Multi-Time-Scale Optimal Scheduling Strategy for Marine Renewable Energy Based on Deep Reinforcement Learning Algorithm.

Surrounded by the Shandong Peninsula, the Bohai Sea and Yellow Sea possess vast marine energy resources. An analysis of actual meteorological data from these regions indicates significant seasonality and intra-day uncertainty in wind and photovoltaic power generation. The challenge of scheduling to leverage the complementary characteristics of various renewable energy sources for maintaining grid stability is substantial. In response, we have integrated wave energy with offshore photovoltaic and wind power generation and propose a day-ahead and intra-day multi-time-scale rolling optimization scheduling strategy for the complementary dispatch of these three energy sources. Using real meteorological data from this maritime area, we employed a CNN-LSTM neural network to predict the power generation and load demand of the area on both day-ahead 24 h and intra-day 1 h time scales, with the DDPG algorithm applied for refined electricity management through rolling optimization scheduling of the forecast data. Simulation results demonstrate that the proposed strategy effectively meets load demands through complementary scheduling of wave power, wind power, and photovoltaic power generation based on the climatic characteristics of the Bohai and Yellow Sea regions, reducing the negative impacts of the seasonality and intra-day uncertainty of these three energy sources on the grid. Additionally, compared to the day-ahead scheduling strategy alone, the day-ahead and intra-day rolling optimization scheduling strategy achieved a reduction in system costs by 16.1% and 22% for a typical winter day and a typical summer day, respectively.

Capillary Manganese Halide Needle-Like Array Scintillator with Isolated Light Crosstalk for Micro-X-Ray Imaging.

The exacerbation of inherent light scattering with increasing scintillator thickness poses a major challenge for balancing the thickness-dependent spatial resolution and scintillation brightness in X-ray imaging scintillators. Herein, a thick pixelated needle-like array scintillator capable of micrometer resolution is fabricated via waveguide structure engineering. Specifically, this involves integrating a straightforward low-temperature melting process of manganese halide with an aluminum-clad capillary template. In this waveguide structure, the oriented scintillation photons propagate along the well-aligned scintillator and are confined within individual pixels by the aluminum reflective cladding, as substantiated from the comprehensive analysis including laser diffraction experiments. Consequently, thanks to isolated light-crosstalk channels and robust light output due to increased thickness, ultrahigh spatial resolutions of 60.8 and 51.7 lp mm-1 at a modulation transfer function (MTF) of 0.2 are achieved on 0.5 mm and even 1 mm thick scintillators, respectively, which both exceed the pore diameter of the capillary arrays' template (Φ = 10 µm). As far as it is known, these micrometer resolutions are among the highest reported metal halide scintillators and are never demonstrated on such thick scintillators. Here an avenue is presented to the demand for thick scintillators in high-resolution X-ray imaging across diverse scientific and practical fields.

Predictive role of impulsivity, anxiety, and depression in the efficacy of intermittent theta burst transcranial magnetic stimulation modalities for treating methamphetamine use disorder: A randomized clinical trial.

INTRODUCTION: Methamphetamine use disorder (MUD) can cause impulsive behavior, anxiety, and depression. Stimulation of the left dorsolateral prefrontal cortex in MUD patients by intermittent theta burst repetitive transcranial magnetic stimulation (iTBS-rTMS) is effective in reducing cravings, impulsive behavior, anxiety, and depression. The purpose of this study was to explore whether these psychological factors helped to predict MUD patients' responses to iTBS-rTMS treatment. METHODS: Fifty MUD patients and sixty healthy subjects matched for general conditions were used as study subjects. The study randomly divided MUD patients into iTBS-rTMS and sham stimulation groups and received 20 sessions of real or sham iTBS-rTMS treatment, and the study collected cue-related evoked craving data before and after treatment. All subjects completed the Barratt Impulsiveness Scale (BIS-11), Self-rating Anxiety Scale (SAS), and Self-rating Depression Scale (SDS). RESULTS: The MUD patients showed significantly higher levels of impulsivity, anxiety, and depression than the healthy subjects. The MUD patients who received the real treatment had significantly lower impulsivity, anxiety, and depression scores, and better treatment effects on cravings than the sham stimulation group. The Spearman rank correlation and stepwise multiple regression analyses showed that the baseline BIS-11 and the reduction rate (RR) of BIS-11 and RR of SDS were positively correlated with the decrease in cravings in the iTBS-rTMS group. ROC curve analysis showed that RR of SDS (AUC = 91.6 %; 95 % CI = 0.804-1.000) had predictive power to iTBS- rTMS therapeutic efficacy, the cutoff value is 15.102 %. CONCLUSIONS: iTBS-rTMS had a good therapeutic effect in MUD patients and the baseline impulsivity, the improved depression and impulsivity were associated with therapeutic effect of iTBS-rTMS. The improved depression had the potential to predict the efficacy of the iTBS-rTMS modality for MUD treatment.

p-Nrf2/HO-1 Pathway Involved in Methamphetamine-induced Executive Dysfunction through Endoplasmic Reticulum Stress and Apoptosis in the Dorsal Striatum.

Methamphetamine (METH) abuse is known to cause executive dysfunction. However, the molecular mechanism underlying METH induced executive dysfunction remains unclear. Go/NoGo experiment was performed in mice to evaluate METH-induced executive dysfunction. Immunoblot analysis of Nuclear factor-E2-related factor 2 (Nrf2), phosphorylated Nrf2 (p-Nrf2), heme-oxygenase-1 (HO-1), Glucose Regulated Protein 78(GRP78), C/EBP homologous protein (CHOP), Bcl-2, Bax and Caspase3 was performed to evaluate the levels of oxidative stress, endoplasmic reticulum (ER) stress and apoptosis in the dorsal striatum (Dstr). Malondialdehyde (MDA) levels and glutathione peroxidase (GSH-Px) activity was conducted to evaluate the level of oxidative stress. TUNEL staining was conducted to detect apoptotic neurons. The animal Go/NoGo testing confirmed that METH abuse impaired the inhibitory control ability of executive function. Meanwhile, METH down-regulated the expression of p-Nrf2, HO-1 and GSH-Px and activated ER stress and apoptosis in the Dstr. Microinjection of Tert-butylhydroxyquinone (TBHQ), an Nrf2 agonist, into the Dstr increased the expression of p-Nrf2, HO-1, and GSH-Px, ameliorated ER stress, apoptosis and executive dysfunction caused by METH. Our results indicated that the p-Nrf2/HO-1 pathway was potentially involved in mediating methamphetamine-induced executive dysfunction by inducing endoplasmic reticulum stress and apoptosis in the dorsal striatum.

Singlet Fission-Based High-Resolution X-Ray Imaging Scintillation Screens.

X-ray imaging technology is critical to numerous different walks of daily life, ranging from medical radiography and security screening all the way to high-energy physics. Although several organic chromophores are fabricated and tested as X-ray imaging scintillators, they generally show poor scintillation performance due to their weak X-ray absorption cross-section and inefficient exciton utilization efficiency. Here, a singlet fission-based high-performance organic X-ray imaging scintillator with near unity exciton utilization efficiency is presented. Interestingly, it is found that the X-ray sensitivity and imaging resolution of the singlet fission-based scintillator are dramatically improved by an efficient energy transfer from a thermally activated delayed fluorescence (TADF) sensitizer, in which both singlet and triplet excitons can be efficiently harnessed. The fabricated singlet fission-based scintillator exhibits a high X-ray imaging resolution of 27.5 line pairs per millimeter (lp mm-1 ), which exceeds that of most commercial scintillators, demonstrating its high potential use in medical radiography and security inspection.

Highly Efficient, Flexible, and Eco-Friendly Manganese(II) Halide Nanocrystal Membrane with Low Light Scattering for High-Resolution X-ray Imaging.

Scintillators enable invisible X-ray to be converted into ultraviolet (UV)/visible light that can be collected using a sensor array and is the core component of the X-ray imaging system. However, combining the excellent properties of high light output, high spatial resolution, flexibility, non-toxicity, and cost effectiveness into a single X-ray scintillator remains a great challenge. Herein, a novel scintillator based on benzyltriphenylphosphonium manganese(II) bromide (BTP2MnBr4) nanocrystal (NC) membranes was developed by the in situ fabrication strategy. The long Mn-Mn distance provided by the large BTP cation allows the nonradiative energy dissipation in this manganese(II) halide to be significantly suppressed. As a result, the flexible BTP2MnBr4 NC scintillator shows an excellent linear response to the X-ray dose rate, a high light yield of ∼71,000 photon/MeV, a low detection limit of 86.2 nGyair/s at a signal-to-noise ratio of 3, a strong radiation hardness, and a long-term thermal stability. Thanks to the low Rayleigh scattering associated with the dense distribution of nanometer-scale emitters, light cross-talk in X-ray imaging is greatly suppressed. The impressively high-spatial resolution X-ray imaging (23.8 lp/mm at modulation transfer function = 0.2 and >20 lp/mm for a standard pattern chart) was achieved on this scintillator. Moreover, well-resolved 3D dynamic rendering X-ray projections were also successfully demonstrated using this scintillator. These results shed light on designing efficient, flexible, and eco-friendly scintillators for high-resolution X-ray imaging.

Heterojunctions on Ta2O5@MWCNT for Ultrasensitive Ethanol Sensing at Room Temperature.

Heterojunctions of Ta2O5 and multiwalled carbon nanotubes (MWCNTs) have been successfully synthesized by a facile and cost-effective hydrothermal method, with a super thin and uniform Ta2O5 shell wrapped around the MWCNT. The combination of Ta2O5 and MWCNTs at the interface not only modifies the morphology but also forms the p-n heterojunction, which contributes to the reconstruction of band structure, as well as the low resistance of matrix and highly chemisorbed oxygen content. The Ta2O5@MWCNT p-n heterojunction exhibits ultrasensitive performance to ethanol at room temperature, with a response of 3.15 toward 0.8 ppm ethanol and a detection limit of 0.173 ppm. The sensor has a high reproducibility at various concentrations of ethanol, superior selectivity to other gases, and long-term stability. The strategy of hybriding metal oxide semiconductors with MWCNT promises to provide a feasible and further developable pathway for high-performance room-temperature gas sensors.

The chain-mediation pathway of gender regarding academic delay of gratification in college students is regulated by anxiety/depressive mood and prospective memory.

Objective: This study investigated the relationship between gender and academic delay of gratification (ADOG) in college students and explored the mediating roles of anxiety/depressive mood and prospective memory to provide a theoretical intervention approach based on internal mechanisms. Methods: Random cluster sampling was conducted on 609 students from three universities situated in the Province of Anhui, China with the use of data from several questionnaires: the general information questionnaire, Generalized Anxiety Disorder Scale, Patient Health Questionnaire, Prospective and Retrospective Memory (PRM) Questionnaire, and ADOG Scale. Results: The females' anxiety and depression levels were lower than that of the males, while the female PRM and ADOG performance improved when compared to that of the males. Anxiety and depression were negatively correlated with PRM and ADOG, respectively, whereas the PRM and ADOG data demonstrated a positive correlation. Depression/anxiety and prospective memory also played a chain intermediary role between gender and ADOG. Conclusion: Gender not only directly affects college students' ADOG, but it also has indirect effects through depression/anxiety and prospective memory. Therefore, it is very important to treat students' mental health differently according to gender to improve prospective memory and delayed academic satisfaction.

Effect of Coupling Medium on Penetration Depth in Microwave Medical Imaging.

In microwave medical imaging, the human skin reflects most of microwave energy due to the impedance mismatch between the air and the body. As a result, only a small portion of the microwave energy can enter the body and work for medical purpose. One solution to tackle this issue is to use a coupling (or matching) medium, which can reduce unwanted reflections on the skin and meanwhile improve spatial imaging resolution. A few types of fluids were measured in this paper for their dielectric properties between 500 MHz and 13.5 GHz. Measurements were performed by a Keysight programmable network analyzer (PNA) with a dielectric probe kit, and dielectric constant and conductivity of the fluids were presented in this paper. Then, quantitative computations were exercised to present the attenuations due to the reflection on the skin and to the loss in each coupling medium, based on the measured liquid dielectric values. Finally, electromagnetic simulations verified that the coupling liquid can allow more microwave energy to enter the body to allow for a more efficient medical examination.

Optoelectronic Artificial Synaptic Device Based on Amorphous InAlZnO Films for Learning Simulations.

Neuromorphic computing, which mimics brain function, can address the shortcomings of the "von Neumann" system and is one of the critical components of next-generation computing. The use of light to stimulate artificial synapses has the advantages of low power consumption, low latency, and high stability. We demonstrate amorphous InAlZnO-based light-stimulated artificial synaptic devices with a thin-film transistor structure. The devices exhibit fundamental synaptic properties, including excitatory postsynaptic current, paired-pulse facilitation (PPF), and short-term plasticity to long-term plasticity conversion under light stimulation. The PPF index stimulated by 375 nm light is 155.9% when the time interval is 0.1 s. The energy consumption of each synaptic event is 2.3 pJ, much lower than that of ordinary MOS devices and other optical-controlled synaptic devices. The relaxation time constant reaches 277 s after only 10 light spikes, which shows the great synaptic plasticity of the device. In addition, we simulated the learning-forgetting-relearning-forgetting behavior and learning efficiency of human beings under different moods by changing the gate voltage. This work is expected to promote the development of high-performance optoelectronic synaptic devices for neuromorphic computing.

Associations of executive function and age of first use of methamphetamine with methamphetamine relapse.

Background and aims: Methamphetamine (MA) is a psychostimulant associated with a high relapse rate among patients with MA use disorder (MUD). Long-term use of MA is associated with mental disorders, executive dysfunction, aggressive behaviors, and impulsivity among patients with MUD. However, identifying which factors may be more closely associated with relapse has not been investigated. Thus, we aimed to investigate the psychological factors and the history of MA use that may influence MA relapse. Methods: This cross-sectional study included 168 male MUD patients (MUD group) and 65 healthy male residents (control group). Each patient was evaluated with self-report measures of executive dysfunction, psychopathological symptoms, impulsiveness, aggressiveness, and history of MA use. Data were analyzed with t-tests, analyses of variance, and correlation and regression analyses. Results: The MUD group reported greater executive dysfunction, psychopathological symptoms, impulsivity, and aggression than the control group. Lower age of first MA use was associated both with having relapsed one or more times and with having relapsed two or more times; greater executive dysfunction was associated only with having relapsed two or more times. Conclusion: Patients with MUD reported worse executive function and mental health. Current results also suggest that lower age of first MA use may influence relapse rate in general, while executive dysfunction may influence repeated relapse in particular. The present results add to the literature concerning factors that may increase the risk of relapse in individuals with MUD.

Generation of Digital Brain Phantom for Machine Learning Application of Dopamine Transporter Radionuclide Imaging.

While machine learning (ML) methods may significantly improve image quality for SPECT imaging for the diagnosis and monitoring of Parkinson's disease (PD), they require a large amount of data for training. It is often difficult to collect a large population of patient data to support the ML research, and the ground truth of lesion is also unknown. This paper leverages a generative adversarial network (GAN) to generate digital brain phantoms for training ML-based PD SPECT algorithms. A total of 594 PET 3D brain models from 155 patients (113 male and 42 female) were reviewed and 1597 2D slices containing the full or a portion of the striatum were selected. Corresponding attenuation maps were also generated based on these images. The data were then used to develop a GAN for generating 2D brain phantoms, where each phantom consisted of a radioactivity image and the corresponding attenuation map. Statistical methods including histogram, Fréchet distance, and structural similarity were used to evaluate the generator based on 10,000 generated phantoms. When the generated phantoms and training dataset were both passed to the discriminator, similar normal distributions were obtained, which indicated the discriminator was unable to distinguish the generated phantoms from the training datasets. The generated digital phantoms can be used for 2D SPECT simulation and serve as the ground truth to develop ML-based reconstruction algorithms. The cumulated experience from this work also laid the foundation for building a 3D GAN for the same application.

Near-Field Microwave Scattering Formulation by A Deep Learning Method.

A deep learning method is applied to modelling electromagnetic (EM) scattering for microwave breast imaging (MBI). The neural network (NN) accepts 2D dielectric breast maps at 3 GHz and produces scattered-field data on an antenna array composed of 24 transmitters and 24 receivers. The NN was trained by 18,000 synthetic digital breast phantoms generated by generative adversarial network (GAN), and the scattered-field data pre-calculated by method of moments (MOM). Validation was performed by comparing the 2,000 NN-produced datasets isolated from the training data with the data computed by MOM. Finally, data generated by NN and MOM were used for image reconstruction. The reconstruction demonstrated that errors caused by NN would not significantly affect the image result. But the computational speed of NN was nearly 104 times faster than the MOM, indicating that deep learning has the potential to be considered as a fast tool for EM scattering computation.

Dielectric Breast Phantoms by Generative Adversarial Network.

In order to conduct the research of machine-learning (ML) based microwave breast imaging (MBI), a large number of digital dielectric breast phantoms that can be used as training data (ground truth) are required but are difficult to be achieved from practice. Although a few dielectric breast phantoms have been developed for research purpose, the number and the diversity are limited and is far inadequate to develop a robust ML algorithm for MBI. This paper presents a neural network method to generate 2D virtual breast phantoms that are similar to the real ones, which can be used to develop ML-based MBI in the future. The generated phantoms are similar but are different from those used in training. Each phantom consists of several images with each representing the distribution of a dielectric parameter in the breast map. Statistical analysis was performed over 10,000 generated phantoms to investigate the performance of the generative network. With the generative network, one may generate unlimited number of breast images with more variations, so the ML-based MBI will be more ready to deploy.

The Influence of Online Game Behaviors on the Emotional State and Executive Function of College Students in China.

Background and Objective: Since the classification of gaming disorder (GD) by the World Health Organization (WHO) as "mental disorder caused by addictive behaviors," there has been controversy regarding whether online game behaviors can lead to mental disorder. This study aims to clarify the correlation between the online game behaviors of college students and anxiety, depression, and executive function of college students in China, from a questionnaire-based investigation. Methods: Based on the whole class random sampling method, a questionnaire survey was conducted among college students in Northern Anhui, China from March 7 to March 27, 2020. The questionnaires included the Internet Game Addiction (IGA) Scale, Behavior Rating Inventory of Executive Function (Adult Version, BRIEF-A), Generalized Anxiety Disorder Scale (GAD-7) and Patient Health Questionnaire Scale (PHQ-9). Results: A total of 850 participants completed the survey, including 353 males (41.53%) and 497 females (58.47%). The primary age group was 18-27 years (91.53%), and the educational background was a bachelor's degree (94.7%). The study found that the online behavior of 17.76% of college students was online game behavior. This study did not identify any students who met the criteria for IGA, and 3% met the criteria for indulgent behavior. A dual role of online games was identified; moderate online game activities can improve the emotional state and executive function of college students, while excessive online game behaviors that may not reach the degree of addiction can also harm emotional state and executive function. Conclusions: This study suggests that although IGA has been regarded as a mental disease, online game behavior should be treated differently. Online game activities should not be entirely denied, but mental disorders caused by excessive gaming activities deserve attention. In particular, the emotional state and executive function of students with excessive online game behaviors should be monitored and intervened in advance to avoid game behaviors turning into indulgent behaviors or addiction. As a cognitive control process, executive function may play a key role in regulating IGA and emotional state.

Performance analysis of an all-dielectric planar Mikaelian lens antenna for 1-D beam-steering application.

This paper investigates the performance of an all-dielectric planar Mikaelian lens based on ray transfer matrices and full-wave analysis for 1-D beam-steering application. This new lens concept has its intrinsic flat shape characteristic allowing for a simple low-cost planar feed technology. To verify the design concept, a lens prototype excited by five rectangular microstrip patch antennas with perforated structure (21×24 holes) is fabricated using stereolithography (SLA) 3-D printing. The simulated and measured results of the proposed lens prototype, operating at 10 GHz, shows that the switched-beam coverage over a certain range of beam-steering angles can be obtained. The intrinsic phase error of lens resulting from comatic aberration exhibits obvious increase as the off-axis angle of beam increases, which leads to further deterioration of the corresponding radiated beam. The beam-steering capabilities from -20° to +20° with around 13.2 dBi of realized gain and side-lobe level (less than -11.5dB), and up to potential steering angles (±30°) with around 10 dBi of realized gain can be steadily achieved. Moreover, the realized gain, efficiency and side-lobe level can be further improved to get better radiation performances by using other materials with lower loss tangent. Due to its intrinsic flat shape characteristic, this lens concept could be a potential alternative to develop a low-cost, low-profile and easy-to-fabricate beam-switching array antenna for microwave communication applications.

SPECTnet: a deep learning neural network for SPECT image reconstruction.

BACKGROUND: Single photon emission computed tomography (SPECT) is an important functional tool for clinical diagnosis and scientific research of brain disorders, but suffers from limited spatial resolution and high noise due to hardware design and imaging physics. The present study is to develop a deep learning technique for SPECT image reconstruction that directly converts raw projection data to image with high resolution and low noise, while an efficient training method specifically applicable to medical image reconstruction is presented. METHODS: A custom software was developed to generate 20,000 2-D brain phantoms, of which 16,000 were used to train the neural network, 2,000 for validation, and the final 2,000 for testing. To reduce development difficulty, a two-step training strategy for network design was adopted. We first compressed full-size activity image (128×128 pixels) to a one-D vector consisting of 256×1 pixels, accomplished by an autoencoder (AE) consisting of an encoder and a decoder. The vector is a good representation of the full-size image in a lower-dimensional space and was used as a compact label to develop the second network that maps between the projection-data domain and the vector domain. Since the label had 256 pixels only, the second network was compact and easy to converge. The second network, when successfully developed, was connected to the decoder (a portion of AE) to decompress the vector to a regular 128×128 image. Therefore, a complex network was essentially divided into two compact neural networks trained separately in sequence but eventually connectable. RESULTS: A total of 2,000 test examples, a synthetic brain phantom, and de-identified patient data were used to validate SPECTnet. Results obtained from SPECTnet were compared with those obtained from our clinic OS-EM method. Images with lower noise and more accurate information in the uptake areas were obtained by SPECTnet. CONCLUSIONS: The challenge of developing a complex deep neural network is reduced by training two separate compact connectable networks. The combination of the two networks forms the full version of SPECTnet. Results show that the developed neural network can produce more accurate SPECT images.

Artificial intelligence in single photon emission computed tomography (SPECT) imaging: a narrative review.

Artificial intelligence (AI) has been widely applied to medical imaging. The use of AI for emission computed tomography, particularly single-photon emission computed tomography (SPECT) emerged nearly 30 years ago but has been accelerated in recent years due to the development of AI technology. In this review, we will describe and discuss the progress of AI technology in SPECT imaging. The applications of AI are dispersed in disease prediction and diagnosis, post-reconstruction image denoising, attenuation map generation, and image reconstruction. These applications are relevant to many disease categories such as the neurological disorders, kidney failure, cancer, heart disease, etc. This review summarizes these applications so that SPECT researchers can have a reference overview of the role of AI in current SPECT studies. For each application, we followed the timeline to present the evolution of AI's usage and offered insights on how AI was combined with the knowledge of underlying physics as well as traditional non-learning techniques. Ultimately, AI applications are critical to the progress of modern SPECT technology because they provide compensations for many deficiencies in conventional SPECT imaging methods and demonstrate unparalleled success. Nonetheless, AI also has its own challenges and limitations in the medical field, including SPECT imaging. These fundamental questions are discussed, and possible future directions and countermeasures are suggested.

Perforated extensible 3-D hyperbolic secant lens antenna for directive antenna applications using additive manufacturing.

In this paper, a novel three-dimensional (3-D) generalized hyperbolic secant (H-S) lens is first introduced using perforated dielectric material. The attractiveness of this new lens is its unique intrinsic flat shape characteristic and extensibility for different configuration scenarios, which provide a potential alternative design for a planar Luneburg and half Maxwell fish-eye lens based on a complex conformal mapping method. A high gain and wideband printed antipodal fermi antenna as a feeding source is employed in the proposed lens antenna prototype. The high radiation performance with low side lobe level of the fabricated lens prototype is validated from 8.2GHz to 12.5GHz, demonstrating 23.8 dBi realized gain at 10 GHz with 3-dB beamwidth of 9° and 2-dB fractional gain bandwidth of 41.6%. Besides, the total radiation efficiency is above ∼40% across all tested frequencies, which suggests the proposed H-S lens itself has a broadband response. The simplicity and low-cost fabrication using additive manufacturing of its lens design indicates great potential in broadband high directive antenna applications.

A Learned Reconstruction Network for SPECT Imaging.

A neural network designed specifically for SPECT image reconstruction was developed. The network reconstructed activity images from SPECT projection data directly. Training was performed through a corpus of training data including that derived from digital phantoms generated from custom software and the corresponding projection data obtained from simulation. When using the network to reconstruct images, input projection data were initially fed to two fully connected (FC) layers to perform a basic reconstruction. Then the output of the FC layers and an attenuation map were delivered to five convolutional layers for signal-decay compensation and image optimization. To validate the system, data not used in training, simulated data from the Zubal human brain phantom, and clinical patient data were used to test reconstruction performance. Reconstructed images from the developed network proved closer to the truth with higher resolution and quantitative accuracy than those from conventional OS-EM reconstruction. To understand better the operation of the network for reconstruction, intermediate results from hidden layers were investigated for each step of the processing. The network system was also retrained with noisy projection data and compared with that developed with noise-free data. The retrained network proved even more robust after having learned to filter noise. Finally, we showed that the network still provided sharp images when using reduced view projection data (retrained with reduced view data).