Bidirectional Transfer Research Articles

Hybrid-supervised bidirectional transfer networks for computer-aided diagnosis

Medical imaging techniques have been widely used for diagnosis of various diseases. However, the imaging-based diagnosis generally depends on the clinical skill of radiologists. Computer-aided diagnosis (CAD) can help radiologists improve diagnostic accuracy as well as the consistency and reproducibility. Although convolutional neural network (CNN) has shown its feasibility and effectiveness in CAD, it generally suffers from the problem of small sample size when training CAD models. Nowadays, self-supervised learning (SSL) has shown its effectiveness in the field of medical image analysis, especially when there are only limited training samples. However, the backbone of downstream task sometimes cannot be well pre-trained in the conventional SSL framework due to the limitation of the pretext task and fine-tuning mechanism. In this work, an improved SSL framework, named Hybrid-supervised Bidirectional Transfer Networks (HBTN), is proposed to improve the performance of CAD models. Specifically, a novel Gray-Scale Image Mapping (GSIM) task is developed, which still takes the widely used image restoration task in SSL as the pretext task, but further embeds the class label information into it to improve discriminative feature learning of its corresponding network model. The proposed HBTN then integrates two different network architectures, i.e. the image restoration network for the pretext task and the classification network for the downstream task, into a unified hybrid-supervised learning (HSL) framework. It jointly trains both networks and collaboratively transfers the knowledge between each other. Consequently, the performance of downstream network is thus improved. The proposed HBTN is evaluated on two medical image datasets for CAD tasks. The experimental results indicate that HBTN outperforms the conventional SSL algorithms for CAD with limited training samples.

(Nanocarbons Division Robert C. Haddon Research Award) Nano Carbons in Mimicking the Early Events of Photosynthetic Reaction Center: ‘Simple’ and ‘Not-so-Simple’ Donor-Acceptor Conjugates

Light-induced electron transfer is a key step in solar energy conversion, and a process with myriad relevance in the fields of artificial photosynthesis, photocatalysis, photoconductivity, and photovoltaics. In a donor-acceptor, D-A system, light absorption by a chromophore results in the formation of an excited state species, followed by a charge transfer between an electron donor (D) and an electron acceptor (A) leading to a charge-separated D .+ -A .- species.The optimization of these processes and comprehending their mutual interplay to ultimately achieve charge-separated species is of paramount importance for the development of efficient systems in solar light conversion schemes.The present talk summarizes our research over the years focused on utilizing nano carbons, viz., fullerenes, endohedral fullerenes, carbon nanotubes, and graphene in the construction of donor-acceptor systems capable of undergoing photoinduced energy and electron transfer events. Both, covalently linked with rigid, semi-rigid, flexible, or conjugated spacers, and self-assembled via metal-ligand axial coordination, ion-pairing, ion-dipole, or hydrogen bonding strategies were utilized in the construction of donor-acceptor conjugates. Multi-modular donor-acceptor constructs were designed, synthesized, and studied to probe sequential energy and electron transfer events, finally resulting in long-lived charge-separated states. In some instances, these molecular systems were also used as photocatalytic energy conversion schemes. The pros and cons of utilizing nanocarbon materials as photo-active functional materials will be highlighted. D’Souza, F.; Sandanayaka, A. S. D.; Ito, O. ‘SWNT-Based Supramolecular Nanoarchitectures with Photosensitizing Donor and Acceptor Molecules’ Phys. Chem. Letts. 2010, 1. 2586-2593.D’Souza, and O. Ito ‘Photosensitized electron transfer processes of nano carbons applicable to solar cells,’ Chem. Soc. Rev. 2012, 41, 86-96.B. KC, F. D’Souza, ‘Design and photochemical study of supramolecular donor-acceptor systems assembled via metal-ligand axial coordination Coord. Chem. Rev. 2016, 322, 104-141.Bottari, M. A. Herranz, L. Wibmer, M. Vollard, L. Rodriguez-Perez, D. M. Guldi, A. Hirsch, N. Martin, F. D’Souza, T. Torres, Chemical functionalization and characterization of graphene-based materials, Chem. Soc. Rev. 2017, 46, 4464-4500.Barrejon, L. M. Arellano, F. D’Souza, F. Langa, ‘Bidirectional charge transfer in carbon-based hybrid nanomaterials’ Nanoscale, 2019, 11, 14978-14992.Kazemi, Y. Jang, A. Liyanage, P. A. Karr, and F. D’Souza, ‘A Carbon Nanotube Binding Bis(pyrenylstyryl)BODIPY-C60 Nano Tweezer: Charge Stabilization through Sequential Electron Transfer’, Angew. Chem. Int. Ed. 2022, in press

Voltage-Oriented Control-Based Three-Phase, Three-Leg Bidirectional AC–DC Converter with Improved Power Quality for Microgrids

Renewable energy sources (RESs) and energy storage schemes (ESSs) integrated into a microgrid (MG) system have been widely used in power generation and distribution to provide a constant supply of electricity. The power electronics converters, particularly the bidirectional power converters (BPCs), are promising interfaces for MG infrastructure because they control the power management of the whole MG system. The controller of BPCs can be designed using several different control strategies. However, all the existing controllers have system stability, dynamics, and power quality issues. Therefore, this study demonstrates the development of an LCL-filtered grid-connected bidirectional AC–DC converter’s (BADC) control strategy based on voltage-oriented control (VOC) to overcome these issues. The proposed VOC-based inner current control loop (ICCL) is implemented in synchronous dq-coordinate with the help of proportional-integral (PI) controllers. An observer-based active damping (AD) is also developed in order to estimate the filter capacitor current from the capacitor voltage instead of directly measuring it. This developed AD system helps to damp the resonance effect of the LCL filter, improves system stability, and also eliminates the practical challenges of measuring capacitor current. The proposed controller with AD is able to realize bidirectional power transfer (BPT) with reduced power losses due to the elimination of passive damping and improved power quality, system dynamics, and stability. The mathematical modeling of the suggested system was developed, and the structure of the system model was established in the MATLAB/Simulink environment. The performance of the proposed system was validated with real-time software-in-the-loop (RT-SIL) simulation using the OPAL-RT simulator for a 16 kVA converter system. The real-time (RT) simulation results show that the BADC with the proposed control scheme can provide better dynamic performance and operate with tolerable total harmonic distortion (THD) of 2.62% and 2.71% for inverter and rectifier modes of operation, respectively.

Bi-directional controlled asymmetric teleportation protocol initiated by a mentor in noisy environments

In this paper we consider the problem of the bidirectional state transfer of a 2-qubit and a 3-qubit quantum state unknown to the parties, respectively, who intend to send the respective states in the site of the other party. These states are intended for transfer but are not possessed by the intender. The protocol has a Mentor whose role is in the fabrication of entangled channels, distribution of qubits through the noisy environment in order to create entanglement sharing and finally fixing one of the several courses of the protocol. The role of the Mentor is thereby completed. The protocol is overseen by a controller whose action signals the final steps to be taken in the protocol. Three types of noises are considered through three different Kraus operators. There is an analysis of the fidelity of the protocol corresponding to three types of noises.

Modulating individual axons and axonal populations in the peripheral nerve using transverse intrafascicular multichannel electrodes.

Objective.A transverse intrafascicular multichannel electrode (TIME) may offer advantages over more conventional cuff electrodes including higher spatial selectivity and reduced stimulation charge requirements. However, the performance of TIME, especially in the context of non-conventional stimulation waveforms, remains relatively unexplored. As part of our overarching goal of investigating stimulation efficacy of TIME, we developed a computational toolkit that automates the creation and usage ofin siliconerve models with TIME setup, which solves nerve responses using cable equations and computes extracellular potentials using finite element method.Approach.We began by implementing a flexible and scalable Python/MATLAB-based toolkit for automatically creating models of nerve stimulation in the hybrid NEURON/COMSOL ecosystems. We then developed a sciatic nerve model containing 14 fascicles with 1,170 myelinated (A-type, 30%) and unmyelinated (C-type, 70%) fibers to study fiber responses over a variety of TIME arrangements (monopolar and hexapolar) and stimulation waveforms (kilohertz stimulation and cathodic ramp modulation).Main results.Our toolkit obviates the conventional need to re-create the same nerve in two disparate modeling environments and automates bi-directional transfer of results. Our population-based simulations suggested that kilohertz stimuli provide selective activation of targeted C fibers near the stimulating electrodes but also tended to activate non-targeted A fibers further away. However, C fiber selectivity can be enhanced by hexapolar TIME arrangements that confined the spatial extent of electrical stimuli. Improved upon prior findings, we devised a high-frequency waveform that incorporates cathodic DC ramp to completely remove undesirable onset responses.Conclusion.Our toolkit allows agile, iterative design cycles involving the nerve and TIME, while minimizing the potential operator errors during complex simulation. The nerve model created by our toolkit allowed us to study and optimize the design of next-generation intrafascicular implants for improved spatial and fiber-type selectivity.

Bidirectional Blockchain-Based Secure Data Transfer Using GALOEEM Routing Protocol in WSN

Bidirectional Blockchain-Based Secure Data Transfer Using GALOEEM Routing Protocol in WSN

Artificial-Intelligence-Based Triple Phase Shift Modulation for Dual Active Bridge Converter With Minimized Current Stress

The dual active bridge (DAB) converter has been popular in many applications for its outstanding power density and bidirectional power transfer capacity. Up to now, triple phase shift (TPS) can be considered as one of the most advanced modulation techniques for DAB converter. It can widen zero voltage switching range and improve power efficiency significantly. Currently, current stress of the DAB converter has been an important performance indicator when TPS modulation is applied for smaller size and higher efficiency. However, to minimize the current stress when the DAB converter is under TPS modulation, two difficulties exist in analysis process and realization process, respectively. Firstly, three degrees of modulation variables in TPS modulation bring challenges to the analysis of current stress in different operating modes. This analysis and deduction process leads to heavy computational burden and also suffers from low accuracy. Secondly, to realize TPS modulation, if a lookup table is adopted after the optimization of modulation variables, modulation performance will be unsatisfactory because of the discrete nature of lookup table. Therefore, an AI-based TPS modulation (AI-TPSM) strategy is proposed in this paper. Neural network (NN) and fuzzy inference system (FIS) are utilized to deal with the two difficulties mentioned above. With the proposed AI-TPSM, the optimization of TPS modulation for minimized current stress will enjoy high degree of automation which can relieve engineers’ working burden and improve accuracy. In the end of this paper, the effectiveness of the proposed AI-TPSM has been experimentally verified with a 1 kW prototype.

Presence of Human DNA on Household Dogs and Its Bi-Directional Transfer.

Awareness of the factors surrounding the transfer of DNA from a person, item, or surface to another person, item, or surface is highly relevant during investigations of alleged criminal activity. Animals in domestic environments could be a victim, offender, or innocent party associated with a crime. There is, however, very limited knowledge of human DNA transfer, persistence, prevalence, and recovery (DNA TPPR) associated with domestic animals. This pilot study aimed to improve our understanding of DNA TPPR associated with domestic dogs by collecting and analysing samples from various external areas of dogs of various breeds, interactions with humans, and living arrangements, and conducting a series of tests to investigate the possibility of dogs being vectors for the indirect transfer of human DNA. Reference DNA profiles from the dog owners and others living in the same residence were acquired to assist interpretation of the findings. The findings show that human DNA is prevalent on dogs, and in the majority of samples, two-person mixtures are present. Dogs were also found to be vectors for the transfer of human DNA, with DNA transferred from the dog to a gloved hand during patting and a sheet while walking.

Intracellular gene transfer and codon usage of cytoplasmic genomes in green plants

Abstract Intracellular gene transfer is widely recognized as one of the most important driving forces for species evolution. Here we investigated transferred cytoplasmic motifs in green plants including spore-bearing plants and seed-bearing plants (hereafter termed spore plants and seed plants) . Our analyses revealed that gene transfer in spore plants was characterized by shorter motifs than that of seed plants. Several spore species did not exhibit intracellular gene transfer. Meanwhile, high frequency transferred tRNA genes were identified with average values of minimum free energy at moderate level. From the chloroplast to the mitochondrial genome, trnP was found to have transferred with high frequency in green plants. In gene transfer from the mitochondrial to the chloroplast genome, trnN was found to be a highly transferred gene. We observed that several tRNA genes including trnF, trnW, and trnN were involved in bidirectional transfer, which may be related to application strategy of functional protein-coding genes in a plant’s adaptive evolution. Codon Adaptation Index (CAI) analysis showed that codon usage was unbalanced in spore and seed plants. CAI values for seed plants were higher than those for spore plants in general, which may reveal rapid divergence adaptability of codon usage in the former. These results provide novel insights into gene transfer and codon usage within cytoplasmic genomes.

Iodanyl Radical Catalysis.

ConspectusHypervalent iodine reagents find application as selective chemical oxidants in a diverse array of oxidative transformations. The utility of these reagents is often ascribed to (1) the proclivity to engage being selective two-electron redox transformations; (2) facile ligand exchange at the three-centered, four-electron (3c-4e) hypervalent iodine-ligand (I-X) bonds; and (3) the hypernucleofugacity of aryl iodides. One-electron redox and iodine radical chemistry is well-precedented in the context of inorganic hypervalent iodine chemistry─for example, in the iodide-triiodide couple that drives dye-sensitized solar cells. In contrast, organic hypervalent iodine chemistry has historically been dominated by the two-electron I(I)/I(III) and I(III)/I(V) redox couples, which results from intrinsic instability of the intervening odd-electron species. Transient iodanyl radicals (i.e., formally I(II) species), generated by reductive activation of hypervalent I-X bonds, have recently gained attention as potential intermediates in hypervalent iodine chemistry. Importantly, these open-shell intermediates are typically generated by activation of stoichiometric hypervalent iodine reagents, and the role of the iodanyl radical in substrate functionalization and catalysis is largely unknown.Our group has been interested in advancing the chemistry of iodanyl radicals as intermediates in the sustainable synthesis of hypervalent I(III) and I(V) compounds and as novel platforms for substrate activation at open-shell main-group intermediates. In 2018, we disclosed the first example of aerobic hypervalent iodine catalysis by intercepting reactive intermediates in aldehyde autoxidation chemistry. While we initially hypothesized that the observed oxidation was accomplished by aerobically generated peracids via a two-electron I(I)-to-I(III) oxidation reaction, detailed mechanistic studies revealed the critical role of acetate-stabilized iodanyl radical intermediates. We subsequently leveraged these mechanistic insights to develop hypervalent iodine electrocatalysis. Our studies resulted in the identification of new catalyst design principles that give rise to highly efficient organoiodide electrocatalysts that operate at modest applied potentials. These advances addressed classical challenges in hypervalent iodine electrocatalysis related to the need for high applied potentials and high catalyst loadings. In some cases, we were able to isolate the anodically generated iodanyl radical intermediates, which allowed direct interrogation of the elementary chemical reactions characteristic of iodanyl radicals. Both substrate activation via bidirectional proton-coupled electron transfer (PCET) reactions at I(II) intermediates and disproportionation reactions of I(II) species to generate I(III) compounds have been experimentally validated.This Account discusses the emerging synthetic and catalytic chemistry of iodanyl radicals. Results from our group have demonstrated that these open-shell species can play a critical role in sustainable synthesis of hypervalent iodine reagents and play a heretofore unappreciated role in catalysis. Realization of I(I)/I(II) catalytic cycles as a mechanistic alternative to canonical two-electron iodine redox chemistry promises to open new avenues to application of organoiodides in catalysis.

Bidirectional electron transfer in Triboelectrification caused by friction-induced change in surface electronic structure

Triboelectrification (TE) plays fundamental roles in a wide range of industrial and scientific fields but its charge transfer mechanisms still remains elusive. It has been widely believed that the electron transfer in TE between two homogeneous surfaces should be always unidirectional. But we found here that the electron transfer during friction process will change its direction as the position of the contact point on the friction surface changes, indicating that bidirectional electron transfer occurs between the contacting surfaces. The experiments further show that the electron transfer direction at any given contact position on the friction surface keeps unchanged over time during the whole friction process if the relative sliding direction keeps unchanged, and when the sliding direction is reversed, the electron transfer direction at a given contact position will also be reversed if the friction surface is crystalline. To explain the observed non-unidirectional electron transfer, a theoretical model for the electron transfer in TE has been proposed, assuming that the changes in the electron transfer direction can be attributed to the changes in the surface work function caused by the friction-induced tangential deformation on the friction surface. The theory has further been validated by the first principles calculations about the influence of the tangential deformation on the surface electronic structure. The present work may offer a new perspective for understanding the charge transfer mechanism in TE and may provide potential applications in numerous fields involving TE such as triboelectric generator, petrochemical industry, power electronics and textile industry.

Multidimensional Mission-Profile-Based Lifetime Estimation Approach for IGBT Modules in MMC–HVdc Application Considering Bidirectional Power Transfer

MMC applied in high voltage direct current (HVdc) field consists of numerous submodules (SMs) and supports bidirectional power transfer, leading to the diverse mission and temperature profiles among SMs and IGBT modules. In this paper, an IGBT lifetime estimation approach in terms of the bidirectional power transfer of MMC is proposed. The temperature profile of each module is translated based on the multi-dimensional mission profile, including arm current, heatsink temperature, active power, and reactive power. In this way, the average junction temperature and junction temperature fluctuation of each SM can be readily estimated according to the power transfer direction. Furthermore, both the line-frequency and low-frequency fatigue are considered in the estimation process by using corresponding lifetime models. The proposed lifetime estimation method is compared with the traditional power loss average method and one-direction power transfer method by the application of the multi-dimensional mission profile of the practical Nanao project. The results show the bidirectional power transfer has a big influence on the damage distribution inside SM. The lower IGBT, upper FWD, and lower FWD all suffer considerable degradation under bidirectional power transfer conditions. Moreover, bond wire fatigue is the dominant degradation mechanism in MMC-HVdc application.

Lifelong Person Re-identification via Knowledge Refreshing and Consolidation

Lifelong person re-identification (LReID) is in significant demand for real-world development as a large amount of ReID data is captured from diverse locations over time and cannot be accessed at once inherently. However, a key challenge for LReID is how to incrementally preserve old knowledge and gradually add new capabilities to the system. Unlike most existing LReID methods, which mainly focus on dealing with catastrophic forgetting, our focus is on a more challenging problem, which is, not only trying to reduce the forgetting on old tasks but also aiming to improve the model performance on both new and old tasks during the lifelong learning process. Inspired by the biological process of human cognition where the somatosensory neocortex and the hippocampus work together in memory consolidation, we formulated a model called Knowledge Refreshing and Consolidation (KRC) that achieves both positive forward and backward transfer. More specifically, a knowledge refreshing scheme is incorporated with the knowledge rehearsal mechanism to enable bi-directional knowledge transfer by introducing a dynamic memory model and an adaptive working model. Moreover, a knowledge consolidation scheme operating on the dual space further improves model stability over the long-term. Extensive evaluations show KRC’s superiority over the state-of-the-art LReID methods with challenging pedestrian benchmarks. Code is available at https://github.com/cly234/LReID-KRKC.

Fast quantum transfer mediated by topological domain walls

The duration of bidirectional transfer protocols in 1D topological models usually scales exponentially with distance. In this work, we propose transfer protocols in multidomain SSH chains and Creutz ladders that lose the exponential dependence, greatly speeding up the process with respect to their single-domain counterparts, reducing the accumulation of errors and drastically increasing their performance, even in the presence of symmetry-breaking disorder. We also investigate how to harness the localization properties of the Creutz ladder---with two localized modes per domain wall---to choose the two states along the ladder that will be swapped during the transfer protocol, without disturbing the states located in the intermediate walls between them. This provides a 1D network with all-to-all connectivity that can be helpful for quantum information purposes.

MI-CAT: A transformer-based domain adaptation network for motor imagery classification

Due to its convenience and safety, electroencephalography (EEG) data is one of the most widely used signals in motor imagery (MI) brain–computer interfaces (BCIs). In recent years, methods based on deep learning have been widely applied to the field of BCIs, and some studies have gradually tried to apply Transformer to EEG signal decoding due to its superior global information focusing ability. However, EEG signals vary from subject to subject. Based on Transformer, how to effectively use data from other subjects (source domain) to improve the classification performance of a single subject (target domain) remains a challenge. To fill this gap, we propose a novel architecture called MI-CAT. The architecture innovatively utilizes Transformer’s self-attention and cross-attention mechanisms to interact features to resolve differential distribution between different domains. Specifically, we adopt a patch embedding layer for the extracted source and target features to divide the features into multiple patches. Then, we comprehensively focus on the intra-domain and inter-domain features by stacked multiple Cross-Transformer Blocks (CTBs), which can adaptively conduct bidirectional knowledge transfer and information exchange between domains. Furthermore, we also utilize two non-shared domain-based attention blocks to efficiently capture domain-dependent information, optimizing the features extracted from the source and target domains to assist in feature alignment. To evaluate our method, we conduct extensive experiments on two real public EEG datasets, Dataset IIb and Dataset IIa, achieving competitive performance with an average classification accuracy of 85.26% and 76.81%, respectively. Experimental results demonstrate that our method is a powerful model for decoding EEG signals and facilitates the development of the Transformer for brain–computer interfaces (BCIs).

Mode Analysis, Transformer Saturation, and Fault Diagnosis Technique for an Open-Circuit Fault in a Three-Phase DAB Converter

The three-phase Dual Active Bridge (DAB3) is a popular dc-dc converter topology for high-power applications, capable of high-efficiency bidirectional power transfer with galvanic isolation. A single point of failure in power converters is the open-circuit fault (OCF) due to failure in a semiconductor device or its gate drive circuit. This study presents detailed waveform analyses for the normal and the fault mode operation. A novel logic-based fault diagnosis scheme is proposed based on the unique pattern in the dc bias of phase currents. Unlike previous schemes, the proposed scheme requires low-bandwidth current sensing only one side of the transformer to detect faults on either side, providing a cost and design benefit. Experimental results verify the analyses and the proposed identification scheme, detecting the fault within a few switching cycles. An in-depth study of the transformer under fault mode is presented for the first time, setting a guideline of time available for fault diagnosis and response. Experimental B-H curves and magnetizing currents of the three-phase transformer illustrate the cycle-by-cycle progression toward core saturation under fault mode. The study also reveals a new potential benefit of the three-phase DAB over the single-phase DAB; i.e., even in the presence of a secondary-side OCF, the DAB3 may continue to operate normally at full load.

Microbial electrosynthetic nitrate reduction to ammonia by reversing the typical electron transfer pathway in Shewanella oneidensis

Ammonia production is a critical industrial process, and mild routes to recycle nitrates in wastewater could be a promising route to ammonia synthesis. In this study, ammonia production is demonstrated in a microbial electrosynthesis system with nitrate and an electrode as electron acceptors and donors, respectively. Based on the bidirectional extracellular electron transfer capability of Shewanella oneidensis MR-1, our microbial electrosynthetic system achieves a maximum ammonia production rate of 24.3 μg h−1·mg protein−1 with 82.5% selectivity and 33.1% cathodic efficiency and functions for several cycles over 30 days. Electrochemical analysis suggests that cytochromes c, flavins, and the flavin/c-cytochrome combination play a pivotal role. Charge transfer resistance weakens over the course of weeks, resulting in easier electron transfer. Parallel reaction monitoring proteomics suggest that reversing a typical “Mtr pathway” plays a role, and a dissimilatory nitrate to ammonia pathway is used. This work proposes a progressive route to carry out ammonia synthesis under mild conditions.

Research on multi-color OLED with bi-directional charge transfer using Ag: Mg as intermediate electrode

Research on multi-color OLED with bi-directional charge transfer using Ag: Mg as intermediate electrode

Bidirectional inductive underwater wireless power transfer converter with variable structure

The undersea sensor networks have an important function in marine resource exploration. To prolong the lifetime of the sensor networks, a bidirectional inductive underwater wireless power transfer (BI-UWPT) system with variable structure is proposed. With the proposed system, the autonomous underwater vehicle (AUV) could receive energy from the undersea docking station and transmit energy to the undersea sensor networks. The two full bridge topologies and the SS compensation networks are adopted in the system to achieve the bidirectional power transfer. To achieve bidirectional energy transmission of different power levels, two different operating frequencies and variable compensation capacitors are employed by the AUV. The system mathematical model is also presented. To verify the theoretical analysis, a 432 W simulation model when the AUV receiving electric energy and a 96 W simulation model when the AUV transmitting electric energy are constructed, respectively. The theoretical analysis is verified by the simulation results. Simulation results show that the variable compensation networks can obtain high transfer efficiency of the BI-UWPT system.

Serratia marcescens Colonization in a Neonatal Intensive Care Unit Has Multiple Sources, with Sink Drains as a Major Reservoir.

Compelling evidence suggests a contribution of the sink environment to the transmission of opportunistic pathogens from the hospital environment to patients in neonatal intensive care units (NICU). In this study, the distribution of the opportunistic pathogen Serratia marcescens in the sink environment and newborns in a NICU was investigated. More than 500 sink drain and faucet samples were collected over the course of five sampling campaigns undertaken over 3 years. Distribution and diversity of S. marcescens were examined with a modified MacConkey medium and a high-throughput short-sequence typing (HiSST) method. Sink drains were an important reservoir of S. marcescens, with an average of 44% positive samples, whereas no faucet sample was positive. The genotypic diversity of S. marcescens was moderate, with an average of two genotypes per drain, while the spatial distribution of S. marcescens was heterogeneous. The genotypic profiles of 52 clinical isolates were highly heterogeneous, with 27 unique genotypes, of which 71% of isolates were found in more than one patient. S. marcescens acquisition during the first outbreaks was mainly caused by horizontal transmissions. HiSST analyses revealed 10 potential cases of patient-to-patient transmission of S. marcescens, five cases of patient-to-sink transmission, and one bidirectional transfer between sink and patient. Environmental and clinical isolates were found in sink drains up to 1 year after the first detection, supporting persisting drain colonization. This extensive survey suggests multiple reservoirs of S. marcescens within the NICU, including patients and sink drains, but other external sources should also be considered. IMPORTANCE The bacterium Serratia marcescens is an important opportunistic human pathogen that thrives in many environments, can become multidrug resistant, and is often involved in nosocomial outbreaks in neonatal intensive care units (NICU). We evaluated the role of sinks during five suspected S. marcescens outbreaks in a NICU. An innovative approach combining molecular and culture methods was used to maximize the detection and typing of S. marcescens in the sink environment. Our results indicate multiple reservoirs of S. marcescens within the NICU, including patients, sink drains, and external sources. These results highlight the importance of sinks as a major reservoir of S. marcescens and potential sources of future outbreaks.