Item Memory Research Articles

Traces of semantization - from episodic to semantic memory in a spiking cortical network model.

Episodic memory is a recollection of past personal experiences associated with particular times and places. This kind of memory is commonly subject to loss of contextual information or “semantization,” which gradually decouples the encoded memory items from their associated contexts while transforming them into semantic or gist-like representations. Novel extensions to the classical Remember/Know (R/K) behavioral paradigm attribute the loss of episodicity to multiple exposures of an item in different contexts. Despite recent advancements explaining semantization at a behavioral level, the underlying neural mechanisms remain poorly understood. In this study, we suggest and evaluate a novel hypothesis proposing that Bayesian–Hebbian synaptic plasticity mechanisms might cause semantization of episodic memory. We implement a cortical spiking neural network model with a Bayesian–Hebbian learning rule called Bayesian Confidence Propagation Neural Network (BCPNN), which captures the semantization phenomenon and offers a mechanistic explanation for it. Encoding items across multiple contexts leads to item-context decoupling akin to semantization. We compare BCPNN plasticity with the more commonly used spike-timing-dependent plasticity (STDP) learning rule in the same episodic memory task. Unlike BCPNN, STDP does not explain the decontextualization process. We further examine how selective plasticity modulation of isolated salient events may enhance preferential retention and resistance to semantization. Our model reproduces important features of episodicity on behavioral timescales under various biological constraints while also offering a novel neural and synaptic explanation for semantization, thereby casting new light on the interplay between episodic and semantic memory processes.

Model‐Based Analysis of Standard Operating Procedures' Role in Abnormal and Emergency Events

AbstractStandard Operating Procedures (SOP) describe the desired flight‐crew response to mission events including normal, abnormal, and emergency events. The SOPs define flight‐crew step‐by‐step actions to perform the SOP appropriate for the mission event. Although the aircraft and automation are tested rigorously for certification, and the flight‐crew are trained and qualified through simulator and flight check‐rides, oversight of the SOPs is marginal. This is especially problematic for SOPs that must be completed in a short Allowable Operational Time Window (AOTW) prior to the occurrence of a hazardous event. The AOTW can be variable due to environmental conditions (e.g. wind, aircraft performance), as is the Time‐on‐Procedure (ToP) (e.g. simultaneous procedures/events, disruptions/distractions in the cockpit, variability in individual human performance). The variance motivates the need for modeling and analysis of SOPs.This paper used a modeling language embedded in a System Engineering Model‐Based tool to evaluate the execution performance of the SOPs for abnormal/emergency mission events with regard to ToP and AOTW. The analysis yielded a classification of the accident scenarios by: (1) Insufficient AOTW, (2) Excessive ToP, or (3) combination of excessive ToP and insufficient AOTW. For the excessive ToP, the analysis highlighted the issue of delay in initiating an SOP due to ambiguity in the triggering cue from the external environment, or missing associations between a cockpit alert and the SOP. The analysis also highlighted issues arising from the presence of memory items in the SOP, or cues that are out of the operator's Field of View (FoV). In all cases, the results highlighted the importance of a method for quantifying the performance of SOPs considering the variability in ToP and AOTW. The implications of these results are discussed.

Altered alpha/beta desynchronization during item-context binding contributes to the associative deficit in older age.

It is proposed that older adults have difficulties to bind item and context and to recruit deep, elaborative processing during encoding. Senescent changes in the oscillatory foundations of these processes are currently unclear. We recorded electroencephalography during item-context memory formation in younger (n= 57) and older (n= 55) adults. At test, we assessed memory for the items and the item-context pairs and examined encoding-related activity based on how much information was recovered at retrieval (miss < item-only < pair). Item memory was comparable between age groups while pair memory was reduced in the older adults. Theta synchronization and alpha/beta desynchronization increased linearly with the amount of information available. Single-trial theta power could not predict subsequent item memory, but predicted pair memory in an age-invariant manner, in line with a mechanism supporting associative memory. In contrast, single-trial alpha/beta power predicted both item and pair memory, in line with a mechanism reflecting the depth of information processing, and predicted pair memory less well in the older than the younger adults. Thus, theta and alpha/beta oscillations contribute differently in shaping the contents of memories and reduced processing capacity contributes to episodic memory decline in older age.

Phonological Working Memory Representations in the Left Inferior Parietal Lobe in the Face of Distraction and Neural Stimulation.

The neural basis of phonological working memory (WM) was investigated through an examination of the effects of irrelevant speech distractors and disruptive neural stimulation from transcranial magnetic stimulation (TMS). Embedded processes models argue that the same regions involved in speech perception are used to support phonological WM whereas buffer models assume that a region separate from speech perception regions is used to support WM. Thus, according to the embedded processes approach but not the buffer approach, irrelevant speech and TMS to the speech perception region should disrupt the decoding of phonological WM representations. According to the buffer account, decoding of WM items should be possible in the buffer region despite distraction and should be disrupted with TMS to this region. Experiment 1 used fMRI and representational similarity analyses (RSA) with a delayed recognition memory paradigm using nonword stimuli. Results showed that decoding of memory items in the speech perception regions (superior temporal gyrus, STG) was possible in the absence of distractors. However, the decoding evidence in the left STG was susceptible to interference from distractors presented during the delay period whereas decoding in the proposed buffer region (supramarginal gyrus, SMG) persisted. Experiment 2 examined the causal roles of the speech processing region and the buffer region in phonological WM performance using TMS. TMS to the SMG during the early delay period caused a disruption in recognition performance for the memory nonwords, whereas stimulations at the STG and an occipital control region did not affect WM performance. Taken together, results from the two experiments are consistent with predictions of a buffer model of phonological WM, pointing to a critical role of the left SMG in maintaining phonological representations.

Hyperspectral anomaly detection via memory‐augmented autoencoders

AbstractRecently, the autoencoder (AE) based method plays a critical role in the hyperspectral anomaly detection domain. However, due to the strong generalised capacity of AE, the abnormal samples are usually reconstructed well along with the normal background samples. Thus, in order to separate anomalies from the background by calculating reconstruction errors, it can be greatly beneficial to reduce the AE capability for abnormal sample reconstruction while maintaining the background reconstruction performance. A memory‐augmented autoencoder for hyperspectral anomaly detection (MAENet) is proposed to address this challenging problem. Specifically, the proposed MAENet mainly consists of an encoder, a memory module, and a decoder. First, the encoder transforms the original hyperspectral data into the low‐dimensional latent representation. Then, the latent representation is utilised to retrieve the most relevant matrix items in the memory matrix, and the retrieved matrix items will be used to replace the latent representation from the encoder. Finally, the decoder is used to reconstruct the input hyperspectral data using the retrieved memory items. With this strategy, the background can still be reconstructed well while the abnormal samples cannot. Experiments conducted on five real hyperspectral anomaly data sets demonstrate the superiority of the proposed method.

Pupillometry and electroencephalography in the digit span task

This dataset consists of raw 64-channel EEG, cardiovascular (electrocardiography and photoplethysmography), and pupillometry data from 86 human participants recorded during 4 minutes of eyes-closed resting and during performance of a classic working memory task – digit span task with serial recall. The participants either memorized or just listened to sequences of 5, 9, or 13 digits presented auditorily every 2 seconds. The dataset can be used for (1) developing algorithms for cognitive load discrimination and detection of cognitive overload; (2) studying neural (event-related potentials and brain oscillations) and peripheral (electrocardiography, photoplethysmography, and pupillometry) physiological signals during encoding and maintenance of each sequentially presented memory item; (3) correlating cognitive load and individual differences in working memory to neural and peripheral physiology, and studying the relationship between the physiological signals; (4) integration of the physiological findings with the vast knowledge coming from behavioral studies of verbal working memory in simple span paradigms. The data are shared in Brain Imaging Data Structure (BIDS) format and freely available on OpenNeuro (https://openneuro.org/datasets/ds003838).

Abstract 4158: High dimensional analysis of sipuleucel T from metastatic castration resistant prostate cancer patients using mass cytometry

Abstract Background: Prostate cancer is the most commonly diagnosed non-skin cancer in men, and metastatic castration-resistant prostate cancer (mCRPC) represents the most lethal form of the disease. Sipuleucel T (sip T) is an autologous active cellular immunotherapy approved by FDA for patients with mCRPC and has been shown effective in improving overall survival (OS) among individuals with mCRPC. To date, detailed analysis of the sip T product has not been studied using an advanced mass cytometry approach. Here, we present high dimensional data describing the specific leukocyte subsets and phenotype of sip-T in detail. Methods: We performed a comprehensive assessment of sip T (n=14 samples, collected under an IRB approved protocol) from 11 different patients with mCRPC, using high-throughput mass cytometry (CyTOF) comprised of 37 different antibodies/markers. Sip T samples were then analyzed for various leukocyte subsets, as well as markers of activation and exhaustion. Furthermore, clinical correlations were drawn between immune cell subsets/markers and patient clinical data such, as prostate specific antigen (PSA) levels. Results: CyTOF analysis revealed that CD3+ T cells (including CD4+ and CD8+) constituted the highest proportion (median, range: 63%, 9-89%) of sip T, followed by CD19+ B cells (4%, 1-82%), CD3-CD14-CD56+ natural killer (NK) cells (4%, 1-18%), and CD3-CD19-CD56-HLA-DR+CD11c+CD14+ monocytes (1%, 1-37%) besides a small proportion (&amp;lt;1%) of conventional dendritic cells (cDC), innate lymphoid cells (ILC), plasmacytoid DC (pDC), myeloid-derived suppressor cells (MDSC) and NK T-cells (NKT). The majority of CD8+ T cells were either CD45RA+CCR7-effector (Teff: 43%, 9-70%), CD45RA-CCR7- effector memory (Tem: 23%, 11-59%) and CD45RA+CCR7+ naive (Tn: 22%, 4-50%). Contrarily, most CD4+ T cells belonged to CD45RA-CCR7+ central memory (Tcm: 45%, 24-78%), effector memory (Tem: 27%, 6-61%) and naive (Tn: 22%, 1-50%) subtypes. Interestingly, the majority of monocytes (86%, 78-97%) were CD86+ suggesting a role in inducing lymphocyte activation. Of activation/exhaustion markers, the highest proportion of PDL1+, PD1+, IFNg+ and ICOS+ cells were observed in monocytes, cDC, NK and CD4+ T cells respectively. Likewise, CTLA4+, VISTA+ and TIGIT+ cells were most commonly found among CD8+ T cells, suggesting a potential regulatory effect on sip T CD8+ T cell proliferation and effector function. Conclusion: This is the first comprehensive study to evaluate the composition of sip T from mCRPC patients using high dimensional CyTOF analysis. Given the increasing use of CyTOF in translational research, this data set serves as an important reference source, which could be used in future studies to help guide therapeutic modulation of activation/exhaustion markers to improve sip T efficacy and enhance beneficial immune responses. Citation Format: Muhammad Saeed, Kevin Kim, Ariel Borkowski, Russell Kent Pachynski. High dimensional analysis of sipuleucel T from metastatic castration resistant prostate cancer patients using mass cytometry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4158.

Auditory Target Detection Enhances Visual Processing and Hippocampal Functional Connectivity.

Though dividing one’s attention between two input streams typically impairs performance, detecting a behaviorally relevant stimulus can sometimes enhance the encoding of unrelated information presented at the same time. Previous research has shown that selection of this kind boosts visual cortical activity and memory for concurrent items. An important unanswered question is whether such effects are reflected in processing quality and functional connectivity in visual regions and in the hippocampus. In this fMRI study, participants were asked to memorize a stream of naturalistic images and press a button only when they heard a predefined target tone (400 or 1,200 Hz, counterbalanced). Images could be presented with a target tone, with a distractor tone, or without a tone. Auditory target detection increased activity throughout the ventral visual cortex but lowered it in the hippocampus. Enhancements in functional connectivity between the ventral visual cortex and the hippocampus were also observed following auditory targets. Multi-voxel pattern classification of image category was more accurate on target tone trials than on distractor and no tone trials in the fusiform gyrus and parahippocampal gyrus. This effect was stronger in visual cortical clusters whose activity was more correlated with the hippocampus on target tone than on distractor tone trials. In agreement with accounts suggesting that subcortical noradrenergic influences play a role in the attentional boost effect, auditory target detection also caused an increase in locus coeruleus activity and phasic pupil responses. These findings outline a network of cortical and subcortical regions that are involved in the selection and processing of information presented at behaviorally relevant moments.

Phonological similarity in the serial recall task hinders item recall, not just order.

The phonological similarity effect in short-term memory (STM) is the finding that serial recall of lists of similar sounding items is poorer than that of dissimilar sounding items. This is traditionally argued to be due to a detrimental effect on memory for the order of the words in the list and not on memory for the words themselves. Models that propose forgetting from STM is due to interference must invoke an additional compensatory process where the shared element of the words acts as a cue to recall, in order to account for the lack of an effect on memory for the words. However, the possibility of a detrimental effect of phonological similarity on item memory when these compensatory processes are not available has not been investigated. Two experiments (n=60 and n=57) in which similarity is operationalized in a way that precludes usage of compensatory processes are reported. The results clearly demonstrate that item recall is poorer for similar lists than dissimilar lists when similarity is defined in this way.

Training modulates memory-driven capture

Attention is captured by information matching the contents of working memory. Though many factors modulate the amount of capture, there is surprising resistance to cognitive control. Capture occurs even when participants are instructed either that an item would never be a target or to drop that item from memory. Does the persistence of capture under these conditions reflect a rigidity in capture, or can properly motivated participants learn to completely suppress distractors and/or completely drop items from memory? Surprisingly, no studies have looked at the influence of extensive training of involuntary capture from working memory items. Here, we addressed whether training leads to a reduction or even elimination of memory-driven capture. After memorizing a single object, participants were cued to remember or to forget this object. Subsequently, they were asked to execute a search task. To measure capture, we compared search performances in displays that did and did not contain a distractor matching the earlier memorized object. Participants completed multiple experimental sessions over four days. The results showed that attentional capture by to-be-remembered distractors was reduced, but not eliminated in subsequent sessions compared with the first session. Training did not impact capture by to-be-forgotten objects. The results suggest observable, but limited, cognitive control over memory-driven capture.

Trial-by-trial mouse trajectory predicts variance in precision across working memory representations: A critical reanalysis of Hao et al. (2021).

Multiple representations in visual working memory (VWM) can vary in mnemonic precision. This inhomogeneity of VWM precision has received some support from recent studies with the whole-report procedure, in which all memory items are recalled in free or forced orders. Recently, Hao et al. (2021, Cognition, 214, 104739) added a novel item-selection stage before each memory recall and found smaller between-trial variance in mouse trajectory during the selection stage in free-recall condition as compared with forced recall, which was taken as evidence for less between-item interference and the resulting precision benefit under free recall. Here, we reanalyzed the original dataset with a different analytic approach and attempted independent hypothesis testing focusing on within-trial trajectory deviations. We found that the direction of trial-by-trial trajectory bias for the first to-be-recalled item was predictive of the relative mnemonic precision of the remaining items. Critically, this relationship was only present for forced recall but not for free recall. Hierarchical Bayesian modeling of recall errors further identified that this relationship was selectively driven by VWM precision. Together, our reanalysis provides evidence for the source of between-item interference and its direct association with variable precision of VWM representations, and further highlights the novel methodological benefits of probing memory decisional processes using mouse trajectory data.

Priority-based transformations of stimulus representation in visual working memory.

How does the brain prioritize among the contents of working memory (WM) to appropriately guide behavior? Previous work, employing inverted encoding modeling (IEM) of electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) datasets, has shown that unprioritized memory items (UMI) are actively represented in the brain, but in a "flipped", or opposite, format compared to prioritized memory items (PMI). To acquire independent evidence for such a priority-based representational transformation, and to explore underlying mechanisms, we trained recurrent neural networks (RNNs) with a long short-term memory (LSTM) architecture to perform a 2-back WM task. Visualization of LSTM hidden layer activity using Principal Component Analysis (PCA) confirmed that stimulus representations undergo a representational transformation-consistent with a flip-while transitioning from the functional status of UMI to PMI. Demixed (d)PCA of the same data identified two representational trajectories, one each within a UMI subspace and a PMI subspace, both undergoing a reversal of stimulus coding axes. dPCA of data from an EEG dataset also provided evidence for priority-based transformations of the representational code, albeit with some differences. This type of transformation could allow for retention of unprioritized information in WM while preventing it from interfering with concurrent behavior. The results from this initial exploration suggest that the algorithmic details of how this transformation is carried out by RNNs, versus by the human brain, may differ.

Neural Patterns in Parietal Cortex and Hippocampus Distinguish Retrieval of Start versus End Positions in Working Memory

Coding serial order of information is a fundamental ability of our cognitive system, and still, little is known about its neural substrate. This study examined the neural substrates involved in the retrieval of information that is serially stored in verbal working memory task using a sensitive multivariate analysis approach. We compared neural activity for memorized items stemming from the beginning versus the end of a memory list assessing the degree of neural pattern discordance between order positions (beginning vs. end). The present results confirmed and refined the role of the intraparietal sulcus in the processing of serial order information in working memory. An important finding is that the hippocampus showed sensitivity to serial order information. Our results indicate that the representation of serial order information relies on a broader set of neural areas and highlight the role of the intraparietal sulcus and the hippocampus, in addition to the supramarginal gyrus and the SMA. The contribution of different neural regions might reflect the involvement of distinct levels of serial order coding (i.e., spatial, attentional, temporal) that support the representation of serial order information.

Striking effect of low-dose radiotherapy combined with PD-1 blockade on small cell lung cancer in mice and refractory patients (Achilles Study).

e20608 Background: Immune checkpoint inhibitors (ICIs) provide a durable and long-term benefit but still unsatisfactory clinical efficacy for extensive-stage small cell lung cancer (ES-SCLC). An optimized dose and schedule of radiotherapy for ES-SCLC remain to be studied. We hypothesized that the addition of low-dose radiotherapy (LDRT) to ICI could improve the efficacy of ES-SCLC. Methods: In SCLC bearing mice, the tumor was irradiated with LDRT in dose-escalation starting at 3Gy×1 fractions (f) up to 3Gy×7f, and combined with PD-1 antibody injection (0.2mg/mouse, i.p, every 3 days). Mice were followed for tumor growth and survival. The tumor microenvironment was dynamically analyzed every 3 days till day 18 by flow cytometry and immunofluorescence. 15 patients (pts) with relapsed ES-SCLC who received the treatment with LDRT and PD-1 blockade were retrospectively reviewed. Results: LDRT with 3Gy×5f delivered over 5 days (LDRT15Gy/5f) was determined to be the optimized dose when combined with PD-1 blockade. Combined group exhibited obvious growth retardation and prolonged survival compared to either monotherapy.The most robust infiltration of T cells in combined group was observed on day 9 after start of treatment. Bulk and single-cell RNA-seq showed significant immune cells infiltration, predominately CD8+ T cells and M1 macrophage. The activation and degranulation of CD8+ T cells in combination group were enhanced. Intra-tumoral CD8+ T cells were mainly assigned to effector memory (Tem) cells, then exhausted precursor (Texp) cells and exhausted (Tex) cells. Pseudo-time analysis supported a state evolution model that CD8+ Tem cells differentiate into intra-tumoral CD8+ Tex cells through an intermediate CD8+ Texp state. Tumor cells were divided into clusters 0, 1, and 2. Cluster 0 nearly disappeared while the proportion of cluster 1 increased in combination therapy. Cluster 1 enriched inflammatory response pathways and higher expression of MHC class I versus clusters 0 and 2. Additionally, the density of intra-tumoral microvessel was decreased, while vascular perfusion and the number of pericyte-covered vessels increased in combined group. Correspondingly, 15 recurrent ES-SCLC pts who treated with up-front LDRT15Gy/5f plus PD-1 blockade showed an ORR of 80%. Median PFS and OS were 4.3 months and 10.9 months. PFS rates at 6, 12 and 24 months were 40%, 20% and 6.7%, and OS rates were 60%, 40% and 24%, respectively. No patient experienced above grade 4 radiation- and immunotherapy-related toxicity. Conclusions: Our study is the first report of the synergistic effect of LDRT15Gy/5f and PD-1 blockade in SCLC mice model and recurrent ES-SCLC pts. The LDRT15Gy/5f demonstrates both immunologic adjuvant and cytotoxic effect on SCLC, and is safe and feasible in clinical pts. Further translational research, Match trial (NCT04622228), is ongoing.

Categorical bias as a crucial parameter in visual working memory: The effect of memory load and retention interval

Visual information can be stored as continuous as well as categorical representations in visual working memory (VWM) to guide subsequent behavior. Yet it is still unclear what determines whether VWM is represented as continuous or categorical information, or as a mix of both. Recent studies have shown that color VWM representations adjust flexibly depending on the number of memory items as well as the duration that these items need to be maintained for. The current study aims to extend and replicate these crucial effects. In a delayed estimation task, participants memorized one to four colored objects presented at different spatial locations, followed by a delay of 100, 500, 1000, or 2000 msec. Next, a probe indicated the location of the color that participants needed to report. We measured the extent to which responses were biased in the direction of prototypical colors. Crucially, we implemented this categorical bias in an extension to the classic mixture model (Zhang & Luck, 2008) in which the center of the error distribution is a crucial parameter that characterizes the extent to which VWM is biased by color categories. We found that VWM shows a strong categorical bias in all cases, and that this bias increases with increasing memory load; strikingly, this effect of memory load on categorical bias is stronger at longer intervals (1000 msec or longer), as compared to shorter intervals, yet it peaks for intermediate memory loads as opposed to the highest memory load. Overall, our results suggest that when visual information needs to be maintained for one second or longer, VWM becomes more reliant on categorical representations as memory load increases.

0106 Unitization Improves Memory for Associations during Sleep Deprivation

Abstract Introduction Total sleep deprivation (TSD) impairs binding, i.e., the ability to form new associations. Unitization – when separate memory items are learned as a single unit (e.g., combining two words into a novel compound word) – reduces the need for binding. Unitization mitigates impaired memory for associations in amnesiacs, but whether it offsets binding problems from TSD is unknown. Methods N=23 healthy adults (ages 19-35, 8 women) participated in an ongoing, double-blind, 4-day/3-night in-laboratory study with a 10h baseline sleep opportunity, 38h TSD, and a 10h recovery sleep opportunity. During TSD, participants were randomized to four administrations of caffeine (200mg), modafinil (alternating between 200mg and 0mg), or placebo at 4h intervals beginning at 01:00. They completed a unitization task at 14:45 on day 2 (baseline, session 1), day 3 (TSD, session 2), and day 4 (recovery, session 3). The task began with a study phase where participants studied 60 pairs of words that were presented individually (e.g., “penny” and “tower”) or as new, unitized words (e.g., “pennytower”) (50% each). Afterward, in the test phase, participants indicated whether 60 presented pairs of individual words were old (presented together at study) or new (recombined into new pairs) (50% each). Results Repeated-measures ANOVA revealed significant effects of study pair type (individual or unitized), session (1–3), and their interaction (p&amp;lt;0.05). Performance did not differ by pair type in session 1 (p=0.46), and performance for pairs of individual words did not change across sessions (p=0.34). However, performance on unitized word pairs improved across sessions (p=0.003), and unitized word pairs were recognized better than individual word pairs in sessions 2 and 3 (p&amp;lt;0.05). Conclusion Across sessions, participants benefitted from practice on unitized word pairs, such that performance improved even during TSD. Although potentially partly attributable to drug condition (to which investigators are still blinded), no such practice effect was seen for word pairs studied individually. Whether this dissociation implies that unitization bypasses the need for binding and thus lessens the impact of TSD requires further investigation. Regardless, unitization may mitigate performance impairment from sleep loss in settings that require forming novel associations, such as eyewitness identifications. Support (If Any) USAMRDC W81XWH-18-1-0100 and CDMRP W81XWH-20-1-0442.

Role of Sleep in Formation of Relational Associative Memory.

Relational memory, the ability to make and remember associations between objects, is an essential component of mammalian reasoning. In relational memory tasks, it has been shown that periods of offline processing, such as sleep, are critical to making indirect associations. To understand biophysical mechanisms behind the role of sleep in improving relational memory, we developed a model of the thalamocortical network to test how slow-wave sleep affects performance on an unordered relational memory task. First, the model was trained in the awake state on a paired associate inference task, in which the model learned to recall direct associations. After a period of subsequent slow-wave sleep, the model developed the ability to recall indirect associations. We found that replay, during sleep, of memory patterns learned in awake increased synaptic connectivity between neurons representing the item that was overlapping between tasks and neurons representing the unlinked items of the different tasks; this forms an attractor that enables indirect memory recall. Our study predicts that overlapping items between indirectly associated tasks are essential for relational memory, and sleep can reactivate pathways to and from overlapping items to the unlinked objects to strengthen these pathways and form new relational memories.SIGNIFICANCE STATEMENT Experimental studies have shown that some types of associative memory, such as transitive inference and relational memory, can improve after sleep. Still, it remains unknown what specific mechanisms are responsible for these sleep-related changes. In this new work, we addressed this problem by building a thalamocortical network model that can learn relational memory tasks and that can be simulated in awake or sleep states. We found that memory traces learned in awake were replayed during slow waves of NREM sleep and revealed that replay increased connections to and from overlapping memory items to form new relational memories. Our work discovered specific mechanisms behind the role of sleep in associative memory and made testable predictions about how sleep augments associative learning.

The Influence of Active Removal from Working Memory on Serial Dependence.

Flexible control of the contents of working memory (WM) includes removing no-longer-relevant information. Although simply withdrawing attention offers a “passive” mechanism, empirical findings suggest that it is also possible to actively remove information from WM. In this Registered Report we tested evidence that the bias (serial dependence) that an item exerts on the subsequent trial will be opposite in sign—attraction vs. repulsion — depending on whether it was passively or actively removed, respectively. A repulsive bias would be consistent with a specific mechanism for active removal: a rapid adaptation-like modification of perceptual circuitry. In a preliminary study, trials of two types were administered in pairs, multi-item WM followed by 1-item delayed recall, and we evaluated serial dependence of the latter on items from the former. In the first trial of each pair, two memoranda were presented, then one was designated irrelevant, then a third memorandum was presented. The critical manipulation was whether the third item was presented at the same location as the now “irrelevant memory item” (IMI). Overlap between the two should prompt the active removal of the IMI, whereas nonoverlap might prompt just the withdrawal of attention. Whereas the IMI exerted the expected attractive bias on 1-item recall in the no-overlap condition, we found an (unexpected) repulsive bias in the overlap condition. Because repulsive biases have been attributed to the adaptation-like modification of perceptual circuitry, replication of this result in this Registered Report would provide independent evidence for this mechanism for active removal from WM. Interpretation of the Stage 2 results are complicated by the fact that the approved Registered Report, carried out online, generated data that failed to meet a basic sanity check, and were therefore uninterpretable. Consequently, a follow-up lab-based experiment using procedures similar to the Registered Report generated results consistent with the hypothesis of principal theoretical interest: The IMI in the overlap condition exerted a repulsive bias on the subsequent trial.

Encoding tasks moderated the reward effect on brain activity during memory retrieval

Previous studies have explored the effects of retrieval reward and depth of processing in encoding on recognition, but it remains unclear whether and how reward and depth of processing during encoding influence recognition. We investigated the effect and neural mechanisms of encoding reward and processing depth on recognition using event-related potentials (ERPs) in this study. In the study phase, participants were asked to perform two encoding tasks: congruity-judgment (deep processing) and size-judgment (shallow processing) in reward and no-reward conditions. The test phases included object (item) and background (source) tests. The results of item retrieval showed that the accuracy of rewarded items was higher than that of unrewarded items only in the congruity-judgment task, and the reward effect (the average amplitudes in the reward condition were significantly more positive than those in the no-reward condition) in the 300–500 and 500–700 ms were greater in the congruity-judgment task than in the size-judgment task. The results of source retrieval showed that the accuracy of rewarded items was higher than that of unrewarded items, that the difference in the size-judgment task was significantly larger, and that the reward effect in the 300–500 and 500–700 ms were greater in the size-judgment task than in the congruity-judgment task. In conclusion, the encoding task moderated the reward effect in item and source memory.

The attentional boost effect facilitates the encoding of contextual details: New evidence with verbal materials and a modified recognition task.

In the attentional boost effect (ABE), words or images encoded with to-be-responded targets are later recalled better than words or images encoded with to-be-ignored distractors. The ABE has been repeatedly demonstrated to improve item memory, whereas evidence concerning contextual memory is mixed, with studies showing both significant and null results. The present three experiments investigated whether the ABE could enhance contextual memory when using a recognition task that allowed participants to reinstate the original study context, by simultaneously manipulating the nature of the instructions provided at encoding. Participants studied a sequence of colored words paired with target (gray circles) or distractor (gray squares) stimuli, under the instructions to remember either the words and their colors (Exps. 1–2) or only the words (Exp. 3) and simultaneously press the space bar whenever a gray circle appeared on the screen. Then, after a brief interval, they were administered a modified recognition task involving two successive stages. First, participants were presented with two different words and had to decide which word was originally encoded; second, they were presented with five colored versions of the (correct) old words and had to remember the color in which they were studied. Results converged in showing that the ABE enhanced contextual memory, although the effect was more robust with intentional encoding instructions.Supplementary InformationThe online version contains supplementary material available at 10.3758/s13414-022-02509-z.