Myc rearrangement redefines the stratification of high-risk multiple myeloma.

BACKGROUND: Myc rearrangement (Myc-R) is a controversial factor linked to adverse outcomes in newly diagnosed multiple myeloma (NDMM). AIMS: This study aimed to evaluate the impact of Myc-R on the prognosis of NDMM patients and its role in risk stratification compared with traditional high-risk cytogenetic abnormalities (HRCAs). MATERIALS & METHODS: A total of 417 NDMM patients enrolled from May 2009 to September 2022 were included. Fluorescence in situ hybridization (FISH) was used to detect Myc-R and other Myc abnormalities (Myc-OA). Median progression-free survival (PFS) and overall survival (OS) were analyzed using Kaplan-Meier methods and log-rank tests. Multivariate Cox regression analysis was used to identify independent risk factors. RESULTS: Myc-R was identified in 13.7% of patients, while 14.6% had Myc-OA. Patients with Myc-R had significantly shorter median PFS (15.9 months) and OS (25.1 months) compared with those with Myc-OA (24.5 months PFS; 29.8 months OS) and Myc-negative (Myc-N) status (29.8 months PFS, 29.8 months OS). Myc-R was independently associated with worse PFS and OS compared to Myc-OA. Patients with Myc-R alone had inferior median PFS (15.9 months vs. 28.1 months, p = 0.032) and OS (25.1 months vs. 61.2 months, p = 0.04) compared to those with traditional single HRCA. DISCUSSION: The study suggests that traditional single HRCA may not significantly impact survival in NDMM patients. However, incorporating Myc rearrangement or traditional double/triple-hit HRCAs into the risk stratification model improves its predictive value, highlighting the importance of Myc rearrangement in risk assessment. CONCLUSION: Myc rearrangement is an independent adverse prognostic factor in NDMM. The incorporation of Myc rearrangement or multiple HRCAs into risk stratification models improves their prognostic value, providing a novel perspective on high-risk factors in NDMM.

Differential Transcription Profiling Reveals the MicroRNAs Involved in Alleviating Damage to Photosynthesis under Drought Stress during the Grain Filling Stage in Wheat.

To explore the possible novel microRNA (miRNA) regulatory pathways in Zhengmai 1860, a newly cultivated drought-tolerant wheat (Triticum aestivum L.) cultivar, miRNA transcriptome sequencing of the flag leaves of Zhengmai 1860, drought-sensitive variety Zhoumai 18, and drought-resistant variety Bainong 207 was performed during the grain filling stage. We also observed changes in the chloroplast ultrastructure, phytohormone levels, and antioxidant- and photosynthesis-related physiological indicators in three wheat varieties. The results showed that the flag leaves of the drought-tolerant variety Zhengmai 1860 had higher chlorophyll contents and net photosynthetic rates than those of Zhoumai 18 under drought stress during the grain filling stage; in addition, the chloroplast structure was more complete. However, there was no significant difference between Zhengmai 1860 and Bainong 207. MiRNA transcriptome analysis revealed that the differential expression of the miRNAs and mRNAs exhibited variable specificity. The KEGG pathway enrichment results indicated that most of the genes were enriched in the MAPK signaling pathway, plant hormone signal transduction, photosynthetic antennae protein, and amino acid and carbohydrate metabolism. In the drought-tolerant cultivar Zhengmai 1860, tae-miR408 was targeted to regulate the allene oxide synthase (AOS) gene, inhibit its expression, reduce the AOS content, and decrease the synthesis of jasmonic acid (JA) and abscisic acid (ABA). The results of this study suggest that Zhengmai 1860 could improve the photosynthetic performance of flag leaves by inhibiting the expression of genes involved in the JA pathway through miRNAs under drought conditions. Moreover, multiple miRNAs may target chlorophyll, antioxidant enzymes, phytohormone signal transduction, and other related pathways; thus, it is possible to provide a more theoretical basis for wheat molecular breeding.

Inferring neural activity before plasticity as a foundation for learning beyond backpropagation.

For both humans and machines, the essence of learning is to pinpoint which components in its information processing pipeline are responsible for an error in its output, a challenge that is known as 'credit assignment'. It has long been assumed that credit assignment is best solved by backpropagation, which is also the foundation of modern machine learning. Here, we set out a fundamentally different principle on credit assignment called 'prospective configuration'. In prospective configuration, the network first infers the pattern of neural activity that should result from learning, and then the synaptic weights are modified to consolidate the change in neural activity. We demonstrate that this distinct mechanism, in contrast to backpropagation, (1) underlies learning in a well-established family of models of cortical circuits, (2) enables learning that is more efficient and effective in many contexts faced by biological organisms and (3) reproduces surprising patterns of neural activity and behavior observed in diverse human and rat learning experiments.

Reversible Emulsions from Polyoxometalate-Polymer: A Robust Strategy to Cyclic Emulsion Catalysis and High-Internal-Phase Emulsion Materials.

Reversible Pickering emulsions, achieved by switchable, interfacially active colloidal particles, that enable on-demand emulsification/demulsification or phase inversion, hold substantial promise for biphasic catalysis, emulsion polymerization, cutting fluids, and crude oil pipeline transportation. However, particles with such a responsive behavior usually require complex chemical syntheses and surface modifications, limiting their extensive use. Herein, we report a simple route to generate emulsions that can be controlled and reversibly undergo phase inversion. The emulsions are prepared and stabilized by the interfacial assembly of polyoxometalate (POM)-polymer, where their electrostatic interaction at the interface is dynamic. The wettability of the POMs that dictates the emulsion type can be readily regulated by tuning the number of polymer chains bound to POMs, which, in turn, can be controlled by varying the concentrations of both components and the water/oil ratio. In addition, the number of polymer chains anchored to the POMs can be varied by controlling the number of negative charges on the POMs through an in situ redox reaction. As such, a reversible inversion of the emulsions can be triggered by switching between exposure to ultraviolet light and the introduction of oxygen. Combining the functions of POM itself, a cyclic interfacial catalysis system was realized. Inversion of the emulsion also affords a pathway to high-internal-phase emulsions. The diversity of the POMs, the polymers, and the responsive switching groups open numerous new, simple strategies for designing a wide range of responsive soft matter for cargo loading, controlled release, and delivery in biomedical and engineering applications without time-consuming particle syntheses.

Long-term observations of NO2 using GEMS in China: Validations and regional transport.

In 2019, South Korea launched the Geostationary Environment Monitoring Spectrometer (GEMS) to observe trace gases with an hourly temporal resolution. Compared to previous payloads on polar-orbiting satellites, the GEMS payload has significant advantages in detecting the diurnal variation characteristics of NO2. However, there is still a lack of ground-based validations regarding the overall accuracy of GEMS in the Chinese region. In this study, we conducted a systematic ground validation of GEMS NO2 data in China for the first time. We validated the accuracy of GEMS NO2 data in four typical pollution regions in China, namely the Beijing-Tianjin-Hebei region (JJJ), the Yangtze River Delta region (YRD), the Pearl River Delta region (PRD), and the Sichuan Basin region (SCB), based on MAX-DOAS and CNEMC data. The averaged correlations using the two datasets for validation were 0.81 and 0.57, respectively, indicating a high level of accuracy for the data in China. Using the GEMS seasonal averaged NO2 data, we studied the distribution of NO2 levels in the four regions. We found that the highest NO2 in all four regions occurred during winter with concentrations of 1.84 × 1016 molecules cm-2, 1.59 × 1016 molecules cm-2, 1.58 × 1016 molecules cm-2 and 9.47 × 1015 molecules cm-2, respectively. The distribution of NO2 was closely related to the terrain. Additionally, we observed a significant underestimation issue with TROPOMI, exceeding 30 % in many regions. Based on MAX-DOAS, we investigated the vertical distribution of NO2 in the four regions and found that NO2 was mainly concentrated below 0.5 km. with the HNU station having the lowest concentration, averaging only 2.12 ppb, which was approximately 41 % of the highest concentration recorded at the CQ station. Furthermore, we conducted a study on regional and cross-regional transport using a combination of MAX-DOAS and GEMS data. We found that the transport flux of NO2 could increase by over 500 % within 1 h, making a significant contribution to local NO2 concentrations. The joint observations of GEMS and MAX-DOAS will provide reliable data support for NO2 research and control in China, making a substantial contribution to environmental protection and sustainable development.

Mask-Free Patterned Perovskite Microcavity Arrays via Inkjet Printing Targeting Laser Emission.

Perovskite materials are promising candidates for the implementation of electrically pumped lasers considering the enhanced performance of perovskite-based light-emitting diodes. Nonetheless, current methods of fabricating perovskite optical microcavities require complex patterning technologies to build suitable resonant cavities for perovskite laser emission, burdening the device structure design. To address this issue, we applied inkjet printing, a maskless patterning technique, to directly create spontaneous formations of polycrystalline perovskite microcavity arrays to explore their laser-emitting action. The substrate surface tension was tuned to modulate the perovskite crystallization process in combination with optimization of printing ink recipes. As a result, polycrystalline perovskite microcavity arrays were achieved, contributing to the laser emission at 528 nm with a lasing threshold of 1.37 mJ/cm2, while simultaneously achieving high-definition patterning of flexible display. These results clearly illustrate the efficiency of inkjet printing technology in the preparation of polycrystalline perovskite optical microcavities and promote the development of flexible laser arrayed displays, providing a facile process toward the realization of perovskite-cavity laser devices.

A review of different breast reconstruction methods.

Breast reconstruction is necessary for the comprehensive treatment of breast cancer. For successful breast reconstruction, the timing of surgery and the surgical methods used are vital. The methods of breast reconstruction can be divided into implant-based breast reconstruction (IBBR) and autologous breast reconstruction (ABR). With the development of acellular dermal matrix (ADM), IBBR has become more common in clinical practice. However, the choice for the position in which the implant should be placed (prepectoral or subpectoral) and the use of ADM is currently controversial. We summarized the differences in indications, complications, advantages, disadvantages, and prognosis between IBBR and ABR. We also compared the indications and complications of different flaps in ABR and found that the LD (latissimus dorsi) flap is suitable for Asian women who have a low body mass index (BMI) and a low incidence of obesity, while the DIEP (deep inferior epigastric perforator) flap can be used in patients with severe breast ptosis. In conclusion, immediate breast reconstruction with an implant or expander is the primary method, as it causes lesser scarring and requires a shorter time compared to ABR. However, for patients with severe breast ptosis or reluctant to receive an implant, ABR can be performed for a satisfying cosmetic result. Indications and complications of different flaps in ABR are also inconsistent. Surgeons should make surgical plans based on the preferences and conditions of each patient. In the future, breast reconstruction methods need to be further refined, and minimally invasive and personalized approaches need to be implemented to provide more benefits to patients.

Phosphine-based metal-organic layers to construct single-site heterogeneous catalysts for arene borylation.

Metal-organic layers (MOLs) are versatile platforms for creating single-site heterogeneous catalysts. Incorporating molecular functionalities into MOLs is crucial for catalysis. In this study, we synthesized phosphine-containing MOLs constructed from Hf6-oxo secondary building units (SBUs) and phosphine ligands. The mono(phosphine)-Ir complexes generated by the metalation of TPP-MOL were highly active as heterogeneous catalysts for the C(sp2)-H borylation of a range of arenes. This research expands the diversity of MOL-based catalysts.

An electrochemiluminescence immunosensor based on multipath signal catalytic amplification integrated in a Cu2+-PEI-Pt/AuNC nanocomposite.

Here, the nanocomposite Cu2+-PEI-Pt/AuNCs with multipath signal catalytic amplification for a peroxydisulfate-dissolved oxygen electrochemiluminescence (ECL) system was prepared to fabricate a sensitive ECL immunosensor. Using polyethyleneimine (PEI), a linear polymer, as the reductant and template, Pt/Au nanochains (Pt/AuNCs) were prepared. Abundant PEI would adsorb on the surface of Pt/AuNCs via Pt-N or Au-N bonds, and further coordinate with Cu2+ to give the final nanocomposite Cu2+-PEI-Pt/AuNCs which possessed multipath signal catalytic amplification for the ECL of the peroxydisulfate-dissolved oxygen system in the presence of H2O2. First, PEI, as an effective co-reactant, could directly enhance the ECL intensity. Second, Pt/AuNCs could not only act as a mimicking enzyme to promote the decomposition of H2O2 to produce more oxygen in situ, but also act as an effective co-reaction accelerator to facilitate the generation of more co-reactive intermediate groups from peroxydisulfate, resulting in an obviously enhanced ECL signal. Then, Cu2+ could also accelerate the decomposition of H2O2 to produce more oxygen in situ, leading to a further improvement of the ECL response. Using Cu2+-PEI-Pt/AuNCs as a loading platform, a sandwiched ECL immunosensor was fabricated. As a result, the obtained ECL immunosensor possessed an ultra-sensitive detection performance for α-1-fetoprotein, providing effective information on the diagnosis and treatment of related diseases.

Unraveling Deposition Atmosphere Impact on Reproducibility of Perovskite Light-Emitting Diodes.

Manipulating the crystallization dynamics of perovskite emitters is an effective strategy for preparing high-performance perovskite light-emitting diodes (PeLEDs). In general, amorphous-like thermodynamically stable intermediates are desirable for a retarded and controllable crystallization process of perovskite emitters. Despite a variety of well-demonstrated strategies for crystallization control, it has been generally realized that perovskite thin-film emitters show problematic reproducibility. Here, we unraveled that the coordinating solvent vapor residues could raise deleterious influences on the formation of amorphous intermediate phases, which thus leads to varying crystal qualities from batch to batch. We demonstrated that undesirable crystalline intermediate phases tend to form with a strong coordination solvent vapor atmosphere, which alters the crystallization process and thus brings about additional ionic defects. By applying an inert gas flush strategy, this detrimental effect could be effectively mitigated, enabling PeLEDs with high reproducibility. This work provides new insight into the fabrication of efficient and reproducible perovskite optoelectronics.

Recurrent predictive coding models for associative memory employing covariance learning.

The computational principles adopted by the hippocampus in associative memory (AM) tasks have been one of the most studied topics in computational and theoretical neuroscience. Recent theories suggested that AM and the predictive activities of the hippocampus could be described within a unitary account, and that predictive coding underlies the computations supporting AM in the hippocampus. Following this theory, a computational model based on classical hierarchical predictive networks was proposed and was shown to perform well in various AM tasks. However, this fully hierarchical model did not incorporate recurrent connections, an architectural component of the CA3 region of the hippocampus that is crucial for AM. This makes the structure of the model inconsistent with the known connectivity of CA3 and classical recurrent models such as Hopfield Networks, which learn the covariance of inputs through their recurrent connections to perform AM. Earlier PC models that learn the covariance information of inputs explicitly via recurrent connections seem to be a solution to these issues. Here, we show that although these models can perform AM, they do it in an implausible and numerically unstable way. Instead, we propose alternatives to these earlier covariance-learning predictive coding networks, which learn the covariance information implicitly and plausibly, and can use dendritic structures to encode prediction errors. We show analytically that our proposed models are perfectly equivalent to the earlier predictive coding model learning covariance explicitly, and encounter no numerical issues when performing AM tasks in practice. We further show that our models can be combined with hierarchical predictive coding networks to model the hippocampo-neocortical interactions. Our models provide a biologically plausible approach to modelling the hippocampal network, pointing to a potential computational mechanism during hippocampal memory formation and recall, which employs both predictive coding and covariance learning based on the recurrent network structure of the hippocampus.

Effects of Tea Plant Varieties with High- and Low-Nutrient Efficiency on Nutrients in Degraded Soil.

Tea plants are widely planted in tropical and subtropical regions globally, especially in southern China. The high leaching and strong soil acidity in these areas, in addition to human factors (e.g., tea picking and inappropriate fertilization methods) aggravate the lack of nutrients in tea garden soil. Therefore, improving degraded tea-growing soil is urgently required. Although the influence of biological factors (e.g., tea plant variety) on soil nutrients has been explored in the existing literature, there are few studies on the inhibition of soil nutrient degradation using different tea plant varieties. In this study, two tea plant varieties with different nutrient efficiencies (high-nutrient-efficiency variety: Longjing43 (LJ43); low-nutrient-efficiency variety: Liyou002 (LY002)) were studied. Under a one-side fertilization mode of two rows and two plants, the tea plant growth status, soil pH, and available nutrients in the soil profiles were analyzed, aiming to reveal the improvement of degraded soil using different tea varieties. The results showed that (1) differences in the phenotypic features of growth (such as dry tea yield, chlorophyll, leaf nitrogen (N), phosphorus (P), and potassium (K) content) between the fertilization belts in LJ43 (LJ43-near and LJ43-far) were lower than those in LY002. (2) RDA results showed that the crucial soil nutrient factors which determine the features of tea plants included available P, slowly available K, and available K. Moreover, acidification was more serious near the fertilization belt. The pH of the soil near LJ43 was higher than that near LY002, indicating an improvement in soil acidification. (3) Soil nutrient heterogeneity between fertilization belts in LJ43 (LJ43-near and LJ43-far) was lower than in LY002. In conclusion, the long-term one-side fertilization mode of two rows and two plants usually causes spatial heterogeneities in soil nutrients and aggravates soil acidification. However, LJ43 can reduce the nutrient heterogeneities and soil acidification, which is probably due to the preferential development of secondary roots. These results are helpful in understanding the influence of tea plant variety on improving soil nutrients and provide a relevant scientific reference for breeding high-quality tea varieties, improving the state of degraded soil and maintaining soil health.

Synergistic Solar-Driven Freshwater Generation and Electricity Output Empowered by Wafer-Scale Nanostructured Silicon.

Electricity generation triggered by the ubiquitous water evaporation process provides an intriguing way to harvest energy from water. Meanwhile, natural water evaporation is also a fundamental way to obtain fresh water for human beings. Here, a wafer-scale nanostructured silicon-based device that takes advantage of its well-aligned configuration that simultaneously realizes solar steam generation (SSG) for freshwater collection and hydrovoltaic effect generation for electricity output is developed. An ingenious porous, black carbon nanotube fabric (CNF) electrode endows the device with sustainable water self-pumping capability, excellent durable conductivity, and intense solar spectrum harvesting. A combined device based on the CNF electrode integrated with nanostructured silicon nanowire arrays (SiNWs) provided an aligned numerous surface-to-volume water evaporation interface that enables a recorded continuous short-circuit current 8.65 mA and a water evaporation rate of 1.31 kg m-2 h-1 under one sun illumination. Such wafer-scale SiNWs-based SSG and hydrovoltaic integration devices would unchain the bottleneck of the weak and discontinuous electrical output of hydrovoltaic devices, which inspires other sorts of semiconductor-based hydrovoltaic device designs to target superior performance.

Multimodal Brain Image Fusion Based on Improved Rolling Guidance Filter and Wiener Filter.

Medical image fusion technology can integrate complementary information from different modality medical images to provide a more complete and accurate description of the specific diagnosed object, which is very helpful for image-guided clinical diagnosis and treatment. This paper proposes an effective brain image fusion framework based on improved rolling guidance filter (IRGF). Firstly, input images are decomposed into base layers and detail layers using the IRGF and Wiener filter. Secondly, the visual saliency maps of the input image are computed by pixel-level saliency value, and the weight maps of detail layers are constructed by max-absolute strategy and are further smoothed with Gaussian filter, the purpose of which is to make the fused image appear more naturally and more suitable for human visual perception. Lastly, base layers are fused by visual saliency map based fusion rule and the corresponding weight maps from detail layers are fused by the weighted least squares optimization scheme. Experimental results testify that our method is superior to some state-of-the-art methods in both subjective and objective assessments.

Preparation of Janus Droplets and Hydrogels with Controllable Morphologies by an Aqueous Two-Phase System on the Superamphiphobic Surface.

Janus particles, having the property integration of each component, have attracted increasing attention due to their considerable potential in the field of material engineering applications. However, organic solvents or sophisticated equipment during the fabrication processes is generally inevitable. Here, we report a facile route to prepare Janus droplets and hydrogels via aqueous two-phase systems (ATPS). Simply merging two polymers, i.e., polyethylene glycol (PEG) and dextran (DEX), as aqueous droplets on a superamphiphobic surface leads to phase separation, provided that their concentrations exceed the threshold in the mixed aqueous droplets, thus generating a Janus structure. Various morphologies of such Janus droplets can be well controlled by manipulating the locations of these two polymers' concentration on the phase diagram, and the evolution of the mixed droplets are deterministic on the basis of the kinetics of their phase separation and the degree of hydrophobicity of the substrate. Introducing monomers and/or nanoparticles, further, into a certain phase of the ATPS droplet followed by photopolymerizing enables Janus hydrogel particles with diverse functionalities to be obtained. The ease and green techniques with which the Janus balance and curvature between two phases of the Janus droplet can be finely tuned point to new directions in designing Janus particles and hold great promises in biological engineering.

Vertical distributions and potential sources of wintertime atmospheric pollutants and the corresponding ozone production on the coast of Bohai Sea.

This study investigated the wintertime vertical distributions and source areas of aerosols, NO2, and HCHO in a coastal city of Dongying from December 2020 to March 2021, using ground-based multi-axis differential optical absorption spectroscopy (MAX-DOAS) and a potential source contribution function (PSCF) model, respectively. Moreover, the chemical production sensitivity of O3 at different height layers was analyzed using HCHO/NO2 ratios. The results revealed that the wintertime averaged highest concentrations of aerosol (1.25 km-1), NO2 (14.81 ppb), and HCHO (2.32 ppb) were mainly distributed at the surface layer, 100-200 m layer, and 200-300 m layer, respectively. Regarding the diurnal cycles, high concentrations of aerosol (>1.4 km-1) and NO2 (>16.0 ppb) usually appeared in the early morning and late afternoon, while high concentrations of HCHO (>2.5 ppb) usually occurred during 12:00-15:00. The PSCF model revealed that the wintertime aerosol mainly originated from Shandong, northern Jiangsu, Korea, and the northwestern Mongolian Plateau. Below 200 m, NO2 was mainly from western Shandong, whereas above 600 m, it was mainly from northern Shandong and the Beijing-Tianjin-Hebei (BTH) region. The corresponding sources for HCHO were central and southern Shandong (below 200 m) and northern Shandong, northern Jiangsu, and southeastern BTH (above 600 m). In addition, the chemical production sensitivity of O3 below 100 m was observed only in the VOC-limited regime. The percentages of O3 production under the NOx-limited, NOx-VOC-limited, and VOC-limited regimes were 10.75% (31.18%), 4.30% (19.35%), and 84.95% (49.47%) at the 500-600 m (900-1000 m) layer. This study has guiding significance for the coordinated control of PM2.5 and O3, and can assist in the implementation of regional joint prevention and control strategies for air pollutants.

Cationic P,O-Coordinated Nickel(II) Catalysts for Carbonylative Polymerization of Ethylene: Unexpected Productivity via Subtle Electronic Variation.

Transition-metal-catalyzed copolymerization of ethylene with carbon monoxide affords polyketones materials with excellent mechanical strength, photodegradability, surface and barrier properties. Unlike the widely used and rather expensive Pd catalysts, Ni-catalyzed carbonylative polymerization is very difficult since the strong binding affinity of CO to Ni deactivates the highly electrophilic metal center easily. In this study, various cationic P,O-coordinated Ni complexes were synthesized using the electronic modulation strategy, and the catalyst with strong electron-donating substituents exhibits an excellent productivity of 104  g polymer (g Ni)-1 , which represents a rare discovery that a Ni complex could operate with such exceptional efficiency in comparison with Pd catalysts. Notably, those Ni catalysts were also efficient for terpolymerization of ethylene, propylene with CO for producing commercial polyketone materials with low melting temperatures and easy processibility.

Learning on Arbitrary Graph Topologies via Predictive Coding.

Training with backpropagation (BP) in standard deep learning consists of two main steps: a forward pass that maps a data point to its prediction, and a backward pass that propagates the error of this prediction back through the network. This process is highly effective when the goal is to minimize a specific objective function. However, it does not allow training on networks with cyclic or backward connections. This is an obstacle to reaching brain-like capabilities, as the highly complex heterarchical structure of the neural connections in the neocortex are potentially fundamental for its effectiveness. In this paper, we show how predictive coding (PC), a theory of information processing in the cortex, can be used to perform inference and learning on arbitrary graph topologies. We experimentally show how this formulation, called PC graphs, can be used to flexibly perform different tasks with the same network by simply stimulating specific neurons. This enables the model to be queried on stimuli with different structures, such as partial images, images with labels, or images without labels. We conclude by investigating how the topology of the graph influences the final performance, and comparing against simple baselines trained with BP.

Reverse Differentiation via Predictive Coding.

Deep learning has redefined AI thanks to the rise of artificial neural networks, which are inspired by neuronal networks in the brain. Through the years, these interactions between AI and neuroscience have brought immense benefits to both fields, allowing neural networks to be used in a plethora of applications. Neural networks use an efficient implementation of reverse differentiation, called backpropagation (BP). This algorithm, however, is often criticized for its biological implausibility (e.g., lack of local update rules for the parameters). Therefore, biologically plausible learning methods that rely on predictive coding (PC), a framework for describing information processing in the brain, are increasingly studied. Recent works prove that these methods can approximate BP up to a certain margin on multilayer perceptrons (MLPs), and asymptotically on any other complex model, and that zerodivergence inference learning (Z-IL), a variant of PC, is able to exactly implement BP on MLPs. However, the recent literature shows also that there is no biologically plausible method yet that can exactly replicate the weight update of BP on complex models. To fill this gap, in this paper, we generalize (PC and) Z-IL by directly defining it on computational graphs, and show that it can perform exact reverse differentiation. What results is the first PC (and so biologically plausible) algorithm that is equivalent to BP in the way of updating parameters on any neural network, providing a bridge between the interdisciplinary research of neuroscience and deep learning. Furthermore, the above results in particular also immediately provide a novel local and parallel implementation of BP.

Graphene-embedded oblique V-shaped silver nanoarrays for hydrophobic pollutants pre-concentration and high-sensitivity SERS detection.

A surface enhanced Raman scattering (SERS) substrate of silver nanorod modified with graphene and silver nanorod (AgNR@Graphene@AgNR) has been fabricated to improve the sensitivity of SERS detection of hydrophobic pollutants, in which, graphene is an interlayer and AgNR is arranged on both sides of the graphene. The embedded graphene could help the oblique V-shaped AgNR structure to improve the sensitivity of SERS detection with a significant electric field enhancement effect. The annealing treatment of the substrate, shortening the nanometer gap between the graphene and AgNR, is benefit for producing a large number of "hot spots" at the fold, which has been verified by the finite difference time domain (FDTD) simulation. The enhancement factor (EF) of AgNR@Graphene@AgNR could reach up to 1.6 × 108 with a good reproducibility. The substrate could achieve high-sensitivity detection of 4-chlorobiphenyl (PCB-3) and 3, 3', 4, 4'-tetrachlorobiphenyl (PCB-77) with the limit of detections (LODs) of 1.72 × 10-10 M and 2.11 × 10-8 M, and the effective identification of PCBs mixture has been realized through principal component analysis (PCA), which means that the AgNR@Graphene@AgNR substrate has a potential significance for the detection and analysis of hydrophobic pollutant mixtures in the environment.