Comparing memory capacity across stimuli requires maximally dissimilar foils: Using deep convolutional neural networks to understand visual working memory capacity for real-world objects

The capacity of visual working and visual long-term memory plays a critical role in theories of cognitive architecture and the relationship between memory and other cognitive systems. Here, we argue that before asking the question of how capacity varies across different stimuli or what the upper bound of capacity is for a given memory system, it is necessary to establish a methodology that allows a fair comparison between distinct stimulus sets and conditions. One of the most important factors determining performance in a memory task is target/foil dissimilarity. We argue that only by maximizing the dissimilarity of the target and foil in each stimulus set can we provide a fair basis for memory comparisons between stimuli. In the current work we focus on a way to pick such foils objectively for complex, meaningful real-world objects by using deep convolutional neural networks, and we validate this using both memory tests and similarity metrics. Using this method, we then provide evidence that there is a greater capacity for real-world objects relative to simple colors in visual working memory; critically, we also show that this difference can be reduced or eliminated when non-comparable foils are used, potentially explaining why previous work has not always found such a difference. Our study thus demonstrates that working memory capacity depends on the type of information that is remembered and that assessing capacity depends critically on foil dissimilarity, especially when comparing memory performance and other cognitive systems across different stimulus sets.

Fermented seeds of Pentaclethra macrophylla mitigate against memory deficit and restored altered enzymatic activity in the brain of streptozotocin-diabetic rats

Memory deficit has been reported as one of the complications of diabetes. Fermented seeds of Pentaclethra Macrophylla (P. macrophylla) have been used in folklore for the management of metabolic diseases. The research aims to evaluate the impact of diets with the inclusion of the fermented seed of P. macrophylla on memory deficit in diabetic rats and its underlying mechanisms. Before the induction, the rats were subjected to training sessions. Thereafter, streptozotocin (50 mg/kg body weight) was administered to the trained rats via intraperitoneal (i.p). 72 hours after, the rats blood glucose level was checked, rats with blood glucose level greater than 250 mg/dl were selected for the memory index evaluation study. The induced rats were randomly distributed into groups: Normal rats (group 1), untreated diabetic rat (Group 2), acarbose treated diabetic rats (group 3); diabetic rats placed on diet supplemented with fermented seed of P. macrophylla (10 & 20% inclusion) were allotted to group 4 & 5. Then, evaluation of memory retention capacity was performed on the day 14 of the experiment. Thereafter, experimental rats were sacrificed, tissue of interest (brain) was excised, homogenized and homogenates were used for biochemical analysis. The cholinergic, angiotensin-1-converting enzyme (ACE), arginase activity and biomarkers for oxidative stress were significantly altered in untreated diabetic rats when compared with non-diabetes rats. Also, the memory capacity of the diabetic rats was significantly reduced when compared with the non-diabetes rats. Meanwhile, diabetic rats placed on diet with fermented seeds of P. macrophylla (10 & 20% inclusion) exhibited significantly higher memory capacity, lower activity of cholinergic, ACE, arginase activity in relation to untreated diabetic rats while the antioxidant status of the brain was enhanced. Nevertheless, fermented seeds of P. macrophylla ameliorated memory deficit in STZ induced diabetes rats. This gave credence to P. macrophylla nutraceutical potential as claimed in folk medicine.

Long-term memory retrieval bypasses working memory

For decades, it has been assumed that when humans retrieve information from long-term memory (LTM), information need first to be brought back into working memory (WM). However, as WM capacity is limited, it is unclear what happens if information from LTM needs to be retrieved while WM is fully engaged? To address this question, observers had to retrieve colors from LTM while WM storage capacity was fully engaged. The behavioral results showed that retrieving information from LTM is possible even when WM capacity is fully occupied. Additional evidence from electroencephalogram (EEG) confirmed that WM was fully engaged as the suppression of alpha oscillation reached its maximum when memorizing the maximum amount of information into WM; yet the suppression in alpha oscillation was even further amplified when items were retrieved simultaneously from LTM, providing a neural signature of additional LTM retrieval capacity above and beyond the maximum WM capacity. Together, our findings indicate that information retrieved from LTM does not always have to be brought back into WM, but instead might be accessed through a different mechanism when WM is fully engaged.

Impact of processing demands at encoding, maintenance and retrieval in visual working memory

There has been surprisingly little examination of how recall performance is affected by processing demands induced by retrieval cues, how manipulations at encoding interact with processing demands during maintenance or due to the retrieval cue, and how these are affected with aging. Here, we investigate these relationships by examining the fidelity of working memory recall across two delayed reproduction tasks with a continuous measure of report across the adult lifespan. Participants were asked to remember and subsequently reproduce from memory the identity and location of a probed item from the encoding display. In Experiment 1, we examined the effect of filtering irrelevant information at encoding and the impact of filtering distracting information at retrieval simultaneously. In Experiment 2, we tested how ignoring distracting information during maintenance or updating current contents with new information during this period affects recall. The results reveal that manipulating processing requirements induced by retrieval cues (by altering the nature of the retrieval foil) had a significant impact on memory recall: the presence of two previously viewed features from the encoding display in the retrieval foil led to a decrease in identification accuracy. Although irrelevant information can be filtered out well at encoding, both ignoring irrelevant information and updating the contents of memory during the maintenance delay had a detrimental effect on recall. These effects were similar across the lifespan, but older individuals were particularly affected by manipulations of processing demands at encoding as well as increasing set size of information to be retained in memory. Finally, analyses revealed that there were no systematic relationships between filtering performance at encoding, maintenance and retrieval suggesting that these processing demands are independent of each other. Rather than filtering being a single, monolithic entity, the data suggest that it is better accounted for as distinctly dissociable cognitive processes that engage and articulate with different phases of working memory.

Sequential Presentation Protects Working Memory From Catastrophic Interference

Neural network models of memory are notorious for catastrophic interference: Old items are forgotten as new items are memorized (French, 1999; McCloskey & Cohen, 1989). While working memory (WM) in human adults shows severe capacity limitations, these capacity limitations do not reflect neural network style catastrophic interference. However, our ability to quickly apprehend the numerosity of small sets of objects (i.e., subitizing) does show catastrophic capacity limitations, and this subitizing capacity and WM might reflect a common capacity. Accordingly, computational investigations (Knops, Piazza, Sengupta, Eger & Melcher, 2014; Sengupta, Surampudi & Melcher, 2014) suggest that mutual inhibition among neurons can explain both kinds of capacity limitations as well as why our ability to estimate the numerosity of larger sets is limited according to a Weber ratio signature. Based on simulations with a saliency map-like network and mathematical proofs, we provide three results. First, mutual inhibition among neurons leads to catastrophic interference when items are presented simultaneously. The network can remember a limited number of items, but when more items are presented, the network forgets all of them. Second, if memory items are presented sequentially rather than simultaneously, the network remembers the most recent items rather than forgetting all of them. Hence, the tendency in WM tasks to sequentially attend even to simultaneously presented items might not only reflect attentional limitations, but also an adaptive strategy to avoid catastrophic interference. Third, the mean activation level in the network can be used to estimate the number of items in small sets, but it does not accurately reflect the number of items in larger sets. Rather, we suggest that the Weber ratio signature of large number discrimination emerges naturally from the interaction between the limited precision of a numeric estimation system and a multiplicative gain control mechanism.

Data from brain activity during visual working memory replicates the correlation between contralateral delay activity and memory capacity

This article provides data from statistical analysis of event-related brain potentials (ERPs) and behavioural performance from 23 participants during a working memory task. Specifically, we used the change detection task from Vogel 2004, using the same timing but a modified size and distance between stimuli. Contralateral delay activity (CDA) was calculated from posterior parieto-occipital electrodes and then it was compared between conditions with different memory load (one, two and four items). Working memory capacity (WMC) was calculated from behavioural data using the formula developed by Pashler (1988). Correlation was performed between WMC and the CDA amplitude difference (from two to four items). The correlation replicates the results from the original paper of Vogel 2004 [1], even though some parameters are different from the original design.

Performance boost of time-delay reservoir computing by non-resonant clock cycle

The time-delay-based reservoir computing setup has seen tremendous success in both experiment and simulation. It allows for the construction of large neuromorphic computing systems with only few components. However, until now the interplay of the different timescales has not been investigated thoroughly. In this manuscript, we investigate the effects of a mismatch between the time-delay and the clock cycle for a general model. Typically, these two time scales are considered to be equal. Here we show that the case of equal or resonant time-delay and clock cycle could be actively detrimental and leads to an increase of the approximation error of the reservoir. In particular, we can show that non-resonant ratios of these time scales have maximal memory capacities. We achieve this by translating the periodically driven delay-dynamical system into an equivalent network. Networks that originate from a system with resonant delay-times and clock cycles fail to utilize all of their degrees of freedom, which causes the degradation of their performance.

Ferulic acid ameliorates pentylenetetrazol-induced seizures by reducing neuron cell death

To investigate the neuroprotective effect of ferulic acid (FA) in a pentylenetetrazol (PTZ)-induced seizures model in rat, the motor response, spatial learning ability and memory capability of the rats were assessed. Both the antioxidation and anti-apoptosis pathways were also investigated. In this study, male Wistar rats were randomly divided into 3 groups (n = 12 in each group). For 28 days, the rats were administered saline alone (i.p. normal saline, NS group), PTZ (40 mg/kg, i.p., PTZ group) once daily to induce seizures, or FA (i.p. 60 mg/kg) 20 min before being given PTZ (40 mg/kg, i.p., FA + PTZ group) to assess the neuroprotective effect of FA. The motor response of the rats was analysed with the Racine scale. The spatial learning and memory capacity of the rats were assessed by the Morris water maze test. The superoxide dismutase (SOD) activity and malondialdehyde (MDA) content were measured, and both in situ staining with the DNA-binding bisbenzimide Hoechst 33258 and TUNEL assays were used to assess apoptosis. Western blotting was used to further analyse the expression of Apaf-1, caspase-9, caspase-3, Bcl-2, Bid, Bax, cleaved caspase-3 and cytochrome c. The results showed that compared to the those of the PTZ group, FA pre-treatment significantly (p <  0.01) reduced the Racine scores starting at day 4, prolonged the latency of the onset of seizure at day 28, reduced the escape latency period starting at day 2, increased the frequency of crossing the platform location, increased the SOD activity, reduced the MDA content and apoptosis percentage, and upregulated the Bcl-2 levels whilst downregulating the Bax, cytochrome c, Apaf-1, caspase-9, caspase-3, cleaved caspase-3 and Bid expression levels. This study demonstrated that pre-treatment with FA exerts strong neuroprotective effects by reducing the motor response and by improving spatial learning ability and memory capacity. The neuroprotective effect may be a result of a reduction in neuron cell death that occurs via the antioxidative and anti-apoptotic pathways.

The cost of proactive interference is constant across presentation conditions

Proactive interference (PI) severely constrains how many items people can remember. For example, Endress and Potter (2014a) presented participants with sequences of everyday objects at 250ms/picture, followed by a yes/no recognition test. They manipulated PI by either using new images on every trial in the unique condition (thus minimizing PI among items), or by re-using images from a limited pool for all trials in the repeated condition (thus maximizing PI among items). In the low-PI unique condition, the probability of remembering an item was essentially independent of the number of memory items, showing no clear memory limitations; more traditional working memory-like memory limitations appeared only in the high-PI repeated condition. Here, we ask whether the effects of PI are modulated by the availability of long-term memory (LTM) and verbal resources. Participants viewed sequences of 21 images, followed by a yes/no recognition test. Items were presented either quickly (250ms/image) or sufficiently slowly (1500ms/image) to produce LTM representations, either with or without verbal suppression. Across conditions, participants performed better in the unique than in the repeated condition, and better for slow than for fast presentations. In contrast, verbal suppression impaired performance only with slow presentations. The relative cost of PI was remarkably constant across conditions: relative to the unique condition, performance in the repeated condition was about 15% lower in all conditions. The cost of PI thus seems to be a function of the relative strength or recency of target items and interfering items, but relatively insensitive to other experimental manipulations.