Search

Reservoir computing (RC) has attracted significant interest as a framework for the implementation of novel neuromorphic computing architectures. Previously attention has been focussed on software-based reservoirs, where it has been demonstrated that reservoir topology plays a role in task performance, and functional advantage has been attributed to small-world and scale-free connectivity. However in hardware systems, such as electronic memristor networks, the mechanisms responsible for the reservoir dynamics are very different and the role of reservoir topology is largely unknown. Here we compare the performance of a range of memristive reservoirs in several RC tasks that are chosen to highlight different system requirements. We focus on percolating networks of nanoparticles (PNNs) which are novel self-assembled nanoscale systems that exhibit scale-free and small-world properties. We find that the performance of regular arrays of uniform memristive elements is limited by their symmetry but that this symmetry can be broken either by a heterogeneous distribution of memristor properties or a scale-free topology. The best perfomance across all tasks is observed for a scale-free network with uniform memistor properties. These results provide insight into the role of topology in neuromorphic reservoirs as well as an overview of the computational performance of scale-free networks of memristors in a range of benchmark tasks.

The connectivity of self-assembled networks of nanowires and nanoparticles is believed to strongly influence their performance in brainlike (neuromorphic) computing applications. Here we present a new perspective on the connectivity of these networks in which their neuronlike active elements are viewed in the same way as the nodes in artificial and biological neuronal networks. We consider two-dimensional and quasi-three-dimensional networks of nanowires and percolating networks of nanoparticles and show that, from this new perspective, they all have similar small-world characteristics. Other characteristics which may impact the computational performance of the networks are also investigated, including their assortativity and the scalefree nature of the nanoparticle networks. Taken together, these results allow comparison of key network characteristics for a variety of self-assembled nanoscale networks, and provide a basis for detailed investigations of computational performance.

The biological brain is a highly efficient computational system in which information processing is performed via electrical spikes. Neuromorphic computing systems that work on similar principles could support the development of the next generation of artificial intelligence and, in particular, enable low-power edge computing. Percolating networks of nanoparticles (PNNs) have previously been shown to exhibit critical spiking behavior, with promise for highly efficient natural computation. Here we employ a rate coding scheme to show that PNNs can perform Boolean operations and image classification. Near perfect accuracy is achieved in both tasks by manipulating the spiking activity using certain control voltages. We demonstrate that the key to successful computation is that nanoscale tunnel gaps within the percolating networks transform input data through a powerful modulus-like nonlinearity. These results provide a basis for implementation of further computational schemes that exploit the brain-like criticality of these networks.

SARS-CoV-2 infection causes diverse outcomes ranging from asymptomatic infection to respiratory distress and death. A major unresolved question is whether prior immunity to endemic, human common cold coronaviruses (hCCCoVs) impacts susceptibility to SARS-CoV-2 infection or immunity following infection and vaccination. Therefore, we analyzed samples from the same individuals before and after SARS-CoV-2 infection or vaccination. We found hCCCoV antibody levels increase after SARS-CoV-2 exposure, demonstrating cross-reactivity. However, a case-control study indicates that baseline hCCCoV antibody levels are not associated with protection against SARS-CoV-2 infection. Rather, higher magnitudes of pre-existing betacoronavirus antibodies correlate with more SARS-CoV-2 antibodies following infection, an indicator of greater disease severity. Additionally, immunization with hCCCoV spike proteins before SARS-CoV-2 immunization impedes the generation of SARS-CoV-2-neutralizing antibodies in mice. Together, these data suggest that pre-existing hCCCoV antibodies hinder SARS-CoV-2 antibody-based immunity following infection and provide insight on how pre-existing coronavirus immunity impacts SARS-CoV-2 infection, which is critical considering emerging variants.

Networks of nanowires are currently being explored for a range of applications in brain-like (or neuromorphic) computing, and especially in reservoir computing (RC). Fabrication of real-world computing devices requires that the nanowires are deposited sequentially, leading to stacking of the wires on top of each other. However, most simulations of computational tasks using these systems treat the nanowires as 1D objects lying in a perfectly 2D plane - the effect of stacking on RC performance has not yet been established. Here we use detailed simulations to compare the performance of perfectly 2D and quasi-3D (stacked) networks of nanowires in two tasks: memory capacity and nonlinear transformation. We also show that our model of the junctions between nanowires is general enough to describe a wide range of memristive networks, and consider the impact of physically realistic electrode configurations on performance. We show that the various networks and configurations have a strikingly similar performance in RC tasks, which is surprising given their radically different topologies. Our results show that networks with an experimentally achievable number of electrodes perform close to the upper bounds achievable when using the information from every wire. However, we also show important differences, in particular that the quasi-3D networks are more resilient to changes in the input parameters, generalizing better to noisy training data. Since previous literature suggests that topology plays an important role in computing performance, these results may have important implications for future applications of nanowire networks in neuromorphic computing.

Exogenous causes, such as alcohol use, and endogenous factors, such as temperament and sex, can modulate developmental trajectories of adolescent neurofunctional maturation. We examined how these factors affect sexual dimorphism in brain functional networks in youth drinking below diagnostic threshold for alcohol use disorder (AUD). Based on the 3-year, annually acquired, longitudinal resting-state functional magnetic resonance imaging (MRI) data of 526 adolescents (12-21 years at baseline) from the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) cohort, developmental trajectories of 23 intrinsic functional networks (IFNs) were analyzed for (1) sexual dimorphism in 259 participants who were no-to-low drinkers throughout this period; (2) sex-alcohol interactions in two age- and sex-matched NCANDA subgroups (N = 76 each), half no-to-low, and half moderate-to-heavy drinkers; and (3) moderating effects of gender-specific alcohol dose effects and a multifactorial impulsivity measure on IFN connectivity in all NCANDA participants. Results showed that sex differences in no-to-low drinkers diminished with age in the inferior-occipital network, yet girls had weaker within-network connectivity than boys in six other networks. Effects of adolescent alcohol use were more pronounced in girls than boys in three IFNs. In particular, girls showed greater within-network connectivity in two motor networks with more alcohol consumption, and these effects were mediated by sensation-seeking only in girls. Our results implied that drinking might attenuate the naturally diminishing sexual differences by disrupting the maturation of network efficiency more severely in girls. The sex-alcohol-dose effect might explain why women are at higher risk of alcohol-related health and psychosocial consequences than men.

There is currently a great deal of interest in the use of nanoscale devices to emulate the behaviors of neurons and synapses and to facilitate brain-inspired computation. Here, it is shown that percolating networks of nanoparticles exhibit stochastic spiking behavior that is strikingly similar to that observed in biological neurons. The spiking rate can be controlled by the input stimulus, similar to "rate coding" in biology, and the distributions of times between events are log-normal, providing insights into the atomic-scale spiking mechanism. The stochasticity of the spiking behavior is then used for true random number generation, and the high quality of the generated random bit-streams is demonstrated, opening up promising routes toward integration of neuromorphic computing with secure information processing.

BACKGROUND: The cerebellum is a target of alcoholism-related brain damage in adults, yet no study has prospectively tracked deviations from normal cerebellar growth trajectories in adolescents before and after initiating drinking. METHODS: Magnetic resonance imaging tracked developmental volume trajectories of 10 cerebellar lobule and vermis tissue constituents in 548 no/low drinking youths age 12 to 21 years at induction into this 5-site, NCANDA (National Consortium on Alcohol and NeuroDevelopment in Adolescence) study. Over the 3- to 4-year longitudinal examination yielding 2043 magnetic resonance imaging scans, 328 youths remained no/low drinkers, whereas 220 initiated substantial drinking after initial neuroimaging. RESULTS: Normal growth trajectories derived from no/low drinkers indicated that gray matter volumes of lobules V and VI, crus II, lobule VIIB, and lobule X declined faster with age in male youths than in female youths, whereas white matter volumes in crus I and crus II and lobules VIIIA and VIIIB expanded faster in female youths than in male youths; cerebrospinal fluid volume expanded faster in most cerebellar regions of male youths than female youths. Drinkers exhibited accelerated gray matter decline in anterior lobules and vermis, accelerated vermian white matter expansion, and accelerated cerebrospinal fluid volumes expansion of anterior lobules relative to youths who remained no/low drinkers. Analyses including both alcohol and marijuana did not support an independent role for marijuana in alcohol effects on cerebellar gray matter trajectories. CONCLUSIONS: Alcohol use-related cerebellar growth trajectory differences from normal involved anterior lobules and vermis of youths who initiated substantial drinking. These regions are commonly affected in alcohol-dependent adults, raising the possibility that cerebellar structures affected by youthful drinking may be vulnerable to age-alcohol interactions in later adulthood.

Self-assembled networks of nanoparticles and nanowires have recently emerged as promising systems for brain-like computation. Here, we focus on percolating networks of nanoparticles which exhibit brain-like dynamics. We use a combination of experiments and simulations to show that the brain-like network dynamics emerge from atomic-scale switching dynamics inside tunnel gaps that are distributed throughout the network. The atomic-scale dynamics emulate leaky integrate and fire (LIF) mechanisms in biological neurons, leading to the generation of critical avalanches of signals. These avalanches are quantitatively the same as those observed in cortical tissue and are signatures of the correlations that are required for computation. We show that the avalanches are associated with dynamical restructuring of the networks which self-tune to balanced states consistent with self-organized criticality. Our simulations allow visualization of the network states and detailed mechanisms of signal propagation.

© 2020 Massachusetts Institute of Technology. Biological neuronal networks are the computing engines of the mammalian brain. These networks exhibit structural characteristics such as hierarchical architectures, small-world attributes, and scale-free topologies, providing the basis for the emergence of rich temporal characteristics such as scale-free dynamics and long-range temporal correlations. Devices that have both the topological and the temporal features of a neuronal network would be a significant step toward constructing a neuromorphic system that can emulate the computational ability and energy efficiency of the human brain. Here we use numerical simulations to show that percolating networks of nanoparticles exhibit structural properties that are reminiscent of biological neuronal networks, and then show experimentally that stimulation of percolating networks by an external voltage stimulus produces temporal dynamics that are self-similar, follow power-law scaling, and exhibit long-range temporal correlations. These results are expected to have important implications for the development of neuromorphic devices, especially for those based on the concept of reservoir computing.

Current influenza A virus (IAV) vaccines stimulate antibody responses that are directed against variable regions of the virus, and are therefore ineffective against divergent strains. As CD8+ T cells target the highly conserved, internal IAV proteins, they have the potential to increase heterosubtypic immunity. Early T-cell priming events influence lasting memory, which is required for long-term protection. However, the early responding, IAV-specific cells are difficult to monitor because of their low frequencies. Here, we tracked the dissemination of endogenous IAV-specific CD8+ T cells during the initial phases of the immune response following IAV infection. We exposed a significant population of recently activated, CD25+ CD43+ IAV-specific T cells that were not detected by tetramer staining. By tracking this population, we found that initial T-cell priming occurred in the mediastinal lymph nodes, which gave rise to the most expansive IAV-specific CD8+ T-cell population. Subsequently, IAV-specific CD8+ T cells dispersed to the bronchoalveolar lavage and blood, followed by spleen and liver, and finally to the lung. These data provide important insight into the priming and tissue dispersion of an endogenous CD8+ T-cell response. Importantly, the CD25+ CD43+ phenotype identifies an inclusive population of early responding CD8+ T cells, which may provide insight into TCR repertoire selection and expansion. A better understanding of this response is critical for designing improved vaccines that target CD8+ T cells.

Longitudinal study provides a robust method for tracking developmental trajectories. Yet inherent problems of retesting pose challenges in distinguishing biological developmental change from prior testing experience. We examined factors potentially influencing change scores on 16 neuropsychological test composites over 1year in 568 adolescents in the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) project. The twice-minus-once-tested method revealed that performance gain was mainly attributable to testing experience (practice) with little contribution from predicted developmental effects. Group mean practice slopes for 13 composites indicated that 60% to ∼100% variance was attributable to test experience; General Ability accuracy showed the least practice effect (29%). Lower baseline performance, especially in younger participants, was a strong predictor of greater gain. Contributions from age, sex, ethnicity, examination site, socioeconomic status, or family history of alcohol/substance abuse were nil to small, even where statistically significant. Recognizing that a substantial proportion of change in longitudinal testing, even over 1-year, is attributable to testing experience indicates caution against assuming that performance gain observed during periods of maturation necessarily reflects development. Estimates of testing experience, a form of learning, may be a relevant metric for detecting interim influences, such as alcohol use or traumatic episodes, on behavior.

During adolescence, problems reflecting cognitive, behavioral and affective dysregulation, such as inattention and emotional dyscontrol, have been observed to be associated with substance use disorder (SUD) risks and outcomes. Prior studies have typically been with small samples, and have typically not included comprehensive measurement of executive dysfunction domains. The relationships of executive dysfunction in daily life with performance based testing of cognitive skills and structural brain characteristics, thought to be the basis for executive functioning, have not been definitively determined. The aims of this study were to determine the relationships between executive dysfunction in daily life, measured by the Behavior Rating Inventory of Executive Function (BRIEF), cognitive skills and structural brain characteristics, and SUD risks, including a global SUD risk indicator, sleep quality, and risky alcohol and cannabis use. In addition to bivariate relationships, multivariate models were tested. The subjects (n = 817; ages 12 through 21) were participants in the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) study. The results indicated that executive dysfunction was significantly related to SUD risks, poor sleep quality, risky alcohol use and cannabis use, and was not significantly related to cognitive skills or structural brain characteristics. In multivariate models, the relationship between poor sleep quality and risky substance use was mediated by executive dysfunction. While these cross-sectional relationships need to be further examined in longitudinal analyses, the results suggest that poor sleep quality and executive dysfunction may be viable preventive intervention targets to reduce adolescent substance use.

BACKGROUND: Abundant cross-sectional evidence links eveningness (a preference for later sleep-wake timing) and increased alcohol and drug use among adolescents and young adults. However, longitudinal studies are needed to examine whether eveningness is a risk factor for subsequent alcohol and drug use, particularly during adolescence, which is marked by parallel peaks in eveningness and risk for the onset of alcohol use disorders. This study examined whether eveningness and other sleep characteristics were associated with concurrent or subsequent substance involvement in a longitudinal study of adolescents. METHODS: Participants were 729 adolescents (368 females; age 12 to 21 years) in the National Consortium on Alcohol and Neurodevelopment in Adolescence study. Associations between the sleep variables (circadian preference, sleep quality, daytime sleepiness, sleep timing, and sleep duration) and 3 categorical substance variables (at-risk alcohol use, alcohol bingeing, and past-year marijuana use [y/n]) were examined using ordinal and logistic regression with baseline age, sex, race, ethnicity, socioeconomic status, and psychiatric problems as covariates. RESULTS: At baseline, greater eveningness was associated with greater at-risk alcohol use, greater bingeing, and past-year use of marijuana. Later weekday and weekend bedtimes, but not weekday or weekend sleep duration, showed similar associations across the 3 substance outcomes at baseline. Greater baseline eveningness was also prospectively associated with greater bingeing and past-year use of marijuana at the 1-year follow-up, after covarying for baseline bingeing and marijuana use. Later baseline weekday and weekend bedtimes, and shorter baseline weekday sleep duration, were similarly associated with greater bingeing and past-year use of marijuana at the 1-year follow-up after covarying for baseline values. CONCLUSIONS: Findings suggest that eveningness and sleep timing may be under recognized risk factors and