We are only beginning to understand the extent and reasons for the individual variation in postprandial response dynamics in key hormones and metabolites such as insulin, glucose, and triacylglycerols (TAGs). More nuanced statistical approaches for postprandial curve analyses, coupled with phenotyping of curve-associated factors, may help uncover the underlying physiology and possible associations with disease risk. In a clinical trial of 190 adults (90 males and 100 females) with abdominal obesity (age: 21-56 yr, BMI: 26-54), we used functional principal component analysis (FPCA) to examine postprandial serum glucose, C-peptide, and TAGs after a 4-h standardized mixed meal test. Correlations were explored between identified curve patterns and more than 100 anthropometric and fasting biochemistry traits. Postprandial curve levels, peaks, and dips varied substantially. Although the primary patterns in postprandial glucose, C-peptide, and TAGs uncovered by FPCA corresponded with area under the curve (AUC), the presence and timing of curve peaks and dips were uncorrelated with AUC. Males had higher postprandial levels and larger variation than females. Over 40% of the participants had nonsynchronized postprandial glucose and C-peptide curves. A postprandial phenotype characterized by elevated and delayed postprandial peaks in glucose and C-peptide, and a high postprandial TAG level and peak, was strongly associated with multiple blood biomarkers and anthropometric traits linked to insulin resistance and liver pathologies (correlation range: -0.49, 0.63). Although our findings support the general usefulness of postprandial AUC, FPCA provided detailed insight into individual postprandial glucose and insulin dynamics, which may inform improved diagnostics and personalized dietary advice.NEW & NOTEWORTHY Functional principal component analysis (FPCA) revealed large individual and sex-specific differences in postprandial glucose, C-peptide, and triacylglycerol dynamics in obesity. Although the primary postprandial patterns correlated strongly with the area under the curve (AUC), the presence and timing of peaks varied independently of AUC. Over 40% of participants showed nonsynchronized postprandial glucose and C-peptide curves. High and delayed peaks were associated with markers of insulin resistance, altered amino acid metabolism, visceral adiposity, and liver dysfunction.

NASCAR is the premier American motorsport, with millions of fans who tune in each week to watch drivers compete in a test of skill, endurance, and above all, speed. As with any sport, winning is paramount, and prior research has focused on modeling driver and team characteristics, as well as strategies to optimize race-day performance. However, no prior work has investigated ways to improve driver qualifying performance, a key factor that would naturally improve the likelihood of achieving a strong finishing position in the race that follows. To address this gap, we analyze qualifying lap data from the NASCAR AdventHealth400 using functional principal components analysis followed by agglomerative hierarchical clustering. This allows us to uncover distinct groups of drivers and extract typical behavior. We identify several braking, throttle, and steering strategies that are differentially associated with qualifying performance. By isolating the highest-performing clusters, we offer actionable insights that can be used to enhance qualifying efforts and, ultimately, race-day results.

Many circulating biomarkers are assessed at different time intervals during clinical studies. Despite of the success of standard joint models in predicting clinical outcomes using low-dimensional longitudinal data (1-2 biomarkers), significant computational challenges are encountered when applying these techniques to high-dimensional biomarker datasets. Modern machine- or deep-learning models show potential for multiple biomarker processes, but systematic evaluations and applications to high-dimensional data in the clinical settings have yet to be reported. We aimed to enhance the scalability of joint modeling and provide guidance on optimal approaches for high-dimensional biomarker data and outcomes. We evaluated multiple deep-learning and machine-learning models using 24 clinical biomarkers and survival data from the SQUIRE trial, a phase 3 randomized clinical trial investigating necitumumab and standard gemcitabine/cisplatin treatment in patients with squamous non-small-cell lung cancer (NSCLC). Overall, we confirmed that longitudinal models enabled more accurate prediction of patients' survival compared to those solely based on baseline information. Coupling multivariate functional principal component analysis (MFPCA) with Cox regression (MFPCA-Cox) provided the highest predictive discrimination and accuracy for the NSCLC patients with AUC values of 0.7 - >0.8 at various landmark time points and prediction timeframes, outperforming recent advanced Transformer and convolutional neural network deep-learning algorithms (TransformerJM and Match-Net, respectively). In conclusion, we identified that MFPCA-Cox represents a robust and versatile joint modeling algorithm for high-dimensional biomarker longitudinal data with irregular and missing data, capturing complex relationships within the data, yielding accurate predictions for both longitudinal biomarkers and survival outcomes, and gaining insights into the underlying dynamics. ClinicalTrials.gov (NCT00981058; first posted on September 22, 2009).

Monitoring equity drawdown risk requires real-time indicators that can be implemented without look-ahead bias and that may add information beyond standard volatility measures. This study develops a leakage-safe residual-stress indicator from cross-sectional PCA reconstruction errors in U.S. sector excess returns. Using daily adjusted prices for SPY and 11 U.S. sector ETFs, sector excess returns are computed relative to SPY, the common component is estimated with principal component analysis (PCA), and residual stress is defined as the cross-sectional root-mean-square magnitude of out-of-sample reconstruction residuals. The PCA mapping is estimated using information available only through t−1, the stress score is computed at t, and high-stress regimes are defined using rolling train-only quantile thresholds shifted forward by one trading day. The results show that realized volatility remains the stronger standalone benchmark in overall early-warning classification performance. Residual stress is therefore not proposed as a replacement for volatility. Instead, it is most useful as a complementary indicator of cross-sectional market dislocation. In the baseline sample, residual-stress spikes cluster around several drawdown-onset episodes, and conditional regime analysis shows that when volatility is low, high residual stress is associated with a higher probability of a drawdown onset within the next H=21 trading days than the low-stress/low-volatility regime. Event-overlap and lead-time diagnostics suggest that residual stress can identify some onset episodes not captured by a simple volatility-threshold rule, although its main incremental value lies in conditional risk stratification rather than systematically earlier triggering. The contribution of the paper is to develop a leakage-safe and interpretable residual-stress diagnostic for conditional drawdown-risk monitoring. The evidence supports a balanced interpretation: residual stress adds state-dependent information beyond standard volatility measures, especially in otherwise low-volatility states, but it does not dominate realized volatility as a standalone predictor.

Cognitive decline is a central challenge of aging, with subtle early changes laying the foundation for broader difficulties later in life. One domain that is particularly challenging to capture with standard assessments is processing efficiency. Previous research has shown age-related differences in processing efficiency using redundant-target detection tasks, but it remains unclear whether individual differences in cognitive ability within the older adults are associated with differences in processing efficiency. In the present study, 65 cognitively healthy older adults (aged 60-79) completed the Montreal Cognitive Assessment (MoCA) and a color-shape redundant-target detection task, from which we estimated resilience capacity (Rz), a processing efficiency metric that quantifies how well a system maintains its target processing speed in the presence of distractors, using Systems Factorial Technology (SFT). MoCA scores were significantly and positively correlated with the standardized resilience capacity summary, Rz (r = 0.35, 95%CI [0.12, 0.55], p = 0.004). This significant association persisted in a partial correlation analysis that controlled for age as a covariate (partial r = 0.35, p = 0.004). In a direct model comparison of four candidate processing-efficiency metrics - inverse efficiency scores (IES), redundancy gains (RG), mean RT of correct responses, and Rz - Rz was the strongest predictor of MoCA. Functional principal component analysis (fPCA) of R(t) identified a temporal component (PC2) on which individuals at the higher end of the MoCA distribution showed a later, more controlled rise in capacity, whereas those at the lower end showed earlier but less efficient processing. Together, these findings indicate that processing efficiency metric - and specifically resilience capacity under distractor interference - is continuously related to cognitive performance in older adults and may reflect aspects of cognitive reserve not captured by global screening scores or summary-statistic-based efficiency measures.

Recurrent health events often involve complex inter-relationships between longitudinal biomarkers and time-to-event outcomes, further complicated by sparse, irregular data collection and time-dependent correlations among events. Traditional statistical methods frequently struggle with these complexities, resulting in biased estimates and suboptimal modeling performance. To address these challenges, we propose the Functional Regression with AutoregressIve fraiLTY (FRAILTY) method, a novel framework designed to jointly model longitudinal measurements and recurrent events, accommodating both scalar and functional covariates while capturing time-dependent correlations among events. The FRAILTY method employs a two-step estimation procedure. First, functional principal component analysis through conditional expectation (PACE) is applied to extract key temporal features from sparse and irregular longitudinal data. Second, the obtained scores are incorporated into a dynamic recurrent frailty model with an autoregressive structure to account for within-subject correlations across recurrent events. Simulation studies demonstrated that the FRAILTY method outperformed existing methods, such as those relying on B-spline basis functions and Bayesian joint modeling, by achieving lower integrated mean squared errors, higher concordance indices, and greater statistical power in detecting functional parameters. Its practical utility was further validated through applications to two datasets: the Systolic Blood Pressure Intervention Trial study and the Multicenter Collaboration to Study Treatment Outcomes in Nephrolithiasis Evaluation cohort.

This article introduces a robust approach to functional principal component analysis (FPCA) for relative data. While recent papers have studied relative functional data within the Bayes space framework, there has been limited focus on developing robust methods to effectively handle anomalous observations and large noise. To address this, we extend the Mahalanobis distance concept to Bayes spaces, proposing its regularized version that accounts for the constraints inherent in relative functional data. Based on this extension, we introduce a new method, robust relative principal component analysis (RRPCA), for more accurate estimation of functional principal components in the presence of outliers. The method’s performance is validated through simulations and real-world applications, showing its ability to improve covariance estimation and principal component analysis compared to traditional methods.

Functional data analysis (FDA) provides a powerful statistical framework for analyzing complex data, such as curves or functions, over high-dimensional domains. In this paper, we focus on functional predictor regression (scalar-on-function) models applied to the prediction of birth weight using maternal dietary patterns during pregnancy. Specifically, we analyze trajectories of nine components of the Alternative Healthy Eating Index (AHEI) as multivariate functional predictors. We adopt Multivariate Functional Principal Component Analysis (MVFPCA) to obtain low-dimensional, uncorrelated representations of the multivariate functional predictors through their FPC scores. Building on MVFPCA, we develop a novel multivariate functional principal component regression (MVFPCR) model to predict birth weight effectively. Our method accommodates various regression approaches, including linear and quantile regression, depending on the distribution of birth weights. Through simulation studies and the application to a fetal growth study dataset, we demonstrate the effectiveness of our proposed model in functional predictor regression.

Abstract Background Physical activity and fatigue patterns in pulmonary arterial hypertension (PAH) are heterogeneous and incompletely captured by traditional clinical assessments. Actigraphy provides objective, continuous measurement of real-world activity. ACTiPH is an ancillary study of the Pulmonary Hypertension Association Registry (PHAR) that uses actigraphy to objectively characterize physical activity patterns in PAH. We applied functional data analysis to identify behavioral activity phenotypes and compared their clinical characteristics, risk profiles, and health-related quality of life (HRQoL). Methods Participants wore the ActiGraph hip based accelerometer during waking hours for 7 consecutive days, with wear periods within 14 days of a standardized PHAR assessment. Minute-level tri-axial accelerometry data were collected, yielding 10,000 minutes of data per participant. To address the high dimensionality and temporal structure of these data, minute-level activity counts were analyzed using functional principal component analysis. The leading components captured variation in overall activity amplitude (PC1), timing of peak activity (PC2), and activity fragmentation or midday fatigue (PC3). Unsupervised k-means clustering was applied to FPCA scores to identify distinct activity phenotypes. Clinical characteristics, patient-reported outcomes, and PH risk scores were compared across clusters using Kruskal-Wallis and chi-square tests. Analyses were restricted to the baseline actigraphy assessment per participant. Results Among 504 participants, four behavioral clusters were identified: Cluster 1 (low activity, early peak), Cluster 2 (low activity, delayed peak), Cluster 3 (moderate activity, early peak), and Cluster 4 (high activity, late peak). Lower activity clusters (Clusters 1 and 2) were older (60-66 vs. 53-55 years;p=0.01) and less likely to be married (36-48% vs. 53-54%; p = 0.0008), with higher proportions reporting never being married or being widowed or divorced. Primary PH diagnosis differed across clusters (p = 0.007), with connective tissue disease-associated PAH more common in lower activity clusters and idiopathic PAH more frequent in higher activity clusters. Functional status and risk profiles varied significantly, with more active clusters showed a lower frequency of elevated EmPHasis-10 scores, a greater proportion with higher SF-12 physical component scores, longer six-minute walk distances, and a lower proportion classified as higher risk by REVEAL Lite 2.0 and COMPERA 2.0 (all p < 0.0001). Conclusion Digital phenotyping using functional data analysis identifies distinct activity patterns in PAH associated with differences in clinical risk, function, and HRQoL. Wearable monitoring offers complementary insights beyond clinic-based assessments and may inform personalized care. Ongoing work will evaluate prognostic significance and temporal stability. Ongoing analyses are evaluating the prognostic significance of these phenotypes and the temporal stability of cluster membership over time. This abstract is funded by: NHLBI – R01 – HL159997 (PI: Kawut), NHLBI – K23 – HL169930 (PI: Minhas)

We and others have shown that maternal hyperglycemia during pregnancy in women with insulin-dependent diabetes mellitus (IDDM) influences fetal growth. Less is understood regarding how trimester-specific glycemic patterns, particularly glucose variability, shape offspring obesity risk across the life course and whether this is mediated by birthweight. Leveraging data from the Diabetes in Pregnancy Program Grant (PPG; 1978-1995) and the Transgenerational Effects on Adult Morbidity (TEAM Study; 2017-2023) cohort, we aimed to evaluate whether the association between maternal glycemic control and adult offspring obesity status was mediated in part through infant birthweight. Maternal glycohemoglobinA1 levels were collected monthly during pregnancy and harmonized across laboratories using standard deviation units (HbA1SD). Functional principal component analysis characterized patterns in blood glucose level and variability. TEAM Study participants, adult offspring of PPG women, completed in-person or online assessments (via Zoom) of body anthropometrics. Linear mixed-effects models and generalized estimating equations estimated associations between maternal glycemia and adult offspring outcomes. Classical mediation methods tested whether birthweight mediated observed relationships. Consistent with prior findings, third-trimester HbA1SD demonstrated the most consistent positive associations with adult offspring body mass index (BMI) outcomes, after adjustment for maternal BMI at last menstrual period, maternal education, gestational weight gain, and sex of offspring (for offspring weight only). There was no evidence that birthweight mediated the relationship between maternal glycemic patterns and adult offspring overweight and obesity; however, birthweight exhibited an independent direct effect on adult offspring BMI in fully adjusted models. Infant birthweight was not shown to be a mediator in the association of maternal gestational glycemic control and overweight/obesity in adult offspring of women with IDDM. Although birthweight does not appear to mediate these long-term associations, the findings underscore the importance of trimester-specific evaluation of blood glucose level and variability, motivating further investigation into transgenerational metabolic risk pathways. · It is unlikely that birthweight mediates the association between gestational blood glucose levels and offspring adult obesity.. · Higher birthweight is an important risk factor for overweight and obesity in the adult offspring.. · Early gestation hyperglycemia may program obesity in adult offspring of IDDM women..

Distinguishing cancerous from non-malignant breast cells using viscoelastic creep and functional principal component analysis.

Statistical analyses of morphological variations of three Larroussius (Diptera: Psychodidae) sister species collected in leishmaniasis endemic foci of Adana in Türkiye.

Abstract The present study aimed to assess the early functional responses of plant communities to extreme anthropogenic disturbances caused by war. To this end, the case of the Kakhovka Dam destruction in June 2023 was examined. The functional structure of plant communities in the first year following the event on Khortytsia Island in the lower Dnipro floodplain (southern Ukraine) was analysed. The research identified 146 species of vascular plants and employed multivariate analysis, utilising functional diversity indices and principal component analysis. Hemeroby and naturalness indices were incorporated to distinguish between anthropogenic and natural influences. The study’s results revealed the presence of five distinct axes of variation in functional community structure. Disturbed areas exhibited increased functional redundancy and evenness, driven by ruderal species dominance and loss of ecological dominants. The phenomenon of functional richness and specialisation exhibited a response to variations in moisture levels, while alterations in functional identity reflected shifts in pollination strategies. The findings indicated a close association between hemeroby and functional redundancy and evenness alterations. The spatial patterns observed across the island reflect a complex interaction between human impacts and natural moisture gradients. This study is among the first to document rapid, trait-based vegetation responses to wartime ecosystem disruption. The study emphasises the efficacy of functional diversity and hemeroby as mechanisms for assessing ecological stability in conflict-affected regions.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder whose molecular mechanisms involve multiple biological pathways. Longitudinal blood-based omics data, such as lipidomics and metabolomics profiles, offer promising noninvasive biomarkers for early detection and prognosis, yet they are high-dimensional, sparse, and exhibit complex temporal and cross-feature correlations. The primary goal of this study is to identify which omics data types are most strongly associated with time to dementia onset in patients with mild cognitive impairment (MCI) at baseline. To address this, we propose a longitudinal sparse single-omics factor analysis (LS-SOFA) framework that models each omics view through view-specific latent factors and feature-weight matrices, with temporal dynamics captured by functional principal component analysis (FPCA). The resulting functional principal component (FPC) scores are incorporated into a survival model to test whether each omics view is associated with time to dementia onset. An efficient covariance-based estimation algorithm substantially reduces computational and memory cost, enabling large-scale application in the Alzheimer's Disease Neuroimaging Initiative (ADNI) study. In simulations, LS-SOFA achieves higher longitudinal estimation accuracy and more stable hypothesis testing than competing methods. Applied to five blood-based omics views from ADNI, LS-SOFA identified plasma lipidomics and serum metabolomics from FIA and UPLC as significantly associated with dementia risk after FDR adjustment, with nominal evidence of association for gut microbial metabolomics from serum. The top features within each omics view reveal biologically interpretable metabolic pathways that may serve as blood-based biomarkers for AD progression.

In this article, we study scalar-on-function regression with functional covariates observed through replicated measurements subject to measurement error. Treating replicated curves as surrogates of an underlying latent process, the proposed framework resolves the identifiability issues commonly encountered in functional measurement error models. Through functional principal component analysis, the model is represented as a finite-dimensional hierarchical linear measurement error model. Parameter estimation is carried out using an expectation-maximization algorithm, and alternative correction strategies based on adjusted regression calibration and simulation extrapolation are also considered for comparison. Simulation studies demonstrate the advantages of explicitly accounting for measurement error in terms of bias reduction and estimation stability. An application to soybean yield prediction in Illinois, using meteorological variables contaminated by measurement error, illustrates the practical value of the proposed approach.

BackgroundSalivary agglutinin, also known as deleted in malignant brain tumors 1 (DMBT1), is an anti-microbial protein. DMBT1 is low in saliva from patients with oral squamous cell carcinoma (OSCC) and dramatically increases after treatment, with accompanying microbial changes. While this suggests an association between DMBT1 suppression and changes in the oral microbiota, causation has not been established. DMBT1 is also a tumor suppressor protein; its loss promotes OSCC progression, but its role in OSCC development is unknown. In this study, OSCC development was investigated in a murine carcinogen model that simulates human OSCC. Microbiota were standardized between Dmbt1 knockout (Dmbt1−/−) and wild-type (Dmbt1+/+) mice via interbreeding and co-housing. Saliva was collected at baseline and at 4, 8, 12, 16, and 22 weeks post-carcinogen initiation (stopped at 16 weeks). Tongues were harvested at week 22 for histopathology, and the salivary microbiome was profiled by 16S rRNA sequencing. Microbial diversity metrics and conditional dependence networks assessed community structure, while longitudinal patterns were analyzed using a locally sparse varying coefficient mixed model and functional principal component analysis (fPCA).ResultsDespite microbiota standardization, Dmbt1−/− and Dmbt1+/+ displayed differences in microbiome composition based on β-diversity metrics. At endpoint, carcinogen-treated Dmbt1−/− showed higher OSCC prevalence and more aggressive invasion than Dmbt1+/+. Several OTUs, including those from Lachnospiraceae, Sphingomonas, Carnobacteriaceae, and Candidatus Saccharibacteria families, demonstrated differential abundance patterns over time, either genotype-specific, diagnosis-specific, or both. Notably, Sphingomonas and Lachnospiraceae exhibited time-dependent abundance differences in mice that developed OSCC. fPCA identified taxa with abundance trajectories that were different between OSCC and precancer and genotype specific. ConclusionsThus, DMBT1 shapes salivary microbiota composition and protects against OSCC development. Dynamic, genotype-specific microbial shifts during carcinogenesis underscore the complex interplay between the oral microbiota and cancer progression.Video Graphical Supplementary InformationThe online version contains supplementary material available at 10.1186/s40168-026-02337-5.

There is an urgent need to identify clinical markers that can help physicians determine when additional treatment is necessary to manage the symptoms of knee osteoarthritis (OA). Patterns of physical activity that occur within a day, eg, low activity in the morning and/or evening, may be a novel means to identify treatment need, given that symptoms may reduce daily activity at specific times of the day. The purpose of this study is to explore the relationship between within-day patterns of physical activity and all-cause mortality in adults with or at high risk for knee OA. We performed a secondary analysis of the Osteoarthritis Initiative (NCT00080171). Our exposure was within-day patterns of physical activity using a multidimensional (14-hour) multilevel (4-day) functional principal component analysis to analyze accelerometer data from analytic baseline. The outcome was all-cause mortality assessed up to eight years. Kaplan-Meier survival curves and Cox proportional hazards regressions were used to calculate adjusted hazard ratios (aHR). There were 1,927 adults with or at high risk for knee OA included in this analysis. We identified four primary within-day activity patterns accounting for around 82% of sample variability. Participants who demonstrated low levels of activity in the morning and evening had 2.09 times the risk of death compared with those demonstrating the average activity pattern of the sample (aHR 2.09, 95% confidence interval 1.15-3.80). Unique within-day patterns of physical activity were associated with risk of death. Those with inactivity in the morning and evening were at increased risk for death.

Background: Infections caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have generated major public health concerns worldwide. Young adults represent a critical group for viral transmission due to their high proportion of asymptomatic infections. Objective: To characterize the dynamics of SARS-CoV-2-specific antibodies in individuals aged 20-29 years from Bogotá, Colombia, across two longitudinal phases. Methods: Phase I assessed seroprevalence, seroconversion, spatial clustering, symptoms associated with seropositivity and antibody kinetics following natural infection. Phase II evaluated vaccine-induced antibodies, immune memory, and neutralizing capacity. Analyses included Functional Principal Component Analysis, survival analysis, clustering, and predictive modeling. Results: In Phase I, a seroprevalence of 15.59% (17/109 participants enrolled) was observed, while seroconversion among those who completed all six sampling points was 30.18% (16/53), with clusters of positive cases in different areas of Bogotá. The symptoms most associated with seropositivity included mucus hypersecretion, fever, and respiratory difficulty. Antibody responses were heterogeneous: naturally infected individuals generally showed high titers during the first 1-2 months, remaining detectable up to 4 months. The reduction in dimensionality suggested dominant humoral patterns, and clustering revealed two immune profiles differing in the risk of seroconversion. Predictive modeling indicated diverse antibody trajectories over 12 months. In Phase II (2024), three long-term immune memory clusters (low, medium, high) were observed; post-vaccination IgG titers were observed, although in most cases they lacked neutralizing activity. Conclusions: This longitudinal exploratory observational study provides an initial characterization of antibody dynamics in young adults, suggesting their potential epidemiological relevance and offering preliminary insights into post-infection and post-vaccination immunity.

SpaFun is a novel, non-model-based method developed to address limitations in existing spatially variable gene detection techniques, particularly for large-scale spatially resolved transcriptomics datasets. These limitations include computational inefficiency, limited statistical power with increasing data size, and the inability to capture spatial heterogeneity and co-expression patterns among genes. Built on functional principal component analysis, SpaFun identifies domain-representative genes with significantly better computational efficiency and greater statistical power while accounting for spatial heterogeneity and co-expression patterns among genes. We applied SpaFun to three spatially resolved transcriptomics datasets and demonstrated that SpaFun outperformed state-of-the-art algorithms for identifying representative genes for tumor regions (e.g. DESeq, edgeR, and limma), as well as recently developed novel algorithms designed for spatial omics to identify the representative genes (e.g. SPARK and CSIDE). This highlights SpaFun’s ability to accurately identify genes most representative of each spatial domain (e.g. tumor, immune, or stroma regions). By uncovering novel disease-relevant genes overlooked by existing algorithms, SpaFun could provide insights into new molecular mechanisms and propose innovative therapeutic strategies to improve patient outcomes.

Background:Early identification of individuals at elevated risk for Alzheimer’s disease (AD) is critical for prevention. Blood-based biomarkers offer scalable alternatives to cerebrospinal fluid and imaging, but the prognostic value of longitudinal plasma lipid trajectories remains unclear.Objective:To evaluate whether multi-year plasma lipid trajectories improve prediction of AD conversion beyond demographic, clinical, genetic, and neuropsychological measures.Methods:We studied 1150 cognitively normal or mildly impaired Alzheimer’s Disease Neuroimaging Initiative participants; 329 progressed to AD dementia over a mean follow-up of 2.3 years. Plasma lipidomics quantified 749 lipid species by high-resolution LC–MS. Trajectories were summarized using functional principal component analysis and related to time to conversion using covariate-adjusted Cox proportional hazards models. Predictive performance was assessed by concordance index and likelihood-ratio tests.Results:Cross-sectional lipid profiles modestly improved prediction beyond demographic and clinical covariates, and longitudinal lipid trajectories yielded small additional gains. Ether-linked triglycerides showed the strongest longitudinal associations with conversion, with TG(O-50:1) [NL-18:1] exhibiting the most robust signal. Neuropsychological measures provided substantially stronger discrimination, and lipid trajectories added limited value once cognitive information was included. Nevertheless, longitudinal lipid changes contributed consistent improvements in models without neuropsychological predictors, supporting their role as complementary blood-based markers in settings where standardized cognitive assessments are unavailable or impractical.Conclusions:Plasma lipid trajectories capture biologically relevant metabolic change associated with AD progression and provide complementary predictive information. Longitudinal lipidomic profiling may support early risk stratification and cohort enrichment when cognitive assessments are unavailable.