlated with CDR‐SB (r range:‐0.72,‐0.65) and MMSE (r range:0.75,0.77), and detected cognitive impairment in the good range (AUCs=0.84‐0.88). Cognitive composites were similarly correlated with neuroimaging metrics, with differences of at most r = 0.02 across comparisons, except for FL, where MMSEComposite had an r 0.04 lower than the others. Composites yielded stronger correlations with most neuroimaging metrics compared to standalone measures except for BV/ICV with CDR‐SB (r = ‐0.27). MTL and VV demonstrated the strongest associations with cognitive composites and other indicators of cognition/functioning.ConclusionRegardless of scoring method, global cognitive composites were similarly related to validity indicators of cognition/functioning and structural neuroimaging in a harmonized sample. MMSEComposite showed the strongest relationships with CDR but was less strongly associated with frontal lobe volume. Composites demonstrated stronger relationships with structural neuroimaging compared to standalone measures. Future work aims to evaluate similarities and differences between cohorts and evaluate relationships longitudinally.

BackgroundDeclining communication is one of the hallmark symptoms of Alzheimer's disease and related dementias (ADRD). Not only is this symptom distressing for the person with ADRD, but care tasks may be more challenging as a result of communication deficits. While it is well known that hearing loss is strongly related to communication capacity, a lesser explored area related to communication is that of cortical‐level sensory processing in ADRD. We hypothesized communication impairment would be positively correlated with caregiver burden and sensory processing abnormalities in older adults with ADRD.MethodUsing data collected from a non‐pharmacological randomized controlled trial aimed at behavior modification in ADRD (participants with ADRD confirmed by Clinical Dementia Rating Scale [CDR] score of 1+ and their primary caregivers), this study conducted secondary data analysis using Pearson correlation to assess relationships among communication impairment as indicated on the CDR, caregiver burden (as measured by the Zarit Burden Inventory), and sensory processing abnormalities (as measured by the Adult Sensory Profile). Demographic data were assessed using summary statistical assessment.ResultData were analyzed from 19 participants with ADRD. Participants with ADRD consisted of 11 females, 8 males, x̄ age of 78.21 (SD=10.15), and x̄ Standard Global CDR of 1.63 (SD=0.84). All participants had functioning sensory acuity with or without aids (i.e., hearing aids). Care partners consisted of 16 females, 3 males, x̄ age of 62.32 (SD=11.56), and spouses were the most frequent care partners. Analyses indicated a significant strong positive correlation between communication impairment and caregiver burden (r = 0.59, p = 0.007). Additionally, communication impairment was significantly positively correlated with sensory processing abnormalities within the domains of sensory sensitivity (r = 0.64, p = 0.004) and sensory avoiding (r = 0.49, p = 0.037).ConclusionBoth receptive (hearing) and expressive (talking) communication is vital to encourage cooperative completion of care tasks. Communication impairment in ADRD is not only linked to increased caregiver burden, but also is associated with cortical‐level sensory processing abnormalities beyond standard hearing loss. Combined with hearing loss, communication is made even more difficult. Additional exploration is warranted to determine causal mechanisms between sensory processing abnormalities and communication impairment in ADRD.

BackgroundLimbic‐predominant age‐related TDP‐43 encephalopathy neuropathological change (LATE‐NC) frequently coexists with Alzheimer's disease (AD), Lewy body disease (LBD), and vascular pathology. While neuropsychiatric symptoms in AD/LATE‐NC are beginning to be explored, the impact of LATE‐NC in mixed AD/LBD and LBD is still underexamined. This study investigated neuropsychiatric symptom profiles in individuals with AD, LBD, and mixed AD/LBD with and without LATE‐NC.MethodIn this retrospective cohort study, individuals from the UBC Hospital Clinic for Alzheimer's and Related Disorders (UBCH‐CARD) autopsy database with AD, LBD or both plus LATE‐NC pathology (LATE+, n = 42) were matched by age and neuropathology to controls without LATE‐NC (LATE‐, n = 38). Neuropsychiatric symptoms were assessed using the Neuropsychiatric Inventory Questionnaire (NPI‐Q) at initial, intermediate, and final visits before death. Outcomes included symptom severity scores at the final visit, odds ratios of symptoms at the final visit and trajectories of symptom progression. Parametric and non‐parametric tests were used to compare symptom severity scores, while linear mixed‐effects models analyzed longitudinal symptom data, adjusting for age, sex, education, and Clinical Dementia Rating Scale (CDR).ResultMean age at death was 80.5 years (LATE+) and 77.9 years (LATE‐), and participants were followed on average 5.8 years. Both groups had a high burden of moderate‐to‐severe AD pathology (LATE+ 97.6% vs. LATE‐ 94.4%) and moderate numbers had neocortical LBD (31.7% vs. 41.7%, p = 0.50). At the final visit, total NPI‐Q severity scores were similar between groups (5.65 vs. 6.12, p = 0.71). Symptom severity scores for apathy (0.73 vs. 1.29, p = 0.01), hallucinations (0.24 vs. 0.82, p = 0.01), and appetite changes (0.08 vs. 0.41, p = 0.04) were lower in LATE+. Odds ratios for symptom presence at the final visit, adjusted for coexisting pathology, remained significant for hallucinations only (0.013, 95% CI 0.001‐0.218, p = 0.002). Longitudinal modeling showed no significant differences in total burden of neuropsychiatric symptoms.ConclusionLATE‐NC did not appear to increase overall neuropsychiatric symptom burden in this autopsy cohort study of individuals with AD and/or LBD. Hallucinations were less frequent in the LATE+ group, but this may be due to incomplete adjustment for coexisting Lewy Body pathology. Further modelling of longitudinal symptom progression will be presented at the conference.

BackgroundEach year, approximately 15% of elderly individuals with mild cognitive impairment (MCI) progress to Alzheimer's disease (AD) or related dementias. Predicting this progression is crucial for informing treatment decisions, identifying protective factors, and modeling treatment effects in clinical trials. Here we trained and tested a 3D convolutional neural network (CNN) to forecast future decline (over two years) in the Clinical Dementia Rating scale sum of boxes score (sobCDR).MethodsWe analyzed data from three independent, publicly‐available cohorts: the (1) Alzheimer's Disease Neuroimaging Initiative (ADNI; n = 1,136), (2) Open Access Series of Imaging Studies (OASIS‐3; n = 241), and (3) National Alzheimer's Coordinating Center (NACC; n = 942). Participants with a 3D T1‐weighted brain MRI, baseline age, sex, & BMI measures, baseline sobCDR (within 90 days of imaging), and a follow‐up sobCDR score (1.75‐2.25 years after baseline) were included.For evaluation, datasets were partitioned into five independent, cross‐validation folds with balanced sex, age at imaging, and change in sobCDR score (scale 0: no impairment – 18: severe impairment) between baseline and 2 years.A 3D DenseNet‐121 CNN was used to predict the sobCDR score 2 years after baseline, first by using only the T1‐weighted MRI. This predicted value (predCDR_2Y) was then included in four linear mixed‐effects models (LMEs) to predict the observed sobCDR at two years from baseline (obsCDR_2Y) that included additional demographic covariates, scanner manufacturer and participant cohort random effects (Table 1).For each linear model, we ran a corresponding null model with all the same covariates, but without the predCDR_2Y variable, to determine the added predictive value of the imaging. From the five cross‐validation folds, we calculated the Pooled, out‐of‐sample, R2 and its standard error (SE), adjusting for the number of explanatory variables in each model.ResultsAll predictive models predicted follow‐up sobCDR with mean absolute error of around 1 point; adding the image‐derived prediction improved accuracy (Table 1).ConclusionDeep learning shows promise for enhancing prognostic models; future work will include additional deep learning methods, more diverse neuroimaging data (amyloid and tau PET and diffusion MRI), and alternative data fusion methods to integrate tabular and neuroimaging data.

BackgroundInstrumental Activities of Daily Living (IADL) are key indicators of functional decline in dementia, but culturally adapted tools for Latin America are scarce. The Amsterdam IADL Questionnaire Short Version (A‐IADL‐Q‐SV) has shown strong psychometric properties globally but has not been validated in Peru. This study aims to assess the validity and robustness of the Peruvian version of Amsterdam (A‐IADL‐Q‐SV‐P) in classifying dementia severity and its agreement with other clinical scales.MethodsA total of 171 participants were recruited a cognitive clinic in Peru. Clinical dementia rating scale (CDR) was used for dementia severity (81 CDR=0, 70 CDR=1, 20 CDR=2) Of the dementia patients (CDR30), 54 had Alzheimer's disease and 36 had Frontotemporal dementia. Validity of the A‐IADL‐Q‐SV‐P was determined by ability to classify CDR level using quadratic discriminant analysis. The robustness of the scale was assessed by Spearman correlations with demographic variables (age, sex, and education). Further association with existing scales for cognitive [Mini‐Mental State Examination (MMSE)], mood [generalized anxiety disorder‐7 (GAD‐7), geriatric depression scale (GDS), and neuropsychiatric inventory (NPI)], and functionality [Pfeffer Functional Activities Questionnaire (PFAQ) and Technology‐Activities of Daily Living Questionnaire (T‐ADLQ)] status.ResultsThe A‐IADL‐Q‐SV‐P achieved excellent accuracy for classifying dementia severity (accuracy: 0.853, AUC>0.92). The ability to classify normal cognition had a sensitivity of 0.95 and a specificity of 0.91. These results were similar when investigating only the participants with either Alzheimer's disease or frontotemporal dementia (accuracy>0.88, sensitivity>0.95, and specificity>0.88, AUC>0.91). We also found weak associations with age, sex, and education (|rho|<0.2). Furthermore, the A‐IADL‐Q‐SV‐P had strong associations with other clinical scales (|rho|>0.4), particularly in relation to cognition (rho=‐0.70) and functionality (rho>0.87).ConclusionA‐IADL‐Q‐SV‐P demonstrated excellent accuracy in classifying dementia severity, with strong correlations to cognitive and other functionality measures. The weak associations with demographic variables and validity in multiple dementia subtypes suggest robustness across diverse populations, supporting its utility as a culturally adapted tool for assessing functional decline in dementia.

BackgroundThe study of the association of biomarkers and neuropsychological assessment can contribute to the understanding of the biological mechanisms associated with the evolution of the neurodegenerative process, as well as to the definition of risk factors for the conversion of patients with cognitive decline into dementia due to Alzheimer's disease (ADD). Our main objective was to determine the relationship between performance on the Dementia Rating Scale (DRS) and the main biomarkers of Alzheimer's Disease.MethodSixty‐four older adults from the Cog‐Aging Brazil cohort with low educational levels were recruited and split into three groups: Cognitively Unimpaired (CU, n = 18), Mild Cognitive Impairment (MCI, n = 21) and Alzheimer's Disease Dementia (ADD, n = 25). Levels of CSF Aβ42 and p‐Tau 181 were assessed by Luminex xMAP technique. Descriptive statistics were carried out according to normality tests. Spearman correlation tests were explored to analyze associations between CSF levels of Aβ42, p‐Tau181, and p‐Tau181/Aβ42 ratio and performance in DRS. Significant associations between p‐Tau181 and Aβ42 and total DRS score (DRST) were explored in linear regression analysis adjusted for age, sex, education, and APOE ε4 carrier status. This study was approved by UFMG Ethics Committee.ResultThe median age of the participants was 75 years (IQR 9.25). The education level was 4 years (IQR 4.25), DRST mean 117.73 (14.04). A negative significant correlation was found between CSF p‐Tau181 and the domains of DRS: Memory (rho ‐0.533) (p < .001) and DRST (rho ‐0.482) (p < .001). No significant correlation was found for Aβ42. A significant correlation between DRS and p‐Tau181/Aβ42 ratio was found in Memory (rho ‐0.430) (p < .001) and DRST (rho ‐0.349) (p < .05). The linear regression model was statistically significant (F = 8.81) (p < .001) with R2 of 0.495. The Linear regression analyses showed a significant association with DRS and p‐Tau181(β = ‐0.548; CI 95%: [ ‐0.260, ‐0.105]; (p < .001) and education (β = 0.420; CI 95%: [0.671, 2.159]; (p < .001).ConclusionThe higher levels of p‐Tau181 seem to correlate with a worse performance in Memory and total DRS. Considering DRST, p‐tau was still associated with worse performance, even when adjusted by sociodemographic factors.

BackgroundThere is a need for brief and scalable measures of dementia severity to determine treatment eligibility and inform clinical decision‐making. We assessed the psychometric properties and optimal cutoff scores of the Quick Dementia Rating Scale (QDRS) to predict Clinical Dementia Rating Scale (CDR) stages in a large, well‐characterized, and clinically diverse sample of older adults with and without neurodegenerative conditions.MethodThe 10‐item informant‐reported QDRS was obtained for 902 participants through the University of California San Francisco Alzheimer's Disease Research Center, 610 of whom had concurrent CDR data and 36 of whom had repeat QDRS data within 6 months. We assessed optimal cut‐points based on our sample using receiver operating characteristic (ROC) analyses including bootstrapped comparison of area under the curve (AUC) against previously published QDRS cut‐points that were based on 267 participants. Reliability (internal, test‐retest) and validity (criterion, convergent, divergent) of the QDRS were also assessed.ResultThe QDRS correlated strongly with the CDR sum of boxes (ρ = .87, p < .001, Figure 1) and significantly differed across CDR stages (F = 548.0, p < .001, Figure 2). It also had strong internal reliability (Cronbach's alpha = .94) and acceptable test‐retest reliability (ρ = .57). Our optimal cut‐point replicated the published cutoff of 1.5 (AUC = .92, Figure 3) to differentiate CDR 0 vs. 0.5 (AUC = .85). When differentiating CDR 0.5 vs. 1, our optimal cutoff was slightly higher (7.5 with AUC = .90 vs. 6.0 with AUC = .84) although when we limited out CDR 1 sample to individuals with an Alzheimer's disease clinical syndrome, a more similar cut‐point of 6.5 was found (AUC = .91, Figure 3). Recommendations for adjusting the cut‐point for less typical neurodegenerative presentations (e.g., linguistic, behavioral, motor predominant) will be presented at AAIC 2025.ConclusionThe QDRS offers an efficient and scalable measure of dementia severity that can be used as a reliable and valid clinical alternative to the CDR, including for potential determination of eligibility and clinical monitoring for emerging AD treatments.

BackgroundSubjective cognitive decline (SCD) and apolipoprotein E (APOE) ε4 allele are known risk factors for Alzheimer's disease (AD). However, the impact of APOE ε4 on cognitive progression in individuals with SCD remains unclear. The aim of this study is to evaluate whether SCD APOE ε4 carriers show greater cognitive decline over one year compared to non‐carriers.MethodWe used data from the BRASCODE cohort, including 142 participants with SCD divided into APOE ε4 carriers (ε4+) and non‐carriers (ε4−). Cognitive function was assessed using the Memory Complaint Scale (MCS) (both self‐reported and informant‐based), the Subjective Cognitive Decline Scale (SCD‐S), the Mini Mental State Examination (MMSE), the Clinical Dementia Rating (CDR) Scale, the Geriatric Anxiety Inventory (GAI) and the Geriatric Depression Scale (GDS). We applied linear mixed models using the baseline and the one year follow‐up data to evaluate the interaction between time and APOE ε4 in the cognitive progression of SCD patients.ResultIn the ε4− group, there were a total number of 97 participants (73.2% females), with a median age of 69 years [67, 72] and 16 years of formal education [11, 18]. In the ε4+ group, there were 42 participants (64.3% females), with a median age of 71 [67.25, 75] and 15 years of formal education [11, 17]. There was no statistically significant difference between the groups in cognitive test outcomes, sleep disturbances, anxiety, depression, or the global Z score. However, a tendency to higher prevalence of family history was observed in the ε4+ group (61.9%) compared to ε4− (42.3%), p = 0.052. Findings from the linear mixed model indicated that ε4+ was not a significant predictor of cognitive decline in SCD patients over one year, p = 0.217.ConclusionEven though APOE ε4 is a well established risk factor for the development of AD, our findings suggest that the presence of this allele was not associated with significant impact on cognitive progression over a one‐year period in individuals with SCD. This suggests that factors other than APOE ε4 may influence SCD trajectories, and longer follow‐up studies are needed to clarify this relationship.