The Berlin Aging Studies (BASE & BASE-II)

Research Scientists
Julia Delius
Ulman Lindenberger (Project Leaders)

Andreas M. Brandmaier

The Berlin Aging Study (BASE)
The Berlin Aging Study II (BASE-II)
Overviews

In the course of the 20th century, average life expectancy almost doubled. What do these added years mean in terms of functional capacity and quality of life? And how do the last years and months preceding death in old age differ from the years before? For well over three decades, the Berlin Aging Studies have helped to answer questions of this sort. Members of the Center have been investigating changes in cognition and other aspects of behavior in the context of the Berlin Aging Study (BASE; Baltes & Mayer, 1999; Lindenberger et al., 2010) and, later, the Berlin Aging Study II (BASE-II; Demuth et al., 2021). Both studies are collaborative and multidisciplinary, involving researchers from institutions inside and outside Berlin. The Berlin Aging Studies also forms part of the Lifebrain consortium, which was funded under Horizon 2020, the EU Framework Programme for Research and Innovation, and continues to be active as a consortium beyond its funding period (e.g., Ravndal et al., 2025).

The Berlin Aging Study (BASE)

More than 35 years ago, 516 Berliners aged 70 years and over participated in the first measurement occasion of BASE. Longitudinal data are available for eight measurement occasions spanning more than 18 years, and mortality-related information was updated at regular intervals until the last participant passed away in 2019 (see below for an overview). BASE data continue to be the source of original publications on individual differences in late-life development (e.g., Ghisletta et al., 2025), and are often included in large-scale multi-site data analyses (e.g., Willroth et al., 2025). Furthermore, the availability of similar or identical measures in BASE and BASE-II has permitted the investigation of cohort differences in various aspects of normal aging, such as cognitive decline (Gerstorf et al., 2023) and loneliness (Suanet et al., 2024).

Verbal Fluency Selectively Predicts Survival in Old and Very Old Age

Paolo Ghisletta and co-authors (2025) made use of the complete mortality data for BASE participants to evaluate the connection between late-life cognition and survival. Using data collected at all eight times of measurement with the BASE cognitive battery across up to 18 years (i.e., nine tasks of perceptual speed, episodic memory, verbal fluency, and verbal knowledge, and a resulting general composite intelligence score), they applied a joint multivariate longitudinal survival model to estimate the unique contribution of each cognitive measure in terms of true current value and current rate of change when predicting survival. Only the two verbal-fluency measures, Categories and Word Beginnings, were uniquely predictive of mortality risk, and more predictive of mortality risk than the composite intelligence score (see Figure 1). The Categories measure required participants to name as many different animals as possible within 90 s, and the Word Beginnings measure required participants to name as many different words starting with the letter s as possible within 90 s. In addition to their advantageous psychometric properties—a wide performance range and a lack of floor and ceiling effects—, fluency tasks span search, word retrieval, and working memory skills that rely on efficient interactions among a variety of brain regions, such as prefrontal and limbic areas, which might explain their superior association to mortality.

Figure 1. Estimated survival probabilities (Kaplan-Meier method) for the entire BASE sample as a function of two tests of verbal fluency, Categories (naming animals) and Word Beginnings (naming word starting with the letter “s”). The upper dotted line represents 1 interquartile range above the median, and the lower dashed line represents 1 interquartile range below the median; the middle, continuous line is for the entire sample. The survival time was 3.01 years for those with low category values (about 11 animals), and 11.99 years for those with high category values (about 33 animals; see lower dashed line and upper dotted line, respectively, in Fig. 1a). For Word Beginnings, the low and high median survival times were estimated at 3.18 and 11.99 years (Fig. 1b), respectively. Thus, there was an almost 9-year predicted difference in median survival time between Berlin Aging Study participants with high versus low values on verbal fluency. IQR = interquartile range.

Image: Psychological Science
Ghisletta et al., 2025
Image licensed under CC BY 4.0

Figure 1. Estimated survival probabilities (Kaplan-Meier method) for the entire BASE sample as a function of two tests of verbal fluency, Categories (naming animals) and Word Beginnings (naming word starting with the letter “s”). The upper dotted line represents 1 interquartile range above the median, and the lower dashed line represents 1 interquartile range below the median; the middle, continuous line is for the entire sample. The survival time was 3.01 years for those with low category values (about 11 animals), and 11.99 years for those with high category values (about 33 animals; see lower dashed line and upper dotted line, respectively, in Fig. 1a). For Word Beginnings, the low and high median survival times were estimated at 3.18 and 11.99 years (Fig. 1b), respectively. Thus, there was an almost 9-year predicted difference in median survival time between Berlin Aging Study participants with high versus low values on verbal fluency. IQR = interquartile range.

Image: Psychological Science
Ghisletta et al., 2025
Image licensed under CC BY 4.0

Key References

Ghisletta, P., Aichele, S., Gerstorf, D., Carollo, A., & Lindenberger, U. (2025). Verbal fluency selectively predicts survival in old and very old age. Psychological Science, 36(2), 87–101. https://doi.org/10.1177/09567976241311923

MPG.PuRe

DOI

publisher-version

supplementary-material

Willroth, E. C., Beck, E., Yoneda, T. B., Beam, C. R., Deary, I. J., Drewelies, J., Gerstorf, D., Huisman, M., Katz, M. J., Lipton, R. B., Muniz Tererra, G., Pedersen, N. L., Reynolds, C. A., Spiro, A., Turiano, N. A., Willis, S., Mroczek, D. K., & Graham, E. K. (2025). Associations of personality trait level and change with mortality risk in 11 longitudinal studies. Journal of Personality and Social Psychology, 128(2), 392–409. https://doi.org/10.1037/pspp0000531

MPG.PuRe

DOI

Open Science: BASE Data at ZPID

In line with the Open Science guidelines of the Institute, access to the BASE data for the academic community has been facilitated greatly by making it available at the Leibniz Institute for Psychology (ZPID) in 2025. ZPID is the public open science institute for psychology in German-speaking countries. Julia Delius and Maike Kleemeyer, a former project member who was appointed the MPI for Human Development’s research data management coordinator in March 2021 (cf. Research Data Management), worked on this project together with Katarina Blask at ZPID. The dataset (with English translations) is now publicly accessible to international researchers. Self-selected data can be provided upon submission of an abstract and approval by the BASE speaker.

BASE at ZPID: https://rdc-psychology.org/en/reuse-data/base

Screenshot of the BASE page — *Figure 2.* Screenshot of the page to search for, and request BASE data at ZPID.

Image: ZPID

*Figure 2.* Screenshot of the page to search for, and request BASE data at ZPID.

Image: ZPID

The Berlin Aging Study II (BASE-II)

The central objective of the multidisciplinary and multi-institutional longitudinal Berlin Aging Study II (BASE-II; see overview below) is to promote a better understanding of individual differences and trajectories in cognitive, psychosocial, and physical functioning. Since 11/2024, the Center for Environmental Neuroscience has taken over the coordination of BASE-II at the Institute from LIP.

During the reporting period, over 30 BASE-II research articles with LIP scientists as authors have been published. Highlights include articles on the contribution of dopaminergic and noradrenergic brain regions to late-life memory performance (Dahl et al., 2023) and the etiology of sex differences in Alzheimer's disease prevalence (Ravndal et al., 2025).

Building on earlier BASE-II work showing that region-specific latent factors of grey-matter integrity are positively associated with episodic memory ability among older adults, Köhncke et al. (2024) investigated cross-sectional associations between general factors of grey-matter integrity and a general factor of cognitive ability in a sample of 1,466 BASE-II participants aged 60–88 years, 319 of whom contributed imaging data. Indicator variables based on T1-weighted images (voxel-based morphometry, VBM), magnetization-transfer imaging (MT), and diffusion tensor imaging-derived mean diffusivity (MD) had sufficient portions of variance in common to establish latent factors of grey-matter structure for a comprehensive set of brain regions. Individual differences in grey-matter factors were positively correlated across neocortical and limbic areas, allowing for the definition of second-order, general factors for neocortical and regions of interest (ROIs), respectively. Both general grey-matter factors were positively correlated with general cognitive ability (see Figures 3a and 3b). For the basal ganglia, the three modality-specific indicators showed heterogenous loading patterns, and no reliable associations of the general grey-matter factor to general cognitive ability were found. BASE-II researchers are currently following up on these cross-sectional analyses by seeking to establish latent change factors for both grey-matter structure and cognition.

*Figure 3. Multi-trait multi-method models of cross-sectional associations between individual differences in grey-matter integrity and general cognition.* Figure 3a shows the submodel for neocortical regions of interest (ROIs), and Figure 3b shows the submodel for limbic ROIs. Rectangles represent observed variables, circles latent variables. Single-headed arrows represent regression parameters and loadings, double-headed arrows variances (where both arrows point to the same variable) and covariances (where they connect two variables). Parameter estimates are standardized. Residual variances of the observed variables, variances and covariances of the covariates, and loadings of the methods factors were freely estimated but estimates are not shown here for sake of clarity. EM = episodic memory; VLMT = Verbal Learning and Memory Task; FP = Face–Profession task; SE = Scene Encoding task; OL = Object Location task; WM = Working Memory; LU = Letter Updating; SU = Spatial Updating; NNB = Number N-back; R = Reasoning; FA = Figural Analogies; PR = Practical Problems; LS = Letter Series; PFC = prefrontal cortex; PRE = precuneus; PAR = parietal cortex; OCC = occipital cortex; HES = Heschl’s gyrus; PHG = parahippocampal gyrus; HC = hippocampus; ACC = anterior cingulate cortex; AMY = amygdala; INS = insula; _V = voxel-based morphometry; _MT = magnetization transfer imaging; _MD = mean diffusivity.

Image: Psychological Science
Köhncke et al., 2024
Image licensed under CC BY 4.0

*Figure 3. Multi-trait multi-method models of cross-sectional associations between individual differences in grey-matter integrity and general cognition.* Figure 3a shows the submodel for neocortical regions of interest (ROIs), and Figure 3b shows the submodel for limbic ROIs. Rectangles represent observed variables, circles latent variables. Single-headed arrows represent regression parameters and loadings, double-headed arrows variances (where both arrows point to the same variable) and covariances (where they connect two variables). Parameter estimates are standardized. Residual variances of the observed variables, variances and covariances of the covariates, and loadings of the methods factors were freely estimated but estimates are not shown here for sake of clarity. EM = episodic memory; VLMT = Verbal Learning and Memory Task; FP = Face–Profession task; SE = Scene Encoding task; OL = Object Location task; WM = Working Memory; LU = Letter Updating; SU = Spatial Updating; NNB = Number N-back; R = Reasoning; FA = Figural Analogies; PR = Practical Problems; LS = Letter Series; PFC = prefrontal cortex; PRE = precuneus; PAR = parietal cortex; OCC = occipital cortex; HES = Heschl’s gyrus; PHG = parahippocampal gyrus; HC = hippocampus; ACC = anterior cingulate cortex; AMY = amygdala; INS = insula; _V = voxel-based morphometry; _MT = magnetization transfer imaging; _MD = mean diffusivity.

Image: Psychological Science
Köhncke et al., 2024
Image licensed under CC BY 4.0

Overviews

Overview of the Berlin Aging Study (BASE) – www.base-berlin.mpg.de
The multidisciplinary Berlin Aging Study (BASE), initially directed by the late Paul B. Baltes and Karl Ulrich Mayer, was started in 1989. Ulman Lindenberger is the current BASE speaker. The study spans eight measurement occasions spaced over 18 years. Its distinguishing features include (1) a focus on the very old (70–100+ years); (2) a locally representative sample, stratified by age and sex; and (3) a broad-based interdisciplinarity (originally involving two research units from the Freie Universität Berlin, Internal Medicine and Psychiatry, and two from this Institute, Sociology and Psychology). In addition to discipline-specific topics, four integrative theoretical orientations guide the study: (1) differential aging, (2) continuity versus discontinuity of aging, (3) range and limits of plasticity and reserve capacity, and (4) aging as a systemic phenomenon.
The initial focus of BASE (1990–1993) was to obtain a heterogeneous sample, stratified by age and sex, of individuals from the western districts of Berlin aged 70 to 100+ years. A core sample of 516 men and women completed the Intensive Protocol comprising detailed measures from all four participating disciplines. Seven longitudinal follow-ups involving different depths of assessment were completed at approximately 2-yearly intervals. Details of the research design and assessment protocols can be found on the BASE website. The core sample formed the basis of the analyses reported in two monographs (see Baltes & Mayer, 1999; Lindenberger et al., 2010). Current work uses the longitudinal data to address issues such as variability and change, mortality prediction, self-related change, and genetic and socioeconomic predictors of cognitive change. The data are available for research purposes at ZPID: https://rdc-psychology.org/en/reuse-data/base.

The Berlin Aging Study: International Research Group

Name	Affiliation
Julia A. M. Delius	MPI for Human Development, Berlin, Germany
Alexandra M. Freund	University of Zurich, Switzerland
Denis Gerstorf	Humboldt-Universität zu Berlin, Germany
Paolo Ghisletta	University of Geneva, Switzerland
Christiane Hoppmann	The University of British Columbia, Vancouver, Canada
Ulman Lindenberger	MPI for Human Development, Berlin, Germany (Speaker)
Nilam Ram	Stanford University, USA
Jacqui Smith	University of Michigan, Ann Arbor, USA
Ursula M. Staudinger	TUD Dresden University of Technology, Germany
Elisabeth Steinhagen-Thiessen	Charité – Universitätsmedizin, Berlin, Germany
Gert G. Wagner^†	MPI for Human Development, Berlin, Germany (Max Planck Senior Research Fellow)

Lifebrain – www.lifebrain.uio.no
BASE and BASE-II participated in Lifebrain, a consortium of European studies funded by the EU Framework Programme Horizon 2020. Its goal was to harmonize, enrich, and fully exploit some of the largest longitudinal studies of age effects on brain, cognition, and mental health in Europe. Its focus was on the effect, over time, of education, socioeconomic status, and lifestyle factors such as sleep, physical activity, and diet on potentially enhancing and protecting brain structure and function. EU funding of the consortium ended in 2023. Lifebrain continues to analyze multi-site data and to publish papers in high-ranking journals.

Key References
Fjell, A. M., Røgeberg, O., Sørensen, Ø., Amlien, I. K., Bartrés-Faz, D., Brandmaier, A. M., Cattaneo, G., Düzel, S., Grydeland, H., Henson, R. N., Kühn, S., Lindenberger, U., Lyngstad, T. H., Mowinckel, A. M., Nyberg, L., Pascual-Leone, A., Solé-Padullés, C., Sneve, M. H., Solana, J., … Vidal-Piñeiro, D. (2025). Reevaluating the role of education in cognitive decline and brain aging: Insights from large-scale longitudinal cohorts across 33 countries. Nature Medicine, 31, 2967–2976. https://doi.org/10.1038/s41591-025-03828-y

Lövdén, M., Pagin, A., Bartrés-Faz, D., Boraxbekk, C.-J., Brandmaier, A. M., Demnitz, N., Drevon, C. A., Ebmeier, K. P., Fjell, A. M., Ghisletta, P., Gorbach, T., Lindenberger, U., Plachti, A., Walhovd, K. B., & Nyberg, L. (2023). No moderating influence of education on the association between changes in hippocampus volume and memory performance in aging. Aging Brain, 4, Article 100082. https://doi.org/10.1016/j.nbas.2023.100082

Vidal-Piñeiro, D., Sørensen, Ø., Strømstrad, M., Amlien, I. K., Baaré, W., Bartrés-Faz, D., Brandmaier, A. M., Cattaneo, G., Düzel, S., Ghisletta, P., Henson, R. N., Kühn, S., Lindenberger, U., Mowinckel, A. M., Nyberg, L., Pascual-Leone, A., Roe, J. M., Solana-Sánchez, J., Solé-Padullés, C., … Fjell, A. M. (2025). Vulnerability to memory decline in aging revealed by a mega-analysis of structural brain change. Nature Communications, 16, Article 11488. https://doi.org/10.1038/s41467-025-66354-y