Inside out: transforming images of lab-grown plants for machine learning applications in agriculture.

Introduction: Machine learning tasks often require a significant amount of training data for the resultant network to perform suitably for a given problem in any domain. In agriculture, dataset sizes are further limited by phenotypical differences between two plants of the same genotype, often as a result of different growing conditions. Synthetically-augmented datasets have shown promise in improving existing models when real data is not available. Methods: In this paper, we employ a contrastive unpaired translation (CUT) generative adversarial network (GAN) and simple image processing techniques to translate indoor plant images to appear as field images. While we train our network to translate an image containing only a single plant, we show that our method is easily extendable to produce multiple-plant field images. Results: Furthermore, we use our synthetic multi-plant images to train several YoloV5 nano object detection models to perform the task of plant detection and measure the accuracy of the model on real field data images. Discussion: The inclusion of training data generated by the CUT-GAN leads to better plant detection performance compared to a network trained solely on real data.

Leveraging Guided Backpropagation to Select Convolutional Neural Networks for Plant Classification.

The development of state-of-the-art convolutional neural networks (CNN) has allowed researchers to perform plant classification tasks previously thought impossible and rely on human judgment. Researchers often develop complex CNN models to achieve better performances, introducing over-parameterization and forcing the model to overfit on a training dataset. The most popular process for evaluating overfitting in a deep learning model is using accuracy and loss curves. Train and loss curves may help understand the performance of a model but do not provide guidance on how the model could be modified to attain better performance. In this article, we analyzed the relation between the features learned by a model and its capacity and showed that a model with higher representational capacity might learn many subtle features that may negatively affect its performance. Next, we showed that the shallow layers of a deep learning model learn more diverse features than the ones learned by the deeper layers. Finally, we propose SSIM cut curve, a new way to select the depth of a CNN model by using the pairwise similarity matrix between the visualization of the features learned at different depths by using Guided Backpropagation. We showed that our proposed method could potentially pave a new way to select a better CNN model.

Ecology and public health burden of Keystone virus in Florida.

Keystone virus (KEYV) is an under-studied orthobunyavirus that is transmitted via both horizontal and vertical cycles involving various mosquito species and vertebrate hosts. Historical evidence indicates that KEYV causes sub-clinical infections in humans, and some case studies draw links between this virus and encephalitis. Given KEYV's potential to cause human infections, it is plausible that it causes an under-appreciated proportion of both generic viral infections and unidentified viral encephalitis cases. This review summarizes current knowledge of KEYV and its disease dynamics in order to better understand the virus' medical and economic burden on human populations.

An embedded system for the automated generation of labeled plant images to enable machine learning applications in agriculture.

A lack of sufficient training data, both in terms of variety and quantity, is often the bottleneck in the development of machine learning (ML) applications in any domain. For agricultural applications, ML-based models designed to perform tasks such as autonomous plant classification will typically be coupled to just one or perhaps a few plant species. As a consequence, each crop-specific task is very likely to require its own specialized training data, and the question of how to serve this need for data now often overshadows the more routine exercise of actually training such models. To tackle this problem, we have developed an embedded robotic system to automatically generate and label large datasets of plant images for ML applications in agriculture. The system can image plants from virtually any angle, thereby ensuring a wide variety of data; and with an imaging rate of up to one image per second, it can produce lableled datasets on the scale of thousands to tens of thousands of images per day. As such, this system offers an important alternative to time- and cost-intensive methods of manual generation and labeling. Furthermore, the use of a uniform background made of blue keying fabric enables additional image processing techniques such as background replacement and image segementation. It also helps in the training process, essentially forcing the model to focus on the plant features and eliminating random correlations. To demonstrate the capabilities of our system, we generated a dataset of over 34,000 labeled images, with which we trained an ML-model to distinguish grasses from non-grasses in test data from a variety of sources. We now plan to generate much larger datasets of Canadian crop plants and weeds that will be made publicly available in the hope of further enabling ML applications in the agriculture sector.

Strong influence of behavioral dynamics on the ability of testing and treating HIV to stop transmission.

Choosing between strategies to control HIV transmission with antivirals requires understanding both the dynamics affecting those strategies' effectiveness and what causes those dynamics. Alternating episodes of high and low contact rates (episodic risk) interact with increased transmission probabilities during early infection to strongly influence HIV transmission dynamics. To elucidate the mechanics of this interaction and how these alter the effectiveness of universal test and treat (UT8T) strategies, we formulated a model of UT8T effects. Analysis of this model shows how and why changing the dynamics of episodic risk changes the fraction of early transmissions (FET) and the basic reproduction number (R0) and consequently causes UT8T to vary from easily eliminating transmission to having little effect. As the length of risk episodes varies from days to lifetimes, FET first increases, then falls. Endemic prevalence varies similarly. R0, in contrast, increases monotonically and is the major determinant of UT8T effects. At some levels of episodic risk, FET can be high, but eradication is easy because R0 is low. At others FET is lower, but a high R0 makes eradication impossible and control ineffective. Thus changes in individual risk over time must be measured and analyzed to plan effective control strategies with antivirals.

Tumor growth increases neuroinflammation, fatigue and depressive-like behavior prior to alterations in muscle function.

Cancer patients frequently suffer from fatigue, a complex syndrome associated with loss of muscle mass, weakness, and depressed mood. Cancer-related fatigue (CRF) can be present at the time of diagnosis, during treatment, and persists for years after treatment. CRF negatively influences quality of life, limits functional independence, and is associated with decreased survival in patients with incurable disease. Currently there are no effective treatments to reduce CRF. The aim of this study was to use a mouse model of tumor growth and discriminate between two main components of fatigue: loss of muscle mass/function and altered mood/motivation. Here we show that tumor growth increased fatigue- and depressive-like behaviors, and reduced body and muscle mass. Decreased voluntary wheel running activity (VWRA) and increased depressive-like behavior in the forced swim and sucrose preference tests were evident in tumor-bearing mice within the first two weeks of tumor growth and preceded the loss of body and muscle mass. At three weeks, tumor-bearing mice had reduced grip strength but this was not associated with altered expression of myosin isoforms or impaired contractile properties of muscles. These increases in fatigue and depressive-like behaviors were paralleled by increased expression of IL-1ß mRNA in the cortex and hippocampus. Minocycline administration reduced tumor-induced expression of IL-1ß in the brain, reduced depressive-like behavior, and improved grip strength without altering muscle mass. Taken together, these results indicate that neuroinflammation and depressed mood, rather than muscle wasting, contribute to decreased voluntary activity and precede major changes in muscle contractile properties with tumor growth.

Successes and shortcomings of polio eradication: a transmission modeling analysis.

Polio eradication is on the cusp of success, with only a few regions still maintaining transmission. Improving our understanding of why some regions have been successful and others have not will help with both global eradication of polio and development of more effective vaccination strategies for other pathogens. To examine the past 25 years of eradication efforts, we constructed a transmission model for wild poliovirus that incorporates waning immunity (which affects both infection risk and transmissibility of any resulting infection), age-mediated vaccination rates, and transmission of oral polio vaccine. The model produces results consistent with the 4 country categories defined by the Global Polio Eradication Program: elimination with no subsequent outbreaks; elimination with subsequent transient outbreaks; elimination with subsequent outbreaks and transmission detected for more than 12 months; and endemic polio transmission. Analysis of waning immunity rates and oral polio vaccine transmissibility reveals that higher waning immunity rates make eradication more difficult because of increasing numbers of infectious adults, and that higher oral polio vaccine transmission rates make eradication easier as adults become reimmunized. Given these dynamic properties, attention should be given to intervention strategies that complement childhood vaccination. For example, improvement in sanitation can reduce the reproduction number in problematic regions, and adult vaccination can lower adult transmission.

Episodic HIV Risk Behavior Can Greatly Amplify HIV Prevalence and the Fraction of Transmissions from Acute HIV Infection.

A deterministic compartmental model was explored that relaxed the unrealistic assumption in most HIV transmission models that behaviors of individuals are constant over time. A simple model was formulated to better explain the effects observed. Individuals had a high and a low contact rate and went back and forth between them. This episodic risk behavior interacted with the short period of high transmissibility during acute HIV infection to cause dramatic increases in prevalence as the differences between high and low contact rates increased and as the duration of high risk better matched the duration of acute HIV infection. These same changes caused a considerable increase in the fraction of all transmissions that occurred during acute infection. These strong changes occurred despite a constant total number of contacts and a constant total transmission potential from acute infection. Two phenomena played a strong role in generating these effects. First, people were infected more often during their high contact rate phase and they remained with high contact rates during the highly contagious acute infection stage. Second, when individuals with previously low contact rates moved into an episodic high-risk period, they were more likely to be susceptible and thus provided more high contact rate susceptible individuals who could get infected. These phenomena make test and treat control strategies less effective and could cause some behavioral interventions to increase transmission. Signature effects on genetic patterns between HIV strains could make it possible to determine whether these episodic risk effects are acting in a population.

Lipopolysaccharide-induced interleukin (IL)-4 receptor-α expression and corresponding sensitivity to the M2 promoting effects of IL-4 are impaired in microglia of aged mice.

In several models of aging, microglia become more inflammatory and reactive to immune challenges. For example, peripheral LPS injection causes exaggerated microglial activation associated with prolonged sickness and depressive-like behavior in aged BALB/c mice. Therefore, the purpose of this study was to determine the extent to which age-related amplified microglial activation was associated with reduced sensitivity to the anti-inflammatory and M2 promoting cytokines interleukin (IL)-10 and IL-4. In initial studies with adult mice, LPS induced a time-dependent increase in M1 and M2 mRNA profiles in microglia. Furthermore, peripheral LPS injection markedly increased surface expression of IL-4 receptor-alpha (IL-4Rα), but not IL-10 receptor-1 (IL-10R1) on microglia. In BV-2 cells, IL-4, but not IL-10, re-directed LPS-activated microglia towards an M2 phenotype. Based on these findings, comparisons of M1 and M2 activation profiles, induction of IL-4Rα, and sensitivity to IL-4 were determined in microglia from adult (3-4 mo) and aged (18-22 mo) mice. In aged microglia, LPS promoted an exaggerated and prolonged M1 and M2 profile compared to adults. Moreover, IL-4Rα protein was not increased on aged microglia following LPS injection. To determine the consequence of impaired IL-4Rα upregulation, adult and aged mice were injected with LPS and activated microglia were then isolated and treated ex vivo with IL-4. While ex vivo IL-4 induced an M2 profile in activated microglia from adult mice, activated microglia from aged mice retained a prominent M1 profile. These data indicate that activated microglia from aged mice are less sensitive to the anti-inflammatory and M2-promoting effects of IL-4.

Protracted downregulation of CX3CR1 on microglia of aged mice after lipopolysaccharide challenge.

Fractalkine (CX(3)CL1) to fractalkine receptor (CX(3)CR1) interactions in the brain are involved in the modulation of microglial activation. Our recent findings indicate that there is microglial hyperactivity in the aged brain during an inflammatory challenge. The underlying cause of this amplified microglial response in the aged brain is unknown. Therefore, the purpose of this study was to determine the degree to which age-associated impairments of CX(3)CL1 and CX(3)CR1 in the brain contribute to exaggerated microglial activation after intraperitoneal (i.p.) injection of lipopolysaccharide (LPS). Here we show that CX(3)CL1 protein was reduced in the brain of aged (18-22 mo) BALB/c mice compared to adult (3-6 mo) controls. CX(3)CL1 protein, however, was unaltered by LPS injection. Next, CX(3)CR1 levels were determined in microglia (CD11b(+)/CD45(low)) isolated by Percoll density gradient separation at 4 and 24h after LPS injection. Flow cytometric and mRNA analyses of these microglia showed that LPS injection caused a marked decrease of CX(3)CR1 and a simultaneous increase of IL-1ß at 4h after LPS injection. While surface expression of CX(3)CR1 was enhanced on microglia of adult mice by 24h, it was still significantly downregulated on a subset of microglia from aged mice. This protracted reduction of CX(3)CR1 corresponded with a delayed recovery from sickness behavior, prolonged IL-1ß induction, and decreased TGFß expression in the aged brain. In the last set of studies BV2 microglia were used to determine effect of TGFß on CX(3)CR1. These results showed that TGFß enhanced CX(3)CR1 expression and attenuated the LPS-induced increase in IL-1ß expression.

Peripheral lipopolysaccharide (LPS) challenge promotes microglial hyperactivity in aged mice that is associated with exaggerated induction of both pro-inflammatory IL-1beta and anti-inflammatory IL-10 cytokines.

In the elderly, systemic infection is associated with an increased frequency of behavioral and cognitive complications. We have reported that peripheral stimulation of the innate immune system with lipopolysaccharide (LPS) causes an exaggerated neuroinflammatory response and prolonged sickness/depressive-like behaviors in aged BALB/c mice. Therefore, the purpose of this study was to determine the degree to which LPS-induced neuroinflammation was associated with microglia-specific induction of neuroinflammatory mediators. Here, we show that peripheral LPS challenge caused a hyperactive microglial response in the aged brain associated with higher induction of inflammatory IL-1beta and anti-inflammatory IL-10. LPS injection caused a marked induction of mRNA expression of both IL-1beta and IL-10 in the cortex of aged mice compared to adults. In the next set of studies, microglia (CD11b(+)/CD45(low)) were isolated from the brain of adult and aged mice following experimental treatments. An age-dependent increase in major histocompatibility complex (MHC) class II mRNA and protein expression was detected in microglia. Moreover, peripheral LPS injection caused a more pronounced increase in IL-1beta, IL-10, Toll-like receptor (TLR)-2, and indoleamine 2,3-dioxygenase (IDO) mRNA levels in microglia isolated from aged mice than adults. Intracellular cytokine protein detection confirmed that peripheral LPS caused the highest increase in IL-1beta and IL-10 levels in microglia of aged mice. Finally, the most prominent induction of IL-1beta was detected in MHC II(+) microglia from aged mice. Taken together, these findings provide novel evidence that age-associated priming of microglia plays a central role in exaggerated neuroinflammation induced by activation of the peripheral innate immune system.

Immune and behavioral consequences of microglial reactivity in the aged brain.

Bidirectional communication between the immune system and the brain is essential for mounting the appropriate immunological, physiological, and behavioral responses to immune activation. Aging, however, may impair this important bi-directional interaction. In support of this notion, peripheral infection in the elderly is associated with an increased frequency of behavioral and cognitive complications. Recent findings in animal models of aging and neurodegenerative disease indicate that microglia, innate immune cells of the brain, become primed or reactive. Understanding age- and disease-associated alterations in microglia is important because glia (microglia and astrocytes) play an integral role in propagating inflammatory signals that are initiated in the periphery. In this capacity, brain glia produce inflammatory cytokines that target neuronal substrates and elicit a sickness-behavior syndrome that is normally beneficial to the host organism. Increased reactivity of microglia sets the stage for an exaggerated neuroinflammatory cytokine response following activation of the peripheral innate immune system, which may underlie subsequent long-lasting behavioral and cognitive deficits. In support of this premise, recent findings indicate that stimulation of the peripheral immune system in aged rodents causes exaggerated neuroinflammation that is paralleled by cognitive impairment, prolonged sickness, and depressive-like complications. Therefore, the purpose of this review is to discuss the new evidence that age-associated priming of microglia could play a pathophysiological role in exaggerated behavioral and cognitive sequelae to peripheral infection.

Minocycline attenuates lipopolysaccharide (LPS)-induced neuroinflammation, sickness behavior, and anhedonia.

BACKGROUND: Activation of the peripheral innate immune system stimulates the secretion of CNS cytokines that modulate the behavioral symptoms of sickness. Excessive production of cytokines by microglia, however, may cause long-lasting behavioral and cognitive complications. The purpose of this study was to determine if minocycline, an anti-inflammatory agent and purported microglial inhibitor, attenuates lipopolysaccharide (LPS)-induced neuroinflammation, sickness behavior, and anhedonia. METHODS: In the first set of experiments the effect of minocycline pretreatment on LPS-induced microglia activation was assessed in BV-2 microglia cell cultures. In the second study, adult (3-6 m) BALB/c mice received an intraperitoneal (i.p.) injection of vehicle or minocycline (50 mg/kg) for three consecutive days. On the third day, mice were also injected (i.p.) with saline or Escherichia coli LPS (0.33 mg/kg) and behavior (i.e., sickness and anhedonia) and markers of neuroinflammation (i.e., microglia activation and inflammatory cytokines) were determined. In the final study, adult and aged BALB/c mice were treated with the same minocycline and LPS injection regimen and markers of neuroinflammation were determined. All data were analyzed using Statistical Analysis Systems General Linear Model procedures and were subjected to one-, two-, or three-way ANOVA to determine significant main effects and interactions. RESULTS: Minocycline blocked LPS-stimulated inflammatory cytokine secretion in the BV-2 microglia-derived cell line and reduced LPS-induced Toll-like-receptor-2 (TLR2) surface expression on brain microglia. Moreover, minocycline facilitated the recovery from sickness behavior (i.e., anorexia, weight loss, and social withdrawal) and prevented anhedonia in adult mice challenged with LPS. Furthermore, the minocycline associated recovery from LPS-induced sickness behavior was paralleled by reduced mRNA levels of Interleukin (IL)-1beta, IL-6, and indoleamine 2, 3 dioxygenase (IDO) in the cortex and hippocampus. Finally, in aged mice, where exaggerated neuroinflammation was elicited by LPS, minocycline pretreatment was still effective in markedly reducing mRNA levels of IL-1beta, TLR2 and IDO in the hippocampus. CONCLUSION: These data indicate that minocycline mitigates neuroinflammation in the adult and aged brain and modulates the cytokine-associated changes in motivation and behavior.