Changes in memory processing across adulthood – Development rather than decline?
Our long-term memory abilities decline with age, but different memory processes are differentially affected. Broadly, processes that rely on gist-like representations (semantic memory, familiarity) appear relatively more spared than verbatim memory (episodic memory, recollection). This project will investigate the hypothesis that a shift towards more integrative and gist-based memory processing is in fact a beneficial adaptation of a memory system faced with the challenge of storing an ever increasing number of memory traces across the lifespan. While being a useful and efficient strategy for most of our everyday memory needs, it would also come with a cost and lead to decrements on certain memory tests. The hypothesis will be investigated by using cognitive testing and functional neuroimaging to study memory processes in midlife, a thus far understudied portion of the lifespan, when major age-related neuropathology has not yet occurred. The results will help differentiate between potentially adaptive changes in memory processing across adulthood, and those arising from true detrimental age-related changes – which, if not carefully considered, may manifest as similar performance decrements on certain memory tests. The results will also highlight possible improvements in our basic cognitive abilities across the lifespan, thereby helping society to capitalize on the competences of middle-aged and older individuals, and counteracting ageism.
Final report
Purpose & development
The project investigated the hypothesis that some apparent age-related memory impairments are in fact the result of beneficial adaptive processes, driven by the challenge of having to store an ever-increasing number of memory traces over the lifespan. This view has been referred to as the information accumulation perspective of cognitive aging (e.g., Qui & Johns, 2020). In parallel, the project aimed to try to identify beneficial changes in memory processes over the lifespan. Therefore, the project focused on individuals in early midlife (40-50 years), a period of life when major age-related brain changes have not usually yet occurred, but the number of memory traces is substantially greater than in younger adults (<30 years). two memory processes were studied: memory integration and mnemonic discrimination. memory integration is a process by which temporally separated but related or partially overlapping memory traces are thought to be linked though allocation to overlapping neural populations in the brain. research has shown that certain forms of memory integration can reduce interference in memory, thereby increasing recollection and reducing forgetting (e.g., anderson and mcculloch, 1999; chanales et al., 2019). integration has also been suggested to help reduce the trade-off between encoding and retrieval states in the brain, which otherwise leads to conflict every time current experiences overlap with previous memories (richter et al., 2016). this state of conflict can be assumed to happen more often with the accumulation of memory traces of life. here, it was hypothesized that increased use of memory integration could be an adaptive brain mechanism to deal with increased interference caused by an increased number of memory traces over life, which could be reflected by improved performance on memory tests intended to measure integration. memory discrimination, on the other hand, is a process that helps keep similar memory traces separate in the brain and allows for the differentiation of similar stimuli from each other in memory. this can be assumed to be more difficult as the number of memory traces grows over the span of life. previous research has shown that memory discrimination deteriorates significantly as early as middle age, and other studies have suggested that there is a trade-off between memory discrimination and integration, so that lower discrimination is associated with higher integration and vice versa (sweegers and talamini, 2014; van kesteren et al., 2016; tompary and davachi, 2017). thus, a pattern of lower memory discrimination and improved memory integration in healthy middle-aged adults could support the information accumulation perspective of cognitive aging.>
The project comprised two sub-studies, a behavioral study to compare younger and middle-aged individuals with respect to memory discrimination and memory integration ability, and a brain imaging study to investigate neural correlates of potential group differences. As described below, no age differences in the standard paradigms used to measure integration and discrimination could be detected in substudy 1. For substudy 2, a new paradigm was therefore developed that tested memory integration over semantically associated stimulus pairs, with the hope that this would enable middle-aged people to take advantage of their presumed more extensive semantic knowledge to perform the test, which also better mimics memory integration in everyday life.
Implementation
The data collection for substudy 1 was delayed due to the Covid-19 pandemic, but was performed in 2020 through a hybrid version of online data collection, where cognitive tests were performed by the participants online with a test leader present over video calls to ensure understanding of the instructions and correct execution. The approach was chosen to provide more control over data quality and participant characteristics (e.g., screening for health-related factors) than anonymous data collection methods via data collection platforms that are otherwise standard for online cognitive testing. A dataset comprising 41 younger participants (20-28 years) and 40 middle-aged participants (40-50 years), with assured data quality, was collected. The second substudy in the project plan was a brain imaging study with functional MRI to compare brain activity during semantically facilitated memory integration between healthy middle-aged individuals and younger adults. Due to replacement of the reseach MRI scanner, the second substudy was also delayed, but data for 72 participants could be collected during 2022-2023. Data processing of the brain imaging data is currently underway.
The project’s three most important results
(1) The results from substudy 1 indicated that healthy younger middle-aged individuals do not necessarily exhibit impaired mnemonic discrimination compared to younger adults, as previous research suggested. Bayes' factors (BFs) showed that the null hypothesis of no age difference was about 4 times as likely compared to the alternative hypothesis that an age difference exists.
(2) In addition, and contrary to our hypothesis, no improvement in memory integration ability could be demonstrated in middle-aged individuals compared to younger adults in substudy 1. Rather, the data were more consistent with age equivalence in this ability (BFs = 5.3 and 1.24 respectively in favor of the null hypothesis that no age difference exists, for two different measures of integration). It is worth noting that the absence of significant age differences in both memory discrimination and memory integration contrasted with significant group differences in the form of more extensive verbal knowledge and slower perceptual processing speed in middle-aged compared to younger participants. These expected cognitive group differences demonstrate that the current sample of individuals exhibit typical age effects in line with previous literature.
(3) The only significant difference in memory abilities between younger and middle-aged individuals was a lower performance on direct associative memory for unrelated stimulus pairs in middle-aged individuals. This finding is in line with established theories that associative memory is particularly age-sensitive, but the current findings suggest that this ability may be a sensitive marker for incipient age effects on memory already in healthy individuals in early midlife.
Taken together, the results of the project have shed light on memory functioning during an up until now relatively understudied part of the lifespan, and shown a pattern that suggests that certain memory processes (integration and discrimination) seem to be relatively more preserved in healthy middle-aged individuals than others (associative memory for unrelated stimuli). However, no evidence was obtained for the hypothesis of improvements in memory abilities in middle-aged people compared to younger adults, nor the information accumulation perspective of cognitive aging. Upcoming analyses of brain imaging data may highlight possible age differences in neural correlates for the use of previous semantic knowledge to support memory integration, where middle-aged adults are hypothesized to be able to make use of their greater semantic world knowledge to support this memory process.
Potential new research questions
Future research could investigate how previously accumulated knowledge over the life span affects memory processes within the linguistic domain, as numerous previous studies have documented more extensive verbal knowledge in middle-aged and older individuals compared to younger individuals. Linguistic parameters of word characteristics can be formally quantified from large-scale text databases as proxy measures of linguistic knowledge in the population, and used as independent variables in verbal paired-associates memory tests. Interference effects of such linguistic parameters on associative memory for word pairs have been shown in older compared to younger individuals (Qiu & Johns, 2020), but the neural correlates of such effects are so far unknown, nor is it known if there are circumstances in which facilitative effects may occur in individuals with more extensive linguistic knowledge.
Dissemination
At the time of writing, one peer-reviewed article with open access has been published from the project (Samrani et al., 2022). Due to the Covid-19 pandemic, the results have mainly been presented at local meetings and seminars, and no external collaboration has taken place. When the results from substudy 2 have been completed, the aim is to present these at national and/or international conferences, as well as publish them in peer-reviewed international journals.
References
Anderson, M. C., and McCulloch, K. C. (1999). Integration as a general boundary condition on retrieval-induced forgetting. J. Exp. Psychol. Learn. Mem. Cogn. 25, 608–629.
Chanales, A. J. H., et al. (2019). Interference between overlapping memories is predicted by neural states during learning. Nat. Commun. 10:5363.
Qiu, M., and Johns, B. T. (2020). Semantic diversity in paired-associate learning: further evidence for the information accumulation perspective of cognitive aging. Psychon. Bull. Rev. 27, 114–121.
Richter, F. R., Chanales, A. J. H., and Kuhl, B. A. (2016). Predicting the integration of overlapping memories by decoding mnemonic processing states during learning. NeuroImage 124, 323–335.
Samrani, G., Lundquist, A., & Pudas, S. (2022). Healthy middle-aged adults have preserved mnemonic discrimination and integration, while showing no detectable memory benefits. Front. Psychol., 12, Article 797387.
Sweegers, C. C. G., and Talamini, L. M. (2014). Generalization from episodic memories across time: a route for semantic knowledge acquisition. Cortex 59, 49–61.
Tompary, A., and Davachi, L. (2017). Consolidation promotes the emergence of representational overlap in the hippocampus and medial prefrontal cortex article consolidation promotes the emergence of representational overlap in the hippocampus and medial prefrontal cortex. Neuron 96, 228–241.e5
van Kesteren, M. T. R., Brown, T. I., and Wagner, A. D. (2016). Interactions between memory and new learning: insights from fMRI multivoxel pattern analysis. Front. Syst. Neurosci. 10:46.
The project investigated the hypothesis that some apparent age-related memory impairments are in fact the result of beneficial adaptive processes, driven by the challenge of having to store an ever-increasing number of memory traces over the lifespan. This view has been referred to as the information accumulation perspective of cognitive aging (e.g., Qui & Johns, 2020). In parallel, the project aimed to try to identify beneficial changes in memory processes over the lifespan. Therefore, the project focused on individuals in early midlife (40-50 years), a period of life when major age-related brain changes have not usually yet occurred, but the number of memory traces is substantially greater than in younger adults (<30 years). two memory processes were studied: memory integration and mnemonic discrimination. memory integration is a process by which temporally separated but related or partially overlapping memory traces are thought to be linked though allocation to overlapping neural populations in the brain. research has shown that certain forms of memory integration can reduce interference in memory, thereby increasing recollection and reducing forgetting (e.g., anderson and mcculloch, 1999; chanales et al., 2019). integration has also been suggested to help reduce the trade-off between encoding and retrieval states in the brain, which otherwise leads to conflict every time current experiences overlap with previous memories (richter et al., 2016). this state of conflict can be assumed to happen more often with the accumulation of memory traces of life. here, it was hypothesized that increased use of memory integration could be an adaptive brain mechanism to deal with increased interference caused by an increased number of memory traces over life, which could be reflected by improved performance on memory tests intended to measure integration. memory discrimination, on the other hand, is a process that helps keep similar memory traces separate in the brain and allows for the differentiation of similar stimuli from each other in memory. this can be assumed to be more difficult as the number of memory traces grows over the span of life. previous research has shown that memory discrimination deteriorates significantly as early as middle age, and other studies have suggested that there is a trade-off between memory discrimination and integration, so that lower discrimination is associated with higher integration and vice versa (sweegers and talamini, 2014; van kesteren et al., 2016; tompary and davachi, 2017). thus, a pattern of lower memory discrimination and improved memory integration in healthy middle-aged adults could support the information accumulation perspective of cognitive aging.>
The project comprised two sub-studies, a behavioral study to compare younger and middle-aged individuals with respect to memory discrimination and memory integration ability, and a brain imaging study to investigate neural correlates of potential group differences. As described below, no age differences in the standard paradigms used to measure integration and discrimination could be detected in substudy 1. For substudy 2, a new paradigm was therefore developed that tested memory integration over semantically associated stimulus pairs, with the hope that this would enable middle-aged people to take advantage of their presumed more extensive semantic knowledge to perform the test, which also better mimics memory integration in everyday life.
Implementation
The data collection for substudy 1 was delayed due to the Covid-19 pandemic, but was performed in 2020 through a hybrid version of online data collection, where cognitive tests were performed by the participants online with a test leader present over video calls to ensure understanding of the instructions and correct execution. The approach was chosen to provide more control over data quality and participant characteristics (e.g., screening for health-related factors) than anonymous data collection methods via data collection platforms that are otherwise standard for online cognitive testing. A dataset comprising 41 younger participants (20-28 years) and 40 middle-aged participants (40-50 years), with assured data quality, was collected. The second substudy in the project plan was a brain imaging study with functional MRI to compare brain activity during semantically facilitated memory integration between healthy middle-aged individuals and younger adults. Due to replacement of the reseach MRI scanner, the second substudy was also delayed, but data for 72 participants could be collected during 2022-2023. Data processing of the brain imaging data is currently underway.
The project’s three most important results
(1) The results from substudy 1 indicated that healthy younger middle-aged individuals do not necessarily exhibit impaired mnemonic discrimination compared to younger adults, as previous research suggested. Bayes' factors (BFs) showed that the null hypothesis of no age difference was about 4 times as likely compared to the alternative hypothesis that an age difference exists.
(2) In addition, and contrary to our hypothesis, no improvement in memory integration ability could be demonstrated in middle-aged individuals compared to younger adults in substudy 1. Rather, the data were more consistent with age equivalence in this ability (BFs = 5.3 and 1.24 respectively in favor of the null hypothesis that no age difference exists, for two different measures of integration). It is worth noting that the absence of significant age differences in both memory discrimination and memory integration contrasted with significant group differences in the form of more extensive verbal knowledge and slower perceptual processing speed in middle-aged compared to younger participants. These expected cognitive group differences demonstrate that the current sample of individuals exhibit typical age effects in line with previous literature.
(3) The only significant difference in memory abilities between younger and middle-aged individuals was a lower performance on direct associative memory for unrelated stimulus pairs in middle-aged individuals. This finding is in line with established theories that associative memory is particularly age-sensitive, but the current findings suggest that this ability may be a sensitive marker for incipient age effects on memory already in healthy individuals in early midlife.
Taken together, the results of the project have shed light on memory functioning during an up until now relatively understudied part of the lifespan, and shown a pattern that suggests that certain memory processes (integration and discrimination) seem to be relatively more preserved in healthy middle-aged individuals than others (associative memory for unrelated stimuli). However, no evidence was obtained for the hypothesis of improvements in memory abilities in middle-aged people compared to younger adults, nor the information accumulation perspective of cognitive aging. Upcoming analyses of brain imaging data may highlight possible age differences in neural correlates for the use of previous semantic knowledge to support memory integration, where middle-aged adults are hypothesized to be able to make use of their greater semantic world knowledge to support this memory process.
Potential new research questions
Future research could investigate how previously accumulated knowledge over the life span affects memory processes within the linguistic domain, as numerous previous studies have documented more extensive verbal knowledge in middle-aged and older individuals compared to younger individuals. Linguistic parameters of word characteristics can be formally quantified from large-scale text databases as proxy measures of linguistic knowledge in the population, and used as independent variables in verbal paired-associates memory tests. Interference effects of such linguistic parameters on associative memory for word pairs have been shown in older compared to younger individuals (Qiu & Johns, 2020), but the neural correlates of such effects are so far unknown, nor is it known if there are circumstances in which facilitative effects may occur in individuals with more extensive linguistic knowledge.
Dissemination
At the time of writing, one peer-reviewed article with open access has been published from the project (Samrani et al., 2022). Due to the Covid-19 pandemic, the results have mainly been presented at local meetings and seminars, and no external collaboration has taken place. When the results from substudy 2 have been completed, the aim is to present these at national and/or international conferences, as well as publish them in peer-reviewed international journals.
References
Anderson, M. C., and McCulloch, K. C. (1999). Integration as a general boundary condition on retrieval-induced forgetting. J. Exp. Psychol. Learn. Mem. Cogn. 25, 608–629.
Chanales, A. J. H., et al. (2019). Interference between overlapping memories is predicted by neural states during learning. Nat. Commun. 10:5363.
Qiu, M., and Johns, B. T. (2020). Semantic diversity in paired-associate learning: further evidence for the information accumulation perspective of cognitive aging. Psychon. Bull. Rev. 27, 114–121.
Richter, F. R., Chanales, A. J. H., and Kuhl, B. A. (2016). Predicting the integration of overlapping memories by decoding mnemonic processing states during learning. NeuroImage 124, 323–335.
Samrani, G., Lundquist, A., & Pudas, S. (2022). Healthy middle-aged adults have preserved mnemonic discrimination and integration, while showing no detectable memory benefits. Front. Psychol., 12, Article 797387.
Sweegers, C. C. G., and Talamini, L. M. (2014). Generalization from episodic memories across time: a route for semantic knowledge acquisition. Cortex 59, 49–61.
Tompary, A., and Davachi, L. (2017). Consolidation promotes the emergence of representational overlap in the hippocampus and medial prefrontal cortex article consolidation promotes the emergence of representational overlap in the hippocampus and medial prefrontal cortex. Neuron 96, 228–241.e5
van Kesteren, M. T. R., Brown, T. I., and Wagner, A. D. (2016). Interactions between memory and new learning: insights from fMRI multivoxel pattern analysis. Front. Syst. Neurosci. 10:46.