Cognitive Tasks Research Articles

Characterizing molecular and synaptic signatures in mouse models of late-onset Alzheimer's disease independent of amyloid and tau pathology.

MODEL-AD (Model Organism Development and Evaluation for Late-Onset Alzheimer's Disease) is creating and distributing novel mouse models with humanized, clinically relevant genetic risk factors to capture the trajectory and progression of late-onset Alzheimer's disease (LOAD) more accurately. We created the LOAD2 model by combining apolipoprotein E4 (APOE4), Trem2*R47H, and humanized amyloid-beta (Aβ). Mice were subjected to a control diet or a high-fat/high-sugar diet (LOAD2+HFD). We assessed disease-relevant outcome measures in plasma and brain including neuroinflammation, Aβ, neurodegeneration, neuroimaging, and multi-omics. By 18 months, LOAD2+HFD mice exhibited sex-specific neuron loss, elevated insoluble brain Aβ42, increased plasma neurofilament light chain (NfL), and altered gene/protein expression related to lipid metabolism and synaptic function. Imaging showed reductions in brain volume and neurovascular uncoupling. Deficits in acquiring touchscreen-based cognitive tasks were observed. The comprehensive characterization of LOAD2+HFD mice reveals that this model is important for preclinical studies seeking to understand disease trajectory and progression of LOAD prior to or independent of amyloid plaques and tau tangles. By 18 months, unlike control mice (e.g., LOAD2 mice fed a control diet, CD), LOAD2+HFD mice presented subtle but significant loss of neurons in the cortex, elevated levels of insoluble Ab42 in the brain, and increased plasma neurofilament light chain (NfL). Transcriptomics and proteomics showed changes in gene/proteins relating to a variety of disease-relevant processes including lipid metabolism and synaptic function. In vivo imaging revealed an age-dependent reduction in brain region volume (MRI) and neurovascular uncoupling (PET/CT). LOAD2+HFD mice also demonstrated deficits in acquisition of touchscreen-based cognitive tasks.

Variants of the P3 event-related potential operate as indicators of distinct mechanisms contributing to problematic alcohol use.

Considerable research has linked relative reduction in the amplitude of the P3 event-related potential (ERP) during cognitive task performance (i.e., Target-P3) with increased risk of alcohol-related problems. A separate literature indicates that a relative increase in the amplitude of the P3 elicited by cues signaling alcohol availability (i.e., ACR-P3) also is associated with alcohol use and problems. To date, no research has integrated these seemingly discrepant findings. Here, we aimed to demonstrate that P3 amplitudes elicited in different task contexts reflect distinct domains of functioning relevant to problematic alcohol involvement (PAI), and therefore can inform heterogeneity in the etiology of PAI. 156 emerging adults (61% women; 88% White/Non-Hispanic) completed a mental rotation task and a picture-viewing task while ERPs were recorded. Participants also completed questionnaire measures of trait disinhibition, alcohol use, and alcohol-related problems. Findings from regression analyses indicated that (a) Target-P3 was negatively associated and ACR-P3 was positively associated with a PAI latent variable; (b) the two P3s accounted for unique variance in PAI, beyond that accounted for by recent drinking; and (c) the association between Target-P3 and PAI-but not ACR-P3 and PAI-was statistically mediated by trait disinhibition. The present findings highlight the unique contributions of distinct functional domains associated with disinhibition and incentive salience in the etiology of PAI. Moreover, findings are consistent with a nuanced understanding of the P3 ERP, whereby its specific meaning varies according to the task context in which it is elicited.

Self-regulated use of retrieval practice: associations with individual differences in non-cognitive and cognitive factors

AbstractRetrieval practice is a learning strategy that has repeatedly been found to have positive effects on memory and learning. However, studies indicate that students rarely use retrieval practice on a voluntary basis. The objective of the present study was to examine students’ self-regulated use of retrieval practice, and to determine whether sex and individual differences in cognitive and non-cognitive aspects are related to optional use of practice testing. A classroom study was conducted with 146 upper-secondary school students taking courses in mathematics and Swedish. An ABAB design was used to compare students’ optional and non-optional use of retrieval practice (i.e., repeated online quizzing). Students performed cognitive tasks to assess working memory capacity and fluid intelligence and completed self-reports of non-cognitive factors related to school achievement, such as grit, need for cognition (NFC), conscientiousness and openness. Quiz use was then compared using paired- and independent-samples t-tests, and hierarchical linear regression analyses explored relations to individual differences. The results showed that students completed significantly fewer quizzes in the optional sections than in the non-optional sections, and that females completed significantly more optional quizzes than males in Swedish, but not in mathematics. Further, the results showed that conscientiousness predicted optional quiz use in mathematics, whereas sex, NFC, conscientiousness, and openness predicted quiz use in Swedish. To conclude, although the findings show a relatively low optional/self-regulated use of practice testing, in line with earlier research, they suggest that sex and non-cognitive factors, such as personality characteristics, can predict optional use of practice testing.

Neural mechanisms of the relationship between aerobic fitness and working memory in older adults: An fNIRS study

Abstract A growing number of studies have revealed that higher aerobic fitness is associated with better working memory (WM) performance in older adults. However, the underlying functional neural mechanisms of this association remain under debate. It has been reported that aging increases recruitment of the prefrontal cortex (PFC) during cognitive tasks, and that this is associated with task performance in a compensatory manner. Therefore, this study aimed to clarify the prefrontal activation pattern that is associated with the relationship between aerobic fitness and WM performance in older adults, focusing on age-related extended prefrontal recruitment. Forty-seven older adults (65–74 years, 29 females) and 49 younger adults (18–24 years, 23 female) performed verbal and spatial n-back tasks, which included 0-, 1-, and 2-back conditions. Reaction time (RT) and accuracy (ACC) were assessed as indices of task performance. Prefrontal activation during the experimental tasks was monitored using functional near-infrared spectroscopy (fNIRS) and analyzed using an adaptive GLM method. We compared task performance and prefrontal activation between age groups to find age-related prefrontal activation patterns. Only older adults underwent a graded exercise test (GXT) to determine their ventilation thresholds (VT) as a measure of aerobic fitness, and, subsequently, the relationships among aerobic fitness, n-back task performance, and prefrontal activation in older adults were examined using correlation analysis and mediation analysis controlling for possible covariates. A comparison of task performance between groups revealed that older adults had slower RT and lower ACC than did younger adults, especially in the higher WM load 2-back condition. Group comparisons of prefrontal activation showed that older adults exhibited additional or greater activation than younger adults mainly in the ventrolateral PFC (VLPFC) and front polar area (FPA) in both the verbal and spatial 2-back conditions. Correlation analysis showed a relationship between higher VT, shorter RT for the verbal 2-back condition, and greater prefrontal activation of the bilateral FPA and right VLPFC during verbal 2-back conditions in older adults. In addition, mediation analyses indicated the possibility of a mediation effect of the prefrontal activation on the relationship between VT and RT for the verbal 2-back condition. These results suggest that older adults with higher aerobic fitness levels recruited more extended PFC regions, possibly for compensatory activation, to enhance their performance of the verbal n-back task. This study sheds light on the neural mechanisms underpinning the relationship between aerobic fitness and cognitive function in older adults.

Testing the Model of Judicial Stress using a COVID-era survey of U.S. federal court personnel

Non-judicial court personnel, critical to a well-functioning justice system, experience overloaded dockets and the responsibility of making significant decisions, contributing to cognitive stress. Understanding and mitigating their stress is essential for maintaining judicial efficiency. We adapted Miller and Richardson’s Model of Judicial Stress to assess stress in a broad sample of non-judicial court personnel (n = 122), including judges, lawyers, and administrative staff. Participants responded to surveys about their stress levels, job performance, and health; they also completed cognitive performance tasks. The findings indicated that stress negatively affected employee outcomes including cognitive performance, job performance, job satisfaction, and health outcomes. Notably, perceived job performance had declined compared to the previous year, suggesting that the pandemic was an additional significant stressor. Based on the data, the Model of Judicial Stress is also applicable to other types of courtroom personnel, underlining its relevance across various judicial roles.

Revisiting the Deviation Effects of Irrelevant Sound on Serial and Nonserial Tasks.

Two types of disruptive effects of irrelevant sound on visual tasks have been reported: the changing-state effect and the deviation effect. The idea that the deviation effect, which arises from attentional capture, is independent of task requirements, whereas the changing-state effect is specific to tasks that require serial processing, has been examined by comparing tasks that do or do not require serial-order processing. While many previous studies used the missing-item task as the nonserial task, it is unclear whether other cognitive tasks lead to similar results regarding the different task specificity of both effects. Kattner et al. (Memory and Cognition, 2023) used the mental-arithmetic task as the nonserial task, and failed to demonstrate the deviation effect. However, there were several procedural factors that could account for the lack of deviation effect, such as differences in design and procedures (e.g., conducted online, intermixed conditions). In the present study, we aimed to investigate whether the deviation effect could be observed in both the serial-recall and mental-arithmetic tasks when these procedural factors were modified. We found strong evidence of the deviation effect in both the serial-recall and the mental-arithmetic tasks when stimulus presentation and experimental design were aligned with previous studies that demonstrated the deviation effect (e.g., conducted in-person, blockwise presentation of sound, etc.). The results support the idea that the deviation effect is not task-specific.

Michelangelo Effect in Cognitive Rehabilitation: Using Art in a Digital Visuospatial Memory Task

The Michelangelo effect is a phenomenon that shows a reduction in perceived effort and an improvement in performance among both healthy subjects and patients when completing a motor task related to artistic stimuli, compared to performing the same task with non-artistic stimuli. It could contribute to the efficacy of art therapy in neurorehabilitation. In this study, the possible occurrence of this effect was tested in a cognitive task by asking 15 healthy subjects and 17 patients with a history of stroke to solve a digital version of the classical memory card game. Three different types of images were used in a randomized order: French cards, artistic portraits, and photos of famous people (to compensate for the possible effects of face recognition). Healthy subjects were involved to test the usability and the load demand of the developed system, reporting no statistically significant differences among the three sessions (p > 0.05). Conversely, patients had a better performance in terms of time (p = 0.014) and the number of attempts (p = 0.007) needed to complete the task in the presence of artistic stimuli, accompanied by a reduction in the perceived effort (p = 0.033). Furthermore, artistic stimuli, with respect to the other two types of images, seemed more associated with visuospatial control than linguistic functions.

Cigarette smoking and cognitive task performance: Experimental effects of very-low nicotine-content cigarettes.

Reductions in the nicotine content of cigarettes decrease smoking rate and dependence severity, but effects on cognition are less well established. The potential impacts of very-low nicotine-content (VLNC) cigarettes on cognitive task performance must be evaluated, especially in vulnerable populations. The aim of the present study is to experimentally examine the effects of VLNC cigarettes on cognitive performance. Adults who smoked daily (n = 775) from three vulnerable populations (socioeconomically disadvantaged reproductive-age women, individuals with opioid use disorder, affective disorders) were examined. Participants were randomly assigned to normal nicotine content (NNC; 15.8 mg nicotine/g tobacco) or VLNC (2.4 mg/g or 0.4 mg/g) cigarettes for 12 weeks. Response inhibition (stop-signal task), working memory (n-back task; n of 2-n of 0), and cognitive interference (nicotine Stroop task) were assessed at baseline, 2, 6, and 12 weeks. Results were analyzed using mixed-model repeated-measures analyses of variance. Extended exposure to VLNC cigarettes produced no significant changes in any measure of cognitive performance compared to NNC cigarettes. Over weeks, response times on the n-back task decreased across doses. No significant effects were observed on the stop-signal or nicotine Stroop tasks. All three vulnerable populations performed comparably on all three cognitive tasks. Extended exposure to VLNC cigarettes produced no impairments in cognitive performance on any of the assessed tasks compared to NNC cigarettes. These findings are consistent with the larger literature detailing other consequences following exposure to VLNC cigarettes and are encouraging for the adoption of a nicotine-reduction policy. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Firearm ownership factors and cognitive functioning: A preliminary study

Recent studies have suggested that owning a firearm for the purpose of protection and intending to acquire a firearm within the next year are associated with disrupted cognitive-affective processes that may increase firearm suicide risk. Prior studies were limited by self-report measures and would benefit from the utilization of objective indicators of cognitive functioning to clarify associations between firearm ownership and cognitive-affective states. The present study examined objective cognitive performance in a sample of U.S. adults in relation to firearm ownership variables. Based on prior studies, we hypothesized that protective owners and individuals with the intention to acquire a firearm within the next year would demonstrate reduced performance on decision-making, mental flexibility, and emotional bias tasks. A total of 112 participants (61% female, 79% white) completed objective cognitive tasks following completion of a larger survey study through ResearchMatch. Results indicated that intention to purchase a firearm within the next year was not associated with cognitive functioning but was associated with an array of heightened psychiatric symptoms. Owning a firearm for the purpose of protection was associated with decreased mental flexibility relative to non-owners, but no differences were seen for decision-making or emotional bias. These findings represent the first attempt to examine underlying objective cognitive processes related to firearm ownership, and implications for future research are discussed.

Reward Sensitivity and Noise Contribute to Negative Affective Bias: A Learning Signal Detection Theory Approach in Decision-Making

In patients with mood disorders, negative affective biases – systematically prioritising and interpreting information negatively – are common. A translational cognitive task testing this bias has shown that depressed patients have a reduced preference for a high reward under ambiguous decision-making conditions. The precise mechanisms underscoring this bias are, however, not yet understood. We therefore developed a set of measures to probe the underlying source of the behavioural bias by testing its relationship to a participant’s reward sensitivity, value sensitivity and reward learning rate. One-hundred-forty-eight participants completed three online behavioural tasks: the original ambiguous-cue decision-making task probing negative affective bias, a probabilistic reward learning task probing reward sensitivity and reward learning rate, and a gambling task probing value sensitivity. We modelled the learning task through a dynamic signal detection theory model and the gambling task through an expectation-maximisation prospect theory model. Reward sensitivity from the probabilistic reward task (β = 0.131, p = 0.024) and setting noise from the probabilistic reward task (β = –0.187, p = 0.028) both predicted the affective bias score in a logistic regression. Increased negative affective bias, at least on this specific task, may therefore be driven in part by a combination of reduced sensitivity to rewards and more variable responses.

Targeting the left DLPFC and right VLPFC in unmarried romantic relationship breakup (love trauma syndrome) with intensified electrical stimulation: A randomized, single-blind, parallel-group, sham-controlled study

BackgroundEnding a romantic relationship is one of the most painful losses an adult experience. Neuroimaging studies suggest that there is a neuropsychological link between breakup experiences and bereaved individuals, and that specific prefrontal regions are involved. The aim of this study was to determine whether enhancement of left DLPFC and right VLPFC activity with a novel intensified anodal transcranial direct current stimulation protocol reduces core symptoms of love trauma syndrome (LTS) and improves treatment-related variables. MethodsIn this randomized, sham-controlled, single-blind parallel trial, we assessed the efficacy of an intensified anodal stimulation protocol (20 min, twice-daily sessions with 20 min intervals, 5 consecutive days) with two montages (left DLPFC vs right VLPFC) to reduce love trauma symptoms. 36 participants with love trauma syndrome were randomized in three tDCS condition (left DLPFC, right VLPFC, sham stimulation). LTS symptoms, treatment-related outcome variables (depressive state, anxiety, emotion regulation, positive and negative affect), and cognitive functions were assessed before, right after, and one month after intervention. ResultsBoth DLPFC and VLPFC protocols significantly reduced LTS symptoms, and improved depressive state and anxiety after the intervention, as compared to the sham group. The improving effect of the DLPFC protocol on love trauma syndrome was significantly larger than that of the VLPFC protocol. For emotion regulation and positive and negative affect, improved regulation of emotions and positive affect and reduced negative affect were revealed after intervention in the two real stimulation conditions compared to the sham. For cognitive functions, no significant difference was observed between the groups, but again a positive effect of intervention within groups in the real stimulation conditions (DLPFC and VLPFC) was found for most components of the cognitive tasks. ConclusionsEnhancement of left DLPFC and right VLPFC activity with intensified stimulation improves LTS symptoms and treatment-related variables. For LTS symptoms, DLPFC stimulation was more efficient than VLPFC stimulation., For the other variables, no significant difference was observed between these two stimulation groups. These promising results require replication in larger trials.

An Investigation into the Utility of Large Language Models in Geotechnical Education and Problem Solving

The study explores the capabilities of large language models (LLMs), particularly GPT-4, in understanding and solving geotechnical problems, a specialised area that has not been extensively examined in previous research. Employing a question bank obtained from a commonly used textbook in geotechnical engineering, the research assesses GPT-4’s performance across various topics and cognitive complexity levels, utilising different prompting strategies like zero-shot learning, chain-of-thought (CoT) prompting, and custom instructional prompting. The study reveals that while GPT-4 demonstrates significant potential in addressing fundamental geotechnical concepts and problems, its effectiveness varies with specific topics, the complexity of the task, and the prompting strategies employed. The paper categorises errors encountered by GPT-4 into conceptual, grounding, calculation, and model inherent deficiencies related to the interpretation of visual information. Custom instructional prompts, specifically tailored to address GPT-4’s shortcomings, significantly enhance its performance. The study reveals that GPT-4 achieved an overall problem-solving accuracy of 67% with custom instructional prompting, significantly higher than the 28.9% with zero-shot learning and 34% with CoT. However, the study underscores the importance of human oversight in interpreting and verifying GPT-4’s outputs, especially in complex, higher-order cognitive tasks. The findings contribute to understanding the potential and limitations of current LLMs in specialised educational fields, providing insights for educators and researchers in integrating AI tools like GPT-4 into their teaching and problem-solving approaches. The study advocates for a balanced integration of AI in education to enrich educational delivery and experience while emphasising the indispensable role of human expertise alongside technological advancements.

Electroencephalogram Functional Connectivity Analysis and Classification of Mental Arithmetic Working Memory Task

Analyzing brain activity during mental arithmetic tasks provides insight into psychological disorders such as ADHD, dyscalculia, and autism. While most research is conducted on the static functional connectivity of the brain while performing a cognitive task, the dynamic changes of the brain, which provide meaningful information for diagnosing individual differences in cognitive tasks, are often ignored. This paper aims to classify electroencephalogram (EEG) signals for rest vs. mental arithmetic task performance, using Bayesian functional connectivity features in the sensor space as inputs into a graph convolutional network. The subject-specific (intrasubject) classification performed on 36 subjects for rest vs. mental arithmetic task performance achieved the highest subject-specific classification accuracy of 98% and an average accuracy of 91% in the beta frequency band, outperforming state-of-the-art methods. In addition, statistical analysis confirms the consistency of Bayesian functional connectivity features compared to traditional functional connectivity features. Furthermore, the graph-theoretical analysis of functional connectivity networks reveals that good-performance subjects had higher global efficiency, betweenness centrality, and closeness centrality than bad-performance subjects. The ablation study on the classification of three cognitive states (subtraction, music, and memory) achieved a classification accuracy of 97%, and visual working memory (n-back task) achieved a classification accuracy of 94%, confirming the consistency and reliability of the proposed methodology.

Sense of self in first-time pregnancy

Pregnancy is a time of profound upheaval of the self, when in addition to undergoing dramatic physical changes to accommodate a developing foetus, significant cognitive and social transformations occur in preparation for birth and parenthood. So far, research into cognitive constructions of the self has been either infant-centric or psychopathology-focused, so our understanding of the healthy, changing self in pregnancy remains relatively poor. This online experiment uses online questionnaires and two cognitive tasks to investigate how constructs relating to the mental self-model, including body representation, self-concept clarity, sense of agency, general self-efficacy and self-attribute learning, differ between first-time pregnant (n = 100) and never-been pregnant (n = 102) women. Results indicate that first-time pregnancy is associated with a significantly higher sense of body agency, body visibility, and body estrangement. Poorer accuracy for newly learned associations was also observed in the pregnant group. Whilst a typical self-processing bias was observed in both groups as expected, an intentional binding effect was absent. Notably, post-hoc exploratory analyses provide initial evidence for trimester effects, with a decisively higher self-reported sense of negative agency in the first trimester compared to the never-pregnant group and other trimesters. Further, body agency and self-efficacy were higher in the second-trimester group compared to the never-pregnant group, suggesting a period of relative recovery and consolidation of the new self. Taken together, our results suggest that aspects of self-representation and agency undergo significant shifts over the course of pregnancy and provide multiple exciting avenues for future research.

Nonlinear EEG Analysis During Motor and Cognitive Tasks in Patients With Long COVID: A Dynamic Systems Approach.

Introduction. Nonlinear EEG provides information about dynamic properties of the brain. This study aimed to compare nonlinear EEG parameters estimated from patients with Long COVID in different cognitive and motor tasks. Materials and Methods. This 12-month prospective cohort study included 83 patients with Long COVID: 53 symptomatic and 30 asymptomatic. Brain electrical activity was evaluated by EEG in 4 situations: (1) at rest, (2) during the Trail Making Test Part A (TMT-A), (3) during the TMT Part B (TMT-B), and (4) during a coordination task: the Box and Blocks Test (BBT). Nonlinear EEG parameters were estimated in the time domain (activity and complexity). Assessments were made at 0 to 3, 3 to 6, and 6 to 12 months after inclusion. Results. There was a decrease in activity and complexity during the TMT-A and TMT-B, and an increase of these parameters during the BBT in both groups. There was an increase in activity at rest and during the TMT-A in the COVID-19 group at 0 to 3 months compared to the control, an increase in activity in the TMT-B in the COVID-19 group at 3 to 6 months compared to the control, and reduced activity and complexity at rest and during the TMT-A at 6 to 12 months compared to the control. Conclusion. The tasks followed a pattern of increased activity and complexity in cognitive tasks, which decreased during the coordination task. It was also observed that an increase in activity at rest and during cognitive tasks in the early stages, and reduced activity and complexity at rest and during cognitive tasks in the late phases of Long COVID.

Edge-Community Entropy is a Novel Neural Correlate of Aging and Moderator of Fluid Cognition.

Decreased neuronal specificity of the brain in response to cognitive demands (i.e., neural dedifferentiation) has been implicated in age-related cognitive decline. Investigations into functional connectivity analogues of these processes have focused primarily on measuring segregation of nonoverlapping networks at rest. Here, we used an edge-centric network approach to derive entropy, a measure of specialization, from spatially overlapping communities during cognitive task fMRI. Using Human Connectome Project Lifespan data (713 participants, 36-100 years old, 55.7% female), we characterized a pattern of nodal despecialization differentially affecting the medial temporal lobe and limbic, visual, and subcortical systems. At the whole-brain level, global entropy moderated declines in fluid cognition across the lifespan and uniquely covaried with age when controlling for the network segregation metric modularity. Importantly, relationships between both metrics (entropy and modularity) and fluid cognition were age-dependent, although entropy's relationship with cognition was specific to older adults. These results suggest entropy is a potentially important metric for examining how neurological processes in aging affect functional specialization at the nodal, network, and whole-brain level.Significant Statement Many potential clinical applications of fMRI necessitate examining localized function to assess and guide behavioral, pharmacological, surgical, and neuromodulatory interventions. Edge-community entropy offers the mathematical and conceptual advantage of calculating the functional specialization of individual nodes within an overlapping and interdependent system of communities using whole-brain, temporally-informed data. We suggest entropy during cognitive tasks may be a more conceptually appropriate functional connectivity analogue of age-related neural dedifferentiation than network segregation metrics, and demonstrate entropy strongly correlates with age and moderates cognition. Notably, entropy was strongly related to age at the whole-brain level and more strongly related than participation coefficient at the nodal level, offering greater potential to interrogate how local age-related dysfunction influences distributed systems supporting cognition and behavior.

Structural, Functional, and Genetic Changes Surrounding Electrodes Implanted in the Brain.

ConspectusImplantable neurotechnology enables monitoring and stimulating of the brain signals responsible for performing cognitive, motor, and sensory tasks. Electrode arrays implanted in the brain are increasingly used in the clinic to treat a variety of sources of neurological diseases and injuries. However, the implantation of a foreign body typically initiates a tissue response characterized by physical disruption of vasculature and the neuropil as well as the initiation of inflammation and the induction of reactive glial states. Likewise, electrical stimulation can induce damage to the surrounding tissue depending on the intensity and waveform parameters of the applied stimulus. These phenomena, in turn, are likely influenced by the surface chemistry and characteristics of the materials employed, but further information is needed to effectively link the biological responses observed to specific aspects of device design. In order to inform improved design of implantable neurotechnology, we are investigating the basic science principles governing device-tissue integration. We are employing multiple techniques to characterize the structural, functional, and genetic changes that occur in the cells surrounding implanted electrodes. First, we have developed a new "device-in-slice" technique to capture chronically implanted electrodes within thick slices of live rat brain tissue for interrogation with single-cell electrophysiology and two-photon imaging techniques. Our data revealed several new observations of tissue remodeling surrounding devices: (a) there was significant disruption of dendritic arbors in neurons near implants, where losses were driven asymmetrically on the implant-facing side. (b) There was a significant loss of dendritic spine densities in neurons near implants, with a shift toward more immature (nonfunctional) morphologies. (c) There was a reduction in excitatory neurotransmission surrounding implants, as evidenced by a reduction in the frequency of excitatory postsynaptic currents (EPSCs). Lastly, (d) there were changes in the electrophysiological underpinnings of neuronal spiking regularity. In parallel, we initiated new studies to explore changes in gene expression surrounding devices through spatial transcriptomics, which we applied to both recording and stimulating arrays. We found that (a) device implantation is associated with the induction of hundreds of genes associated with neuroinflammation, glial reactivity, oligodendrocyte function, and cellular metabolism and (b) electrical stimulation induces gene expression associated with damage or plasticity in a manner dependent upon the intensity of the applied stimulus. We are currently developing computational analysis tools to distill biomarkers of device-tissue interactions from large transcriptomics data sets. These results improve the current understanding of the biological response to electrodes implanted in the brain while producing new biomarkers for benchmarking the effects of novel electrode designs on responses. As the next generation of neurotechnology is developed, it will be increasingly important to understand the influence of novel materials, surface chemistries, and implant architectures on device performance as well as the relationship with the induction of specific cellular signaling pathways.

Extended wakefulness alters the relationship between EEG oscillations and performance in a sustained attention task.

During drowsiness, maintaining consistent attention becomes difficult, leading to behavioural lapses. Bursts of oscillations in the electroencephalogram (EEG) might predict such lapses, given that alpha bursts increase during inattention and theta bursts increase with time spent awake. Paradoxically, however, alpha bursts decrease with time awake and theta bursts increase during focussed attention and cognitive tasks. Therefore, we investigated to what extent theta and alpha bursts predicted performance in a sustained attention task, either when well rested (baseline, BL) or following 20 h of extended wakefulness (EW). High-density EEG was measured in 18 young adults, and the timing of bursts was related to trial outcomes (fast, slow, and lapse trials). To increase the likelihood of lapses, the task was performed under soporific conditions. Against expectations, alpha bursts were more likely before fast trials and less likely before lapses at baseline, although the effect was substantially reduced during extended wakefulness. Theta bursts showed no significant relationship to behavioural outcome either at baseline or extended wakefulness. However, following exploratory analyses, we found that large-amplitude theta and alpha bursts were more likely to be followed by lapse trials during extended wakefulness but not baseline. In summary, alpha bursts during baseline anticipated better trial outcomes, whereas large-amplitude theta and alpha bursts during extended wakefulness anticipated worse outcomes. Therefore, neither theta nor alpha bursts maintain a consistent relationship with behaviour under different levels of overall vigilance.

Post-error adjustments occur in both reaching and grasping.

Post-error slowing (PES), the tendency to slow down a behavioral response after a previous error, has typically been investigated during simple cognitive tasks using response time as a measure of PES magnitude. More recently, PES was investigated during a single reach-to-grasp task to determine where post-error adjustments are employed in a more ecological setting. Kinematic analyses in the previous study detected PES during pre-movement planning and within the grasping component of movement execution. In the current study (N = 22), we increased the cognitive demands of a reach-to-grasp task by adding a choice between target and distractor locations to further explore PES, and other post-error adjustments, under different task conditions. We observed a significant main effect of task condition on overall reaction time (RT); however, it did not significantly impact PES or other post-error adjustments. Nonetheless, the results of this study suggest post-error adjustment is a flexible process that can be observed during pre-movement planning and within the onset and magnitude of the reaching component, as well as in the magnitudes of the grasping component. Considering the sum of the results in the context of existing literature, we conclude that the findings add support to a functional account of error reactivity, such that post-error adjustments are implemented intentionally to improve performance.

Data Pruning-enabled High Performance and Reliable Graph Neural Network Training on ReRAM-based Processing-in-Memory Accelerators

Graph Neural Networks (GNNs) have achieved remarkable accuracy in cognitive tasks such as predictive analytics on graph-structured data. Hence, they have become very popular in diverse real-world applications. However, GNN training with large real-world graph datasets in edge-computing scenarios is both memory- and compute-intensive. Traditional computing platforms such as CPUs and GPUs do not provide the energy efficiency and low latency required in edge intelligence applications due to their limited memory bandwidth. Resistive random-access memory (ReRAM)-based processing-in-memory (PIM) architectures have been proposed as suitable candidates for accelerating AI applications at the edge, including GNN training. However, ReRAM-based PIM architectures suffer from low reliability due to their limited endurance, and low performance when they are used for GNN training in real-world scenarios with large graphs. In this work, we propose a learning-for-data-pruning framework, which leverages a trained Binary Graph Classifier (BGC) to reduce the size of the input data graph by pruning subgraphs early in the training process to accelerate the GNN training process on ReRAM-based architectures. The proposed light-weight BGC model reduces the amount of redundant information in input graph(s) to speed up the overall training process, improves the reliability of the ReRAM-based PIM accelerator, and reduces the overall training cost. This enables fast, energy-efficient, and reliable GNN training on ReRAM-based architectures. Our experimental results demonstrate that using this learning for data pruning framework, we can accelerate GNN training and improve the reliability of ReRAM-based PIM architectures by up to 1.6 ×, and reduce the overall training cost by 100 × compared to state-of-the-art data pruning techniques.