Deep Learning-Based Eye-Tracking Analysis for Diagnosis of Alzheimer's Disease Using 3D Comprehensive Visual Stimuli.

Alzheimer's Disease (AD) is a neurodegenerative disorder that causes a continuous decline in cognitive functions and eventually results in death. An early AD diagnosis is important for taking active measures to slow its deterioration. Traditional diagnoses are usually based on clinical experience, which is limited by several realistic factors. In this paper, we focus on exploiting deep learning techniques to diagnose AD based on eye-tracking behaviors. Visual attention, as a typical eye-tracking behavior, is of great clinical value in detecting cognitive abnormalities in AD patients. To better analyze the differences in visual attention between AD patients and normals, we first conducted a 3D comprehensive visual task on a noninvasive eye-tracking system to collect visual attention heatmaps. Then a multilayered comparison convolutional neural network (MC-CNN) is proposed to distinguish the visual attention differences between AD patients and normals. In MC-CNN, the multilayered feature representations of heatmaps were obtained by hierarchical residual blocks to better encode eye movement behaviors, which were further integrated into a distance vector to benefit the comprehensive visual task. From evaluation, MC-CNN can distinguish AD patients from normals with 0.84 accuracy, 0.86 recall, 0.82 precision, 0.83 F1-score and 0.90 area under the curve (AUC). The above results demonstrate the effectiveness of the proposed MC-CNN in AD diagnosis based on the comprehensive 3D visual task.

A novel deep learning approach for diagnosing Alzheimer's disease based on eye-tracking data.

Eye-tracking technology has become a powerful tool for biomedical-related applications due to its simplicity of operation and low requirements on patient language skills. This study aims to use the machine-learning models and deep-learning networks to identify key features of eye movements in Alzheimer's Disease (AD) under specific visual tasks, thereby facilitating computer-aided diagnosis of AD. Firstly, a three-dimensional (3D) visuospatial memory task is designed to provide participants with visual stimuli while their eye-movement data are recorded and used to build an eye-tracking dataset. Then, we propose a novel deep-learning-based model for identifying patients with Alzheimer's Disease (PwAD) and healthy controls (HCs) based on the collected eye-movement data. The proposed model utilizes a nested autoencoder network to extract the eye-movement features from the generated fixation heatmaps and a weight adaptive network layer for the feature fusion, which can preserve as much useful information as possible for the final binary classification. To fully verify the performance of the proposed model, we also design two types of models based on traditional machine-learning and typical deep-learning for comparison. Furthermore, we have also done ablation experiments to verify the effectiveness of each module of the proposed network. Finally, these models are evaluated by four-fold cross-validation on the built eye-tracking dataset. The proposed model shows 85% average accuracy in AD recognition, outperforming machine-learning methods and other typical deep-learning networks.

Perovskite-Based Memristor with 50-Fold Switchable Photosensitivity for In-Sensor Computing Neural Network.

In-sensor computing can simultaneously output image information and recognition results through in-situ visual signal processing, which can greatly improve the efficiency of machine vision. However, in-sensor computing is challenging due to the requirement to controllably adjust the sensor's photosensitivity. Herein, it is demonstrated a ternary cationic halide Cs0.05FA0.81MA0.14 Pb(I0.85Br0.15)3 (CsFAMA) perovskite, whose External quantum efficiency (EQE) value is above 80% in the entire visible region (400-750 nm), and peak responsibility value at 750 nm reaches 0.45 A/W. In addition, the device can achieve a 50-fold enhancement of the photoresponsibility under the same illumination by adjusting the internal ion migration and readout voltage. A proof-of-concept visually enhanced neural network system is demonstrated through the switchable photosensitivity of the perovskite sensor array, which can simultaneously optimize imaging and recognition results and improve object recognition accuracy by 17% in low-light environments.

The relationship between fibrinogen-to-albumin ratio and in-stent restenosis in patients with coronary artery disease undergoing drug-eluting stenting.

BACKGROUND: In-stent restenosis (ISR) remains a significant clinical problem in patients with coronary artery disease (CAD) treated with percutaneous coronary intervention (PCI). Recent studies identified the fibrinogen-to-albumin ratio (FAR) as a novel inflammatory marker to predict inflammation in chronic diseases. This study aimed to investigate the relationship between FAR and ISR in patients with DES implantation. METHODS: A total of 506 consecutive CAD patients were enrolled. Subjects history of successful native vessel PCI with DES at least 12 months prior to undergoing repeat angiography for chest pain. Patients were divided between ISR group (n = 125) and no-ISR group (n = 381). ISR was defined as luminal stenosis ≥50% located within the stent or up to 5 mm beyond the stent edges by the quantitative coronary analysis. Laboratory parameters were measured before angiography. Significant factors associated with ISR were evaluated by multivariate logistic regression analysis. RESULTS: Baseline characteristics were similar between the ISR and no-ISR groups. The ISR group had significantly higher FAR level compared with the no-ISR group (73.26 ± 17.68 vs. 64.90 ± 15.88, P < 0.05). Furthermore, the ISR group had significantly lower albumin level and higher prevalence of diabetes mellitus compared to no-ISR (P < 0.05). In a multivariate analysis, FAR (odds ratio [OR] = 1.039, 95% confidence interval (CI) = 1.024-1.054), albumin (OR = 0.923, 95% CI = 0.389-0.977) and diabetes mellitus (OR = 2.663, 95% CI = 1.587-4.468) were significantly associated with ISR. CONCLUSION: FAR is significantly associated with the development of ISR in CAD patients undergoing PCI with DES implantation.

A Quartz Crystal Microbalance dew point sensor without frequency measurement.

This work deals with the design of a dew point sensor based on Quartz Crystal Microbalance (QCM) without measuring the frequency. This idea is inspired by the fact that the Colpitts oscillation circuit will stop oscillating when the QCM works in the liquid media. The quartz crystal and the electrode are designed through the finite element simulation and the stop oscillating experiment is conducted to verify the sensibility. Moreover, the measurement result is calibrated to approach the true value. At last a series of dew points at the same temperature is measured with the designed sensor. Results show that the designed dew point sensor is able to detect the dew point with the proper accuracy.

Optical fiber-based system for continuous measurement of in-bore projectile velocity.

This paper reports the design of an optical fiber-based velocity measurement system and its application in measuring the in-bore projectile velocity. The measurement principle of the implemented system is based on Doppler effect and heterodyne detection technique. The analysis of the velocity measurement principle deduces the relationship between the projectile velocity and the instantaneous frequency (IF) of the optical fiber-based system output signal. To extract the IF of the fast-changing signal carrying the velocity information, an IF extraction algorithm based on the continuous wavelet transforms is detailed. Besides, the performance of the algorithm is analyzed by performing corresponding simulation. At last, an in-bore projectile velocity measurement experiment with a sniper rifle having a 720 m/s muzzle velocity is performed to verify the feasibility of the optical fiber-based velocity measurement system. Experiment results show that the measured muzzle velocity is 718.61 m/s, and the relative uncertainty of the measured muzzle velocity is approximately 0.021%.

Human immunodeficiency virus type 1 infection increases the in vivo capacity of peripheral monocytes to cross the blood-brain barrier into the brain and the in vivo sensitivity of the blood-brain barrier to disruption by lipopolysaccharide.

Human immunodeficiency virus type 1 (HIV-1), introduced into the brain by HIV-1-infected monocytes which migrate across the blood-brain barrier (BBB), infects resident macrophages and microglia and initiates a process that causes HIV-1-associated neurocognitive disorders. The mechanism by which HIV-1 infection circumvents the BBB-restricted passage of systemic leukocytes into the brain and disrupts the integrity of the BBB is not known. Circulating lipopolysaccharide (LPS), which can compromise the integrity of the BBB, is significantly increased in HIV-1-infected individuals. We hypothesized that HIV-1 infection increases monocyte capacity to migrate across the BBB, which is further facilitated by a compromise of BBB integrity mediated by the increased systemic LPS levels present in HIV-1-infected individuals. To investigate this possibility, we examined the in vivo BBB migration of monocytes derived from our novel mouse model, JR-CSF/EYFP mice, which are transgenic for both a long terminal repeat-regulated full-length infectious HIV-1 provirus and ROSA-26-regulated enhanced yellow fluorescent protein. We demonstrated that JR-CSF/EYFP mouse monocytes displayed an increased capacity to enter the brain by crossing either an intact BBB or a BBB whose integrity was partially compromised by systemic LPS. We also demonstrated that the JR-CSF mouse BBB was more susceptible to disruption by systemic LPS than the control wild-type mouse BBB. These results demonstrated that HIV-1 infection increased the ability of monocytes to enter the brain and increased the sensitivity of the BBB to disruption by systemic LPS, which is elevated in HIV-1-infected individuals. These mice represent a new in vivo system for studying the mechanism by which HIV-1-infected monocytes migrate into the brain.

Increased in vivo activation of microglia and astrocytes in the brains of mice transgenic for an infectious R5 human immunodeficiency virus type 1 provirus and for CD4-specific expression of human cyclin T1 in response to stimulation by lipopolysaccharides.

Inflammatory mediators and viral products produced by human immunodeficiency virus (HIV)-infected microglia and astrocytes perturb the function and viability of adjacent uninfected neuronal and glial cells and contribute to the pathogenesis of HIV-associated neurocognitive disorders (HAND). In vivo exposure to lipopolysaccharide (LPS) activates parenchymal microglia and astrocytes and induces cytokine and chemokine production in the brain. HIV-infected individuals display increased circulating LPS levels due to microbial translocation across a compromised mucosa barrier. We hypothesized that HIV-infected microglia and astrocytes display increased sensitivity to the proinflammatory effects of LPS, and this combines with the increased levels of systemic LPS in HIV-infected individuals to contribute to the development of HAND. To examine this possibility, we determined the in vivo responsiveness of HIV-infected microglia and astrocytes to LPS using our mouse model, JR-CSF/human cyclin T1 (JR-CSF/hu-cycT1) mice, which are transgenic for both an integrated full-length infectious HIV type 1 (HIV-1) provirus derived from the primary R5-tropic clinical isolate HIV-1(JR-CSF) regulated by the endogenous HIV-1 long terminal repeat and the hu-cycT1 gene under the control of a CD4 promoter. In the current report, we demonstrated that in vivo-administered LPS more potently activated JR-CSF/hu-cycT1 mouse microglia and astrocytes and induced a significantly higher degree of monocyte chemoattractant protein production by JR-CSF/hu-cycT1 astrocytes compared to that of the in vivo LPS response of control littermate mouse microglia and astrocytes. These results indicate that HIV infection increases the sensitivity of microglia and astrocytes to inflammatory stimulation and support the use of these mice as a model to investigate various aspects of the in vivo mechanism of HIV-induced neuronal dysfunction.

CD4-specific transgenic expression of human cyclin T1 markedly increases human immunodeficiency virus type 1 (HIV-1) production by CD4+ T lymphocytes and myeloid cells in mice transgenic for a provirus encoding a monocyte-tropic HIV-1 isolate.

Human immunodeficiency virus type 1 (HIV-1)-encoded Tat provides transcriptional activation critical for efficient HIV-1 replication by interacting with cyclin T1 and recruiting P-TEFb to efficiently elongate the nascent HIV transcript. Tat-mediated transcriptional activation in mice is precluded by species-specific structural differences that prevent Tat interaction with mouse cyclin T1 and severely compromise HIV-1 replication in mouse cells. We investigated whether transgenic mice expressing human cyclin T1 under the control of a murine CD4 promoter/enhancer cassette that directs gene expression to CD4(+) T lymphocytes and monocytes/macrophages (hu-cycT1 mice) would display Tat responsiveness in their CD4-expressing mouse cells and selectively increase HIV-1 production in this cellular population, which is infected primarily in HIV-1-positive individuals. To this end, we crossed hu-cycT1 mice with JR-CSF transgenic mice carrying the full-length HIV-1(JR-CSF) provirus under the control of the endogenous HIV-1 long terminal repeat and demonstrated that human cyclin T1 expression is sufficient to support Tat-mediated transactivation in primary mouse CD4 T lymphocytes and monocytes/macrophages and increases in vitro and in vivo HIV-1 production by these stimulated cells. Increased HIV-1 production by CD4(+) T lymphocytes was paralleled with their specific depletion in the peripheral blood of the JR-CSF/hu-cycT1 mice, which increased over time. In addition, increased HIV-1 transgene expression due to human cyclin T1 expression was associated with increased lipopolysaccharide-stimulated monocyte chemoattractant protein 1 production by JR-CSF mouse monocytes/macrophages in vitro. Therefore, the JR-CSF/hu-cycT1 mice should provide an improved mouse system for investigating the pathogenesis of various aspects of HIV-1-mediated disease and the efficacies of therapeutic interventions.

Microglia from mice transgenic for a provirus encoding a monocyte-tropic HIV type 1 isolate produce infectious virus and display in vitro and in vivo upregulation of lipopolysaccharide-induced chemokine gene expression.

A large body of evidence has indicated that microglia are the predominant cellular location for HIV-1 in the brains of HIV-1-infected individuals and play a direct role in the development of HIV-1-associated dementia (HAD). Therefore, investigation of the mechanism by which HIV-1-infected microglia contribute to the development of HIV-associated dementia should be facilitated by the creation of a mouse model wherein microglia carry replication-competent HIV-1. To circumvent the inability of HIV-1 to infect mouse cells, we developed a mouse line that is transgenic for a full-length proviral clone of a monocyte-tropic HIV-1 isolate, HIV-1(JR-CSF) (JR-CSF mice), whose T cells and monocytes produce infectious HIV-1. We detected expression of the long terminal repeat-regulated proviral transgene in the microglia of these transgenic mice and demonstrated that it was increased by in vitro and in vivo stimulation with lipopolysaccharide. Furthermore, microglia isolated from JR-CSF mouse brains produced HIV-1 that was infectious in vitro and in vivo. We examined the effect that carriage of the HIV-1 provirus had on chemokine gene regulation in the brains of these mice and demonstrated that MCP-1 gene expression by JR-CSF mouse microglia and brains was more responsive to in vitro and in vivo stimulation with lipopolysaccharide than were microglia and brains from control mice. Thus, this study indicates that the JR-CSF mice may represent a new mouse model to study the effect of HIV-1 replication on microglia function and its contribution to HIV-1-associated neurological disease.

[Impact of salt stress on peroxidase activity in Populus deltoides cambium and its consequence].

Studies on the impact of salt stress on the peroxidase (POD) activity in Populus deltoides cambium and on the anatomic characteristics of its secondary xylem indicated that the changes of POD activity caused by soil salinity stress behaved differently in dormant and growing period. In low salinity soil, the POD activity of dormant Populus deltoides showed the trend of increasing first, and then, decreasing as the soil salinity was rising gradually. Namely, with the rising of soil salinity in the range of 0.024-->0.094-->0.145%, the POD activity varied in the range of 83.7-->132.1-->63.2 units.min-1.g-1FW accordingly. However, in growing season, with the gradual increase of soil salinity from 0.036 to 0.289%, the POD activity decreased gradually from 405.2 to 107.2 units.min-1.g-1FW. There were regular changes in wood anatomic characteristics of Populus deltoides. Namely, with the increase of soil salinity 0.036-->0.125-->0.289%, the vessel diameters (VD) of both early wood and late wood varied accordingly in the range of 41.8-->56.6-->43.4 microns, and 29.1-->33.1-->33.1 microns, respectively. The vessel frequencies (VF) were 141.8-->113.8-->144.2 and 160.0-->134.8-->206.7 entries mm-2, respectively. Along with the gradual increase of soil salinity from 0.036 to 0.289%, the fiber length gradually decreased from 693.8 to 570.4 microns, and the fiber width decreased from 14.9 to 13.5 microns.