Svenja Schwarck

Episodic memory declines early with ageing, reflecting reduced neural resources and diminished memory specificity. However, few studies have created a cognitive challenge with multiple levels of task demand to investigate this early, subtle decline. Furthermore, it is unclear whether the genetic Alzheimer’s disease risk factor APOE4 and early Alzheimer’s pathology constrain the neural resources required to cope with increasing task demands. In this preregistered behavioural study named EMCOMP (Episodic Memory & COMPensation), we conducted a sentence-based episodic memory retrieval task where we manipulated retrieval difficulty by varying the required degree of memory specificity, thereby increasing task demands on episodic retrieval with five demand levels at recognition (novel foils, ‘old’ target sentences and three lure levels) and two demand levels at free and cued recall (gist and details). We collected data from 100 cognitively unimpaired adults [37 young (mean age 24 years), 63 older (mean age 72 years)] with additional neuropsychological testing and blood-based measures available for the older group. First, we investigated differences in EMCOMP episodic memory retrieval accuracy and confidence between young and older adults in linear mixed-effects models. Second, we investigated differences associated with the APOE4 genotype and plasma-derived Alzheimer’s pathology. Third, we assessed correlations between EMCOMP recognition and recall and established cognitive tests. Young adults showed higher recognition accuracy and confidence as well as a higher recall score and a lower recall error rate compared to older adults. As target-lure similarity increased, there was a steeper decline in recognition accuracy and a steeper increase in high-confidence errors in older compared to younger adults and in older APOE4 carriers compared to older non-carriers. However, we found no associations with Alzheimer’s pathology. EMCOMP performance was positively correlated with established cognitive tests. Our study demonstrates age- and APOE4-related differences in episodic memory retrieval under increasing retrieval difficulty—differences that may not be detectable in paradigms with only one or two levels of variation. By manipulating sentence-based retrieval specificity, we provide a novel approach to detect subtle cognitive deficits and potential functional compensation in cognitively unimpaired older adults with risk factors or early Alzheimer’s pathology. Future research should extend this work to more diverse samples and combine behavioural assessment with functional magnetic resonance imaging (MRI) to examine underlying brain activation patterns and functional connectivity.

The medial temporal lobe (MTL) is crucial for episodic memory. Tau pathology is a hallmark of Alzheimer's disease (AD) and accumulates in layer-specific patterns in the MTL during aging. It is, however, unclear whether early AD pathology relates to mesoscale network signatures distinct from non-pathological aging. To address this gap, we acquired 7 Tesla submillimeter-resolution resting-state fMRI, plasma-based AD biomarkers, glial fibrillary acidic protein (GFAP) levels, APOE genotype, regional [18F]PI-2620 tau PET burden, and longitudinal episodic memory data in 75 cognitively unimpaired older adults. Older age was associated with lower perirhinal-hippocampal connectivity and lower network segregation, whereas higher plasma-based AD pathology was associated with higher perirhinal-hippocampal connectivity. Furthermore, temporal-lobe tau burden was related to altered connectivity patterns in tau-vulnerable MTL subfields and layers, dependent on GFAP levels. Retrosplenial tau burden was associated with higher hippocampal-retrosplenial connectivity consistent with tau spread along canonical hippocampal output pathways. Finally, higher connectivity within the hippocampus attenuated the negative association between temporal-lobe tau burden and memory performance but predicted unfavorable memory trajectories. Our findings show differential associations of age and AD pathology with mesoscale MTL-connectivity patterns. Importantly, increased hippocampal connectivity may support memory function in the short term while contributing to subsequent memory decline.

BACKGROUND: Episodic memory declines early with aging, reflecting reduced neural resources and diminished memory specificity. However, few studies have created a cognitive challenge with multiple levels of task demand to investigate this early subtle decline. Furthermore, it is unclear whether the genetic Alzheimer’s disease risk factor APOE4 and early Alzheimer’s pathology constrain the neural resources required to cope with increasing task demands. METHODS: In this preregistered behavioral study named EMCOMP (Episodic Memory & COMPensation), we conducted a semantic episodic memory retrieval task using a sentence-based task demand manipulation with five demand levels at recognition (novel foils, “old” target sentences, and three lure levels) and two demand levels at free and cued recall (gist and details). We collected data from 100 cognitively unimpaired adults (37 young (mean age 24 years), 63 older (mean age 72 years)) with additional neuropsychological testing and blood-based measures available for the older group. First, we investigated differences in EMCOMP episodic memory retrieval accuracy and confidence between young and older adults in linear mixed-effects models. Second, we investigated differences associated with the APOE4 genotype and plasma-derived Alzheimer’s pathology. Third, we assessed correlations between EMCOMP recognition and recall and established cognitive tests. RESULTS: Young adults showed higher recognition accuracy and confidence as well as a higher recall score and a lower recall error rate compared to older adults. As recognition-task demand increased there was a steeper decline in recognition accuracy and a steeper increase in high-confidence errors in older compared to younger adults and in older APOE4 carriers compared to older non-carriers. However, we found no associations with Alzheimer’s pathology. EMCOMP performance was positively correlated with established cognitive tests. CONCLUSION: Our study demonstrates age– and APOE4 -related differences in episodic memory retrieval under increasing task demands – differences that may not be detectable in paradigms with only one or two levels of task demand. By manipulating semantic retrieval specificity, we provide a novel approach to detect subtle cognitive deficits and potential functional compensation in cognitively unimpaired older adults with risk factors or early Alzheimer’s pathology. Future research should extend this work to more diverse samples and combine behavioral assessment with fMRI to examine underlying brain activation patterns and functional connectivity.

Background: Cognitive reserve (CR) and brain maintenance (BM) support preserved cognition by mitigating or slowing neural decline in aging. Physical fitness represents a plausible pathway to both CR and BM, yet the underlying neurobiological mechanisms remain insufficiently understood. This study examined the role of physical fitness in BM and CR by testing its associations with Alzheimer's disease related brain pathology and by assessing its moderating and mediating effects on the relationship between pathology and cognition. Methods: Data were obtained from 345 cognitively unimpaired older adults (mean age = 73.11, SD 8.03 years; 177 females) participating in the ongoing SFB1436 study. We assessed global cognition and delayed verbal memory performance, aerobic fitness (VO2max), and muscular capacity. Blood biomarkers included plasma Aβ1-42/Aβ1-40, p-tau217, and GFAP (glial fibrillary acidic protein), as well as serum BDNF (brain derived neurotrophic factor), VEGF (vascular endothelial growth factor), and Cathepsin-B (plasticity). Neuroimaging measures comprised medial temporal lobe (MTL) tau burden ([18F]PI-2620 PET), MRI-derived hippocampal volumes, white matter hyperintensities, perivascular spaces (PVS) in basal ganglia (BG) and centrum semiovale, total gray matter volume (GMV), and MTL thickness. To examine BM, we tested associations between fitness and brain pathology. To examine CR, we conducted moderation analyses assessing whether fitness attenuated the negative impact of MTL tau burden on cognition. Mediation analyses further evaluated whether hippocampal volume, total GMV, or MTL thickness mediated a potential association between fitness and cognition. All models controlled for age and sex (and education in mediation analyses), and multiple comparisons were FDR-corrected. Results: Physical fitness was not related to cognitive performance. Higher muscular capacity was associated with lower BG PVS volumes, while higher aerobic fitness was related to higher MTL thickness and total GMV. Elevated MTL tau burden was associated with poorer verbal memory performance. Although physical fitness did not significantly moderate the tau-memory relationship, model comparison indicated weak evidence for CR effects (p = .015). Conclusion: Muscular capacity was linked to lower BG PVS volumes, supporting resistance against pathology, while aerobic fitness related to preserved cortical integrity and tended to act as a CR proxy against MTL tau pathology. Physical fitness may support cerebrovascular and glymphatic function, providing weak CR-like protection against tau-related cognitive decline.

Background: Cognitive reserve (CR) and brain maintenance enable the brain to maintain performance despite injury and disease while also reducing neural decline by safeguarding brain structure and function. Physical activity is a potential pathway to BM and CR, as fitness relates to better cognition in older adults, though the underlying mechanisms remain unclear. To explore the role of physical fitness in BM and CR, we tested its association with brain pathology and its potential moderation of pathology's impact on cognitive performance. Method: We collected data from 167 cognitively unimpaired participants (mean age 71.57±7.50 years; 70 females) of the ongoing SFB1436 study ( www.sfb1435.de ; Figure 1). We collected many markers, including global and verbal cognitive performance; aerobic (VO _2max ) and muscular capacity; blood‐based biomarkers of Alzheimer's disease (plasma Aβ1‐42/1‐40, ptau217) and plasticity (serum BDNF, VEGF and Cathepsin‐B); PET‐derived medial temporal lobe tau burden (MTL DVR, ¹⁸ F‐PI‐2620 PET); and MRI‐derived volumes of hippocampi, white matter hyperintensities, and perivascular spaces (PVS) in the basal ganglia (BG) and centrum semiovale regions. The tests were two‐fold. We first tested whether fitness was associated with lower MTL DVR values, reduced MRI‐derived volumes of brain pathology, and better cognition. Using moderation analysis, we then tested whether physical fitness moderated the relationship between pathology and cognition. We relied on ANOVA for model comparison. We adjusted models for age and sex and FDR‐corrected multiple comparisons. Result: Participants with better aerobic capacity (VO _2max ) had lower BG‐PVS volumes (Figure 2a) and better global cognitive performance. Those with higher MTL tau burden had worse verbal memory (Figure 2b). We found no evidence of a relationship between physical fitness and Alzheimer's markers, plasticity‐related markers, or hippocampal volume. Moderation analysis revealed that physical fitness did not moderate the relationship between MTL tau burden and verbal memory, but model comparison revealed weak evidence for CR against MTL tau. Conclusion: We demonstrated that aerobic fitness is related to lower BG‐PVS volumes in old age and showed that aerobic fitness tends to act as CR proxy against MTL tau pathology. Aerobic fitness may help maintain cerebrovascular and glymphatic dysfunction in old age, thereby mitigating cognitive decline.

Resistance to age-related pathological changes (brain maintenance), including Alzheimer’s disease, cerebrovascular disease, and neurodegeneration may promote cognitive resilience in aging. However, how lifestyle and health profiles relate to successful cognitive and brain aging remains poorly understood. In a novel, deeply phenotyped cohort of 211 cognitively unimpaired older adults (age = 71.0 ± 7.4 years, 46% female), we characterized principal components of lifestyle and health using questionnaire, fitness, and blood data. We estimated cognitive age gap (CAG) based on comprehensive neuropsychological data and brain age gap (BAG) based on brain-pathology markers, including plasma biomarkers of Alzheimer’s pathology (pTau ₂₁₇ and Aβ _1-42 /Aβ _1-40 ), MRI-based measures of white matter hyperintensities, perivascular spaces, and brain atrophy. Regression analyses tested how the observed lifestyle-health profiles were related to CAG and BAG. Seven principal components explained 49% of the variance in health and lifestyle. The second component, characterized by a mentally and physically active life and low cardiovascular risk, was associated with lower CAG ( β = − 0.66, p < 0.001) and BAG ( β = − 0.52, p = 0.003), reflecting a younger-than-expected brain and cognitive age, respectively. The association of an active lifestyle and lower CAG was partially mediated by BAG. Higher CAG was also associated with other lifestyle components characterized by low mental stimulation. APOE-ε4 carriers exhibited higher BAG. In conclusion, a lifestyle combining low cardiovascular risk, high mental engagement throughout life and high physical activity/fitness is jointly associated with less-than-expected brain pathology and better-than-expected cognitive performance, supporting its involvement in brain maintenance and cognitive resilience to aging.

Background: Training studies report beneficial effects of physical (PP) on cognitive performance (COG) in older adults, but are often accompanied by potentially biased parameters, conclusions, and lack of directionality. To address these issues, we used a dynamic Bayesian approach to analyse the dynamic session‐to‐session change and coupling of PP and COG over time.Methods: We used two studies (N = 17 each): Study 1 contained 24‐weeks (72 sessions) of training of older adults with suspected Alzheimer’s disease (AD). Study 2 included four months (40 sessions) of training of older adults at metabolic risk. The hierarchical Bayesian continuous‐time dynamic modeling approach (R package ctsem) comprised: (i) A subject‐level latent dynamic model, (ii) subject‐level measurement model, and (iii) population model. The dynamic model was specified with two fully connected state variables enabling bidirectional coupling between PP and COG using default priors and starting values (4 chains and 8.000 iterations). Intercept and drift parameters were set to vary freely. Population and individual level parameters are estimated simultaneously using all data from all subjects. Second‐level models included MMSE, SF12 (physical health questionnaire) in study 1 and whole hippocampal volumes (wHCV: T2‐ASHS, longitudinal) and white matter hyperintensities (SAMSEG, longitudinal) in study 2 as time independent covariates.Results: Study 1: Higher PP was dynamically linked to COG (‐1.335, 95%‐BCI [‐1.725, ‐0.954]). The effect was short‐term, lasting up to five days (‐0.368, 95%‐BCI [‐0.479, ‐0.266]). Study 2: PP improved COG in subsequent sessions (‐1.11, 95%‐BCI[‐1.22, ‐0.99]), which lasted for up to 15 days. Higher wHCV was associated with a stronger coupling‐effect of PP on COG (‐0.15, 95%‐BCI[‐0.18, ‐0.11] ) and lower persistence of COG (‐0.145, 95%‐BCI[‐0.21, ‐0.08]).Discussion: Our results show immediate exercise‐induced improvements in COG, with a longer‐lasting effect for participants at metabolic risk compared to older adults with suspected AD. Despite impaired cognitive performance, the cognitive system was still able to fluctuate and change favorably. Physical exercise seems to mobilize capacities that would otherwise remain unused. Utilizing a Bayesian approach to examine time‐locked effects can provide insights for implementing individualized training approaches and guiding clinical decision‐making.

Background: Training studies report beneficial effects of physical (PP) on cognitive performance (COG) in older adults, but are often accompanied by potentially biased parameters, conclusions, and lack of directionality. To address these issues, we used a dynamic Bayesian approach to analyse the dynamic session‐to‐session change and coupling of PP and COG over time.Methods: We used two studies (N = 17 each): Study 1 contained 24‐weeks (72 sessions) of training of older adults with suspected Alzheimer’s disease (AD). Study 2 included four months (40 sessions) of training of older adults at metabolic risk. The hierarchical Bayesian continuous‐time dynamic modeling approach (R package ctsem) comprised: (i) A subject‐level latent dynamic model, (ii) subject‐level measurement model, and (iii) population model. The dynamic model was specified with two fully connected state variables enabling bidirectional coupling between PP and COG using default priors and starting values (4 chains and 8.000 iterations). Intercept and drift parameters were set to vary freely. Population and individual level parameters are estimated simultaneously using all data from all subjects. Second‐level models included MMSE, SF12 (physical health questionnaire) in study 1 and whole hippocampal volumes (wHCV: T2‐ASHS, longitudinal) and white matter hyperintensities (SAMSEG, longitudinal) in study 2 as time independent covariates.Results: Study 1: Higher PP was dynamically linked to COG (‐1.335, 95%‐BCI [‐1.725, ‐0.954]). The effect was short‐term, lasting up to five days (‐0.368, 95%‐BCI [‐0.479, ‐0.266]). Study 2: PP improved COG in subsequent sessions (‐1.11, 95%‐BCI[‐1.22, ‐0.99]), which lasted for up to 15 days. Higher wHCV was associated with a stronger coupling‐effect of PP on COG (‐0.15, 95%‐BCI[‐0.18, ‐0.11] ) and lower persistence of COG (‐0.145, 95%‐BCI[‐0.21, ‐0.08]).Discussion: Our results show immediate exercise‐induced improvements in COG, with a longer‐lasting effect for participants at metabolic risk compared to older adults with suspected AD. Despite impaired cognitive performance, the cognitive system was still able to fluctuate and change favorably. Physical exercise seems to mobilize capacities that would otherwise remain unused. Utilizing a Bayesian approach to examine time‐locked effects can provide insights for implementing individualized training approaches and guiding clinical decision‐making.

Training studies typically investigate the cumulative rather than the analytically challenging immediate effect of exercise on cognitive outcomes. We investigated the dynamic interplay between single‐session exercise intensity and time‐locked recognition speed‐accuracy scores in older adults with Alzheimer's dementia (N = 17) undergoing a 24‐week dual‐task regime. We specified a state‐of‐the‐art hierarchical Bayesian continuous‐time dynamic model with fully connected state variables to analyze the bi‐directional effects between physical and recognition scores over time. Higher physical performance was dynamically linked to improved recognition (−1.335, SD = 0.201, 95% Bayesian credible interval [BCI] [−1.725, −0.954]). The effect was short‐term, lasting up to 5 days (−0.368, SD = 0.05, 95% BCI [−0.479, −0.266]). Clinical scores supported the validity of the model and observed temporal dynamics. Higher physical performance predicted improved recognition speed accuracy in a day‐by‐day manner, providing a proof‐of‐concept for the feasibility of linking exercise training and recognition in patients with Alzheimer's dementia.Highlights: Hierarchical Bayesian continuous‐time dynamic modeling approach A total of 72 repeated physical exercise (PP) and integrated recognition speed‐accuracy (IRSA) measurements PP is dynamically linked to session‐to‐session variability of IRSA Higher PP improved IRSA in subsequent sessions in subjects with Alzheimer's dementia Short‐term effect: lasting up to 4 days after training session

Background: Training studies typically investigate the cumulative rather than the analytically challenging immediate effect of exercise on cognitive outcomes. Methods: We investigated the dynamic interplay between single-session exercise intensity and time-locked cognition in older adults with suspected Alzheimer’s dementia (N = 17) undergoing a 24-week dual-task regime. We specified a state of the art hierarchical Bayesian continuous- time dynamic model with fully connected state variables to analyze the bidirectional effects between physical and cognitive performance over time. Results: Higher physical performance was dynamically linked to improved memory recognition (-1.335, SD = 0.201, 95 BCI [-1.725, -0.954]). The effect was short-term, lasting up to five days (-0.368, SD = 0.05, 95 BCI [-0.479, -0.266]). Clinical scores supported the validity of the model and observed temporal dynamics. Conclusion: Higher physical performance predicted improved memory recognition in a day- by-day manner, providing a proof-of-concept for the feasibility of linking exercise training and cognition in patients with Alzheimer’s dementia. Highlights: Hierarchical Bayesian continuous-time dynamic modelling approach 72 repeated physical exercise (PP) and cognitive (COG) performance measurements PP is dynamically linked to session-to-session variability of COG Higher PP improved COG in subsequent sessions in subjects with Alzheimer’s dementia Short-term effect: lasting up to four days after training session Research in Context: Systematic review: Training-induced effects on cognitive outcomes in Alzheimer’s dementia and/or associated dynamic Bayesian modelling approaches were reviewed. Although studies showed exercise-induced cognitive improvements or maintenance, most of these studies fail to capture the dynamic nature of the change and interplay of physical (PP) and cognitive (COG) performance. Interpretation: Using a sophisticated hierarchical Bayesian continuous-time dynamic modelling approach, a fully connected state variable model was specified. PP is dynamically linked to COG, i.e. higher PP predicted improved COG in subsequent sessions. This effect was rather short term, lasting for up to five days. Future direction: Our results support exercise-induced effects on cognition. The cognitive system was still able to fluctuate and change favourably even in a sample with Alzheimer’s dementia. Further studies using dynamic modelling are necessary to replicate findings and examine other contributors to cognitive volatility in dementia.