Group-specific repetition suppression (A, B) and repetition enhancement (C, D) effects. A. Positive posterior suppression effect (RS1). B. Negative frontocentral suppression effect (RS2). C. Negative frontal enhancement effect (RE1). D. Positive centro-parietal enhancement effect (RE2). ERPs are averaged over all trials in which objects were shown for the first time (solid line) or for the second time (dashed line) for children (left), young adults (middle), and older adults (right). The x-axis shows trial time (s) with stimulus onset at 0 (origin) and offset at 1.5 s, the y-axis shows amplitude (μV) with negative values plotted downwards. The time windows in which reliable differences between first and second presentation were identified (cluster-based permutation analysis) are shaded in gray. All ERPs are averaged over the respective electrodes in which the effects were identified, highlighted by asterisks in the respective topographical distributions plotted next to the ERPs. Topographies show the resulting t-values from contrasting ERPs of first and second presentations, averaged over the respective significant time windows. The pvalues from the cluster-based permutation analysis are provided for each time-electrode cluster.

Group-specific repetition suppression (A, B) and repetition enhancement (C, D) effects. A. Positive posterior suppression effect (RS1). B. Negative frontocentral suppression effect (RS2). C. Negative frontal enhancement effect (RE1). D. Positive centro-parietal enhancement effect (RE2). ERPs are averaged over all trials in which objects were shown for the first time (solid line) or for the second time (dashed line) for children (left), young adults (middle), and older adults (right). The x-axis shows trial time (s) with stimulus onset at 0 (origin) and offset at 1.5 s, the y-axis shows amplitude (μV) with negative values plotted downwards. The time windows in which reliable differences between first and second presentation were identified (cluster-based permutation analysis) are shaded in gray. All ERPs are averaged over the respective electrodes in which the effects were identified, highlighted by asterisks in the respective topographical distributions plotted next to the ERPs. Topographies show the resulting t-values from contrasting ERPs of first and second presentations, averaged over the respective significant time windows. The pvalues from the cluster-based permutation analysis are provided for each time-electrode cluster.

Source publication
Article
Full-text available
The specificity with which past experiences can be remembered varies across the lifespan, possibly due to differences in how precisely information is encoded. Memory formation can be investigated through repetition effects, the common finding that neural activity is altered when stimuli are repeated. However, whether differences in this indirect me...

Contexts in source publication

Context 1
... this procedure, we identified two clusters in children, three in younger adults, and four in older adults (all clusters p ≤ 0.002) with broadly overlapping topography and latency (see Fig. 5 for details). For all age groups, there was a cluster over posterior electrodes showing lower positive amplitudes for the second versus first stimulus presentation (i.e., repetition suppression; RS1; Fig. 5A). This cluster appeared earliest in children (380-868 ms after stimulus onset), slightly later and with shorter duration in young ...
Context 2
... in children, three in younger adults, and four in older adults (all clusters p ≤ 0.002) with broadly overlapping topography and latency (see Fig. 5 for details). For all age groups, there was a cluster over posterior electrodes showing lower positive amplitudes for the second versus first stimulus presentation (i.e., repetition suppression; RS1; Fig. 5A). This cluster appeared earliest in children (380-868 ms after stimulus onset), slightly later and with shorter duration in young adults (492-840 ms), and latest and with the shortest duration in older adults (568-884 ms). Within similar time windows another repetition suppression cluster (RS2) was identified at frontal and central ...
Context 3
... ms after stimulus onset), slightly later and with shorter duration in young adults (492-840 ms), and latest and with the shortest duration in older adults (568-884 ms). Within similar time windows another repetition suppression cluster (RS2) was identified at frontal and central electrodes which showed reduced negativity for repeated stimuli (Fig. 5B). This cluster also appeared earliest in children (192-864 ms), followed by young adults (224-756 ms), and latest, as well as with much shorter duration in older adults (676-848 ms). For all age groups, both of these suppression effects occurred mainly while the ERP deflections came back to baseline, starting right at or after the peak ...
Context 4
... we identified a repetition enhancement effect (RE1) for both young and older adults that showed stronger (more negative) activity for the second than for the first object presentations over mainly frontal and temporal electrode sites (young adults: 244-368 ms; older adults: 308-548 ms; Fig. 5C). An opposite enhancement effect (RE2) was only identified for older adults over centro-parietal regions at 308-640 ms after stimulus onset (Fig. 5D). In analogy to the two suppression effects, the opposite enhancement clusters identified in older adults could be two reflections of the same effect, which would be indicated by a high ...
Context 5
... showed stronger (more negative) activity for the second than for the first object presentations over mainly frontal and temporal electrode sites (young adults: 244-368 ms; older adults: 308-548 ms; Fig. 5C). An opposite enhancement effect (RE2) was only identified for older adults over centro-parietal regions at 308-640 ms after stimulus onset (Fig. 5D). In analogy to the two suppression effects, the opposite enhancement clusters identified in older adults could be two reflections of the same effect, which would be indicated by a high correlation between the effect ...
Context 6
... effects were identified for all age groups (see Fig. 5). In a next step, we examined whether the size of the effects differed between age groups (see Fig. 6 left). Because of overall age differences in EEG amplitudes, rather than computing and comparing the raw amplitude difference between the ERPs to first and second presentations, it is more appropriate to compare the effect sizes, i.e., ...

Citations

... The sample dataset includes EEG data from ten 7-9-year-old children (6 female, 4 male) and ten 18-30-year-old adults (5 female, 5 male) during the encoding phase of an episodic memory study ( Fig. 1; for a detailed description of the task and data preprocessing, see Sommer et al., 2021). The data made available stem from a subsample of the original participants and selected conditions of the original experiment that allow the interested reader to run this tutorial. ...
... In the recognition task, exact item repetitions, similar lures (new exemplars from the same object categories), and entirely new objects were presented (cf. Stark et al., 2019), allowing for an estimation of precise item memory (for details, see Sommer et al., 2021). Here, as an example, we correlate item memory performance with neural item specificity (computed above) to identify between-person associations between brain and behavior. ...
Article
Full-text available
The human brain encodes information in neural activation patterns. While standard approaches to analyzing neural data focus on brain (de-)activation (e.g., regarding the location, timing, or magnitude of neural responses), multivariate neural pattern similarity analyses target the informational content represented by neural activity. In adults, a number of representational properties have been identified that are linked to cognitive performance, in particular the stability, distinctiveness, and specificity of neural patterns. However, although growing cognitive abilities across childhood suggest advancements in representational quality, developmental studies still rarely utilize information-based pattern similarity approaches, especially in electroencephalography (EEG) research. Here, we provide a comprehensive methodological introduction and step-by-step tutorial for pattern similarity analysis of spectral (frequency-resolved) EEG data including a publicly available pipeline and sample dataset with data from children and adults. We discuss computation of single-subject pattern similarities and their statistical comparison at the within-person to the between-group level as well as the illustration and interpretation of the results. This tutorial targets both novice and more experienced EEG researchers and aims to facilitate the usage of spectral pattern similarity analyses, making these methodologies more readily accessible for (developmental) cognitive neuroscientists.
... Researchers have shown that, even though there exist slight variations, humans show a level of consistency when remembering the same kind of images with a very similar probability irrespective of the time delay (Sommer et al., 2021). This research has led to the inference that it is possible to measure an individual's probability of remembering an image. ...
... So, the approximate memorability scores will be obtained for the same image from multiple people. Image memorability is a reflection of individual viewing the image, but however, the level of memorability of an image is quite similar across individuals most of the time (Sommer et al., 2021). So, since the memorability of the image is only going to slightly vary for most people, these approximate measures are taken as the ground truth memorability score. ...
... Factors like color harmony and object interestingness are generally agreed upon by people as factors that improve image memorability Khosla et al., 2015. Few methods have been proposed (Perera, Tal & Zelnik-Manor, 2019;Fajtl et al., 2018;Squalli-Houssaini et al., 2018) to predict the memorability of an image using deep learning methods. Those methods either used handcrafted features or ensemble models to predict the memorability score. ...
Article
Full-text available
Image memorability is a very hard problem in image processing due to its subjective nature. But due to the introduction of Deep Learning and the large availability of data and GPUs, great strides have been made in predicting the memorability of an image. In this paper, we propose a novel deep learning architecture called ResMem-Net that is a hybrid of LSTM and CNN that uses information from the hidden layers of the CNN to compute the memorability score of an image. The intermediate layers are important for predicting the output because they contain information about the intrinsic properties of the image. The proposed architecture automatically learns visual emotions and saliency, shown by the heatmaps generated using the GradRAM technique. We have also used the heatmaps and results to analyze and answer one of the most important questions in image memorability: ''What makes an image memorable?''. The model is trained and evaluated using the publicly available Large-scale Image Memorability dataset (LaMem) from MIT. The results show that the model achieves a rank correlation of 0.679 and a mean squared error of 0.011, which is better than the current state-of-the-art models and is close to human consistency (p = 0.68). The proposed architecture also has a significantly low number of parameters compared to the state-of-the-art architecture, making it memory efficient and suitable for production.
... Initially, such an approach was used in MRI research with adults where it is known as cluster mass test (Bullmore et al., 1999). This by now well-established method for fMRI, MEG, and EEG analysis of adult data (e.g., Pernet et al., 2015) is still used seldom for developmental EEG (but see e.g., Meyer et al., 2020;Meyer and Hunnius, 2021;Sommer et al., 2021). The cluster-based permutation test not only corrects for the multiple comparison problem and thereby reduces false positive results, it also reduces the potential for false negative effects. ...
Article
Full-text available
Developmental research using electroencephalography (EEG) offers valuable insights in brain processes early in life, but at the same time, applying this sensitive technique to young children who are often non-compliant and have short attention spans comes with practical limitations. It is thus of particular importance to optimally use the limited resources to advance our understanding of development through reproducible and replicable research practices. Here, we describe methodological approaches that help maximize the reproducibility of developmental EEG research. We discuss how to transform EEG data into the standardized Brain Imaging Data Structure (BIDS) which organizes data according to the FAIR data sharing principles. We provide a tutorial on how to use cluster-based permutation testing to analyze developmental EEG data. This versatile test statistic solves the multiple comparison problem omnipresent in EEG analysis and thereby substantially decreases the risk of reporting false discoveries. Finally, we describe how to quantify effect sizes, in particular of cluster-based permutation results. Reporting effect sizes conveys a finding’s impact and robustness which in turn informs future research. To demonstrate these methodological approaches to data organization, analysis and report, we use a publicly accessible infant EEG dataset and provide a complete copy of the analysis code.
... The sample dataset includes EEG data from ten 7-9-year-old children (6 female, 4 male) and ten 18-30-year-old adults (5 female, 5 male) during the encoding phase of an episodic memory study ( Figure 1; for a detailed description of the task and data preprocessing, see Sommer et al., 2021). The data made available stem from a subsample of the original participants and selected conditions of the original experiment that allow the interested reader to run this tutorial. ...
... In the recognition task, exact item repetitions, similar lures (new exemplars from the same object categories), and entirely new objects were presented (cf. Stark et al., 2019), allowing for an estimation of precise item memory (for details, see Sommer et al., 2021). Here, as an example, we correlate item memory performance with neural item specificity (computed above) to identify between-person associations between brain and behavior. ...
Preprint
Full-text available
The human brain encodes information in neural activation patterns. While standard approaches to analyzing neural data focus on brain (de-)activation (e.g., regarding the location, timing, or magnitude of neural responses), multivariate neural pattern similarity analyses target the informational content represented by neural activity. In adults, a number of representational properties have been identified that are linked to cognitive performance, in particular the stability, distinctiveness, and specificity of neural patterns. However, although growing cognitive abilities across childhood suggest advancements in representational quality, developmental studies still rarely utilize information-based pattern similarity approaches, especially in electroencephalography (EEG) research. Here, we provide a comprehensive methodological introduction and step-by-step tutorial for pattern similarity analysis of spectral (frequency-resolved) EEG data including a publicly available pipeline and sample dataset with data from children and adults. We discuss computation of single-subject pattern similarities and their statistical comparison at the within-person to the between-group level as well as the illustration and interpretation of the results. This tutorial targets both novice and more experienced EEG researchers and aims to facilitate the usage of spectral pattern similarity analyses, making these methodologies more readily accessible for (developmental) cognitive neuroscientists.
... Although the current study clearly identifies a relationship between representational specificity and stability during encoding with later memory performance, the older adults group, who showed overall reduced specificity and stability, did not perform overall significantly worse than the young adult group, similar to findings by . Less pronounced or no age-related deficits in item recognition memory, compared with, for example, associative memory, are not uncommon, especially under incidental encoding conditions (compare Old and Naveh-Benjamin, 2008; Sommer et al., 2021). Nevertheless, older adults often tend to respond "old" more frequently than younger adults, contributing to their higher rates of false memories Schacter et al., 1997;Fandakova et al., 2013Fandakova et al., , 2020, which was also observed in the present study. ...
Article
The distinctiveness of neural information representation is crucial for successful memory performance but declines with advancing age. Computational models implicate age-related neural dedifferentiation on the level of item representations, but previous studies mostly focused on age differences of categorical information representation in higher-order visual regions. In an age-comparative fMRI study, we combined univariate analyses and whole-brain searchlight pattern similarity analyses to elucidate age differences in neural distinctiveness at both category and item levels and their relation to memory. Thirty-five younger (18-27 years old) and 32 older (67-75 years old) women and men incidentally encoded images of faces and houses, followed by an old/new recognition memory task. During encoding, age-related neural dedifferentiation was shown as reduced category-selective processing in ventral visual cortex and impoverished item specificity in occipital regions. Importantly, successful subsequent memory performance built upon high item stability, that is, high representational similarity between initial and repeated presentation of an item, which was greater in younger than older adults. Overall, we found that differences in representational distinctiveness coexist across representational levels and contribute to interindividual and intraindividual variability in memory success, with item specificity being the strongest contributor. Our results close an important gap in the literature, showing that older adults' neural representation of item-specific information in addition to categorical information is reduced compared to younger adults.SIGNIFICANCE STATEMENTA long-standing hypothesis links age-related cognitive decline to a loss of neural specificity. While previous evidence supports the notion of age-related neural dedifferentiation of category-level information in ventral visual cortex, whether or not age differences exist at the item level was a matter of debate. Here, we observed age group differences at both levels as well as associations between both categorical distinctiveness and item specificity to memory performance, with item specificity being the strongest contributor. Importantly, age differences in occipital item specificity were largely due to reduced item stability across repetitions in older adults. Our results suggest that age differences in neural representations can be observed across the entire cortical hierarchy and are not limited to category-level information.
Article
Full-text available
Although the human brain is adapted to function within three-dimensional environments, conventional laboratory research commonly investigates cognitive mechanisms in a reductionist approach using two-dimensional stimuli. However, findings regarding mnemonic processes indicate that realistic experiences in Virtual Reality (VR) are stored in richer and more intertwined engrams than those obtained from the conventional laboratory. Our study aimed to further investigate the generalizability of laboratory findings and to differentiate whether the processes underlying memory formation differ between VR and the conventional laboratory already in early encoding stages. Therefore, we investigated the Repetition Suppression (RS) effect as a correlate of the earliest instance of mnemonic processes under conventional laboratory conditions and in a realistic virtual environment. Analyses of event-related potentials (ERPs) indicate that the ERP deflections at several electrode clusters were lower in VR compared to the PC condition. These results indicate an optimized distribution of cognitive resources in realistic contexts. The typical RS effect was replicated under both conditions at most electrode clusters for a late time window. Additionally, a specific RS effect was found in VR at anterior electrodes for a later time window, indicating more extensive encoding processes in VR compared to the laboratory. Specifically, electrotomographic results (VARETA) indicate multimodal integration involving a broad cortical network and higher cognitive processes during the encoding of realistic objects. Our data suggest that object perception under realistic conditions, in contrast to the conventional laboratory, requires multisensory integration involving an interconnected functional system, facilitating the formation of intertwined memory traces in realistic environments.
Article
Long-standing theories of cognitive aging suggest that memory decline is associated with age-related differences in the way information is neurally represented. Multivariate pattern similarity analyses enabled researchers to take a representational perspective on brain and cognition, and allowed them to study the properties of neural representations that support successful episodic memory. Two representational properties have been identified as crucial for memory performance, namely the distinctiveness and the stability of neural representations. Here, we review studies that used multivariate analysis tools for different neuroimaging techniques to clarify how these representational properties relate to memory performance across adulthood. While most evidence on age differences in neural representations involved stimulus category information , recent studies demonstrated that particularly item-level stability and specificity of activity patterns are linked to memory success and decline during aging. Overall, multivariate methods offer a versatile tool for our understanding of age differences in the neural representations underlying memory.
Article
This article has been retracted: please see Elsevier Policy on Article Withdrawal (https://www.elsevier.com/about/our-business/policies/article-withdrawal). This article has been retracted at the request of the authors. During follow-up analyses for a different manuscript, the authors recognized a serious error in their analysis pipeline that raises doubts about the results presented in this paper. Most likely, the results presented in the paper do not hold after fixing the error. The authors need some more time to fully understand the consequences of this error. Therefore, the paper is retracted at the request of authors. The authors regret this and apologize to readers of Neurobiology of Aging.
Article
Episodic memory decline is a hallmark of cognitive aging and a multifaceted phenomenon. We review studies that target age differences across different memory processing stages, i.e., from encoding to retrieval. The available evidence suggests that age differences during memory formation may affect the quality of memory representations in an age-graded manner with downstream consequences for later processing stages. We argue that low memory quality in combination with age-related neural decline of key regions of the episodic memory network puts older adults in a double jeopardy situation that finally results in broader memory impairments in older compared to younger adults.