Real-time Approach Research Articles

A Novel Real-Time Control Approach for Sparse and Safe Frequency Regulation in Inverter Intensive Microgrids

A Novel Real-Time Control Approach for Sparse and Safe Frequency Regulation in Inverter Intensive Microgrids

A real-time application-based convolutional neural network approach for tomato leaf disease classification

A real-time application-based convolutional neural network approach for tomato leaf disease classification

Nonlinear Optical Responses of Janus MoSSe/MoS2 Heterobilayers Optimized by Stacking Order and Strain.

Nonlinear optical responses in second harmonic generation (SHG) of van der Waals heterobilayers, Janus MoSSe/MoS2, are theoretically optimized as a function of strain and stacking order by adopting an exchange-correlation hybrid functional and a real-time approach in first-principles calculation. We find that the calculated nonlinear susceptibility, χ(2), in AA stacking (550 pm/V) becomes three times as large as AB stacking (170 pm/V) due to the broken inversion symmetry in the AA stacking. The present theoretical prediction is compared with the observed SHG spectra of Janus MoSSe/MoS2 heterobilayers, in which the peak SHG intensity of AA stacking becomes four times as large as AB stacking. Furthermore, a relatively large, two-dimensional strain (4%) that breaks the C3v point group symmetry of the MoSSe/MoS2, enhances calculated χ(2) values for both AA (900 pm/V) and AB (300 pm/V) stackings 1.6 times as large as that without strain.

Enhancing Sensor-Based Mobile User Authentication in a Complex Environment by Deep Learning

With the advent of smart mobile devices, end users get used to transmitting and storing their individual privacy in them, which, however, has aroused prominent security concerns inevitably. In recent years, numerous researchers have primarily proposed to utilize motion sensors to explore implicit authentication techniques. Nonetheless, for them, there are some significant challenges in real-world scenarios. For example, depending on the expert knowledge, the authentication accuracy is relatively low due to some difficulties in extracting user micro features, and noisy labels in the training phrase. To this end, this paper presents a real-time sensor-based mobile user authentication approach, ST-SVD, a semi-supervised Teacher–Student (TS) tri-training algorithm, and a system with client–server (C-S) architecture. (1) With S-transform and singular value decomposition (ST-SVD), we enhance user micro features by transforming time-series signals into 2D time-frequency images. (2) We employ a Teacher–Student Tri-Training algorithm to reduce label noise within the training sets. (3) To obtain a set of robust parameters for user authentication, we input the well-labeled samples into a CNN (convolutional neural network) model, which validates our proposed system. Experimental results on large-scale datasets show that our approach achieves authentication accuracy of 96.32%, higher than the existing state-of-the-art methods.

PrunedYOLO-Tracker: An efficient multi-cows basic behavior recognition and tracking technique

Recognition and tracking of basic cow behaviors in natural environments are critical technologies for cows' health monitoring in smart farming. This study proposes an efficient cow behavior recognition and tracking method (PrunedYOLO-Tracker) to address the main problem of the lack of a real-time and identity-linked approach for analyzing and monitoring cow behaviors. Firstly, the channel pruning algorithm is employed to compress the size and parameters of the base model, YOLO v5l. Secondly, a multi-object tracking (MOT) method, Cascaded-Buffered IoU (C-BIoU), which expands the detection and trajectory matching space by increasing the buffer zone, is proposed to combine the behavior information from detection with the trajectory information from tracking, achieving multi-cows' behavior recognition and tracking. Through experimental verification, the pruned model maintains high detection accuracy while reducing the model size, floating-point operations (FLOPs) and parameters by 73.5%, 76.7% and 74.0%, respectively. Compared to the original model, the pruned model only experiences a slight decrease of 0.2% in F1 score, while achieving a 0.3% increase in mean Average Precision (mAP). In terms of cow tracking performance, when compared to six other multi-object tracking algorithms including DeepSort, DeepMot, OC-Sort, BotSort, StrongSort, and ByteTrack, the proposed method demonstrates the highest High Order Tracking Accuracy (HOTA), Multi-Object Tracking Precision (MOTP) and Identification F1 (IDF1) scores, reaching 72.4%, 86.1% and 80.3%, respectively. The results obtained from testing in multiple cow activity environments demonstrate that the proposed method exhibits excellent performance in behavior recognition and tracking, as well as high-speed processing capabilities, with an average video processing speed of 81 frames per second (FPS). This method possesses the ability to reliably monitor and manage cow behavior in real-time, providing technical support for promptly anomalies detection and cow health status monitoring for dairy farming managers.

Learning human inverse kinematics solutions for redundant robotic upper-limb rehabilitation

This paper proposes a real-time human inverse kinematics approach based on the Gaussian Process (GP) to address the open problem of finding appropriate inverse kinematic solutions for upper-limb rehabilitation using redundant robotic exoskeletons. The proposed approach is validated using recorded data from a Motion Capture (MoCap) system and is shown to achieve accurate and unique upper-limb natural solutions for different rehabilitation exercises. This approach is a computationally tractable alternative to existing solutions, which are not always feasible and unsuitable for real-time rehabilitation tests, as they are based on optimizing discomfort indices online or computationally demanding learning methods. A comparison with existing inverse kinematics solutions confirms the superior performance of the proposed approach. Additionally, a robust Model Predictive Control (MPC) with an Integral Sliding Mode (ISM) combination is proposed for safe trajectory tracking during rehabilitation exercises. Several real-time experiments were conducted on a 7-DoF exoskeleton robot with and without the presence of a healthy subject to verify the effectiveness of the proposed methods.

An Intelligent Deep Learning Technique for Predicting Hobbing Tool Wear Based on Gear Hobbing Using Real-Time Monitoring Data

Industry 4.0 has been an impactful and much-needed revolution that has not only influenced different aspects of life but has also changed the course of manufacturing processes. The main purpose of the manufacturing industry is to increase productivity, reduce manufacturing costs, and improve the quality of the product. This has helped to drive economic growth and improve people’s standards. The gear-hobbing industry, being the most efficient one, has not received much attention in terms of Industry 4.0. In prior works, simulation-based approaches with individual parameters, e.g., temperature, current, and vibration, or a few of these parameters, were considered with different approaches, This work presents a real-time experimental approach that involves raw data collection on three different parameters together, i.e., temperature, current, and vibration, using sensors placed on an industrial machine during gear hobbing process manufacturing. The data are preprocessed and then utilised for training an artificial neural network (ANN) to predict the remaininguseful life (RUL) of a tool. It is demonstrated that an ANN with multiple hidden layers can predict the RUL of the tool with high accuracy. The compared results show that tool wear prediction using an ANN with multiple layers has better prediction accuracy during worm gear hobbing.

Pseudospectral convex optimization for on-ramp merging control of connected vehicles

It can be a daunting task for human drivers to merge into highways because of the intricate vehicle negotiations and potential risk within limited time and space. Connected vehicle (CV) technologies could be a solution to this problem and offer many benefits to the road safety, traffic mobility, and energy efficiency. However, real-time optimal control of CVs is still an open challenge, due to the nonlinear vehicle dynamics, non-convex fuel consumption model, and highly dynamic uncertain inter-vehicle interactions. To tackle these issues, a novel real-time optimal control approach that balances the computational efficiency and solution optimality is proposed for the purpose of onboard application. To this end, the pseudospectral collocation method is integrated with a sequential convex programming approach to develop two new optimization algorithms, which are implemented within a model predictive control (MPC) framework to allow for real-time generation of optimal merging speed profiles. One algorithm leverages the line search technique to improve convergence, and the other benefits from the trust region method for better computational efficiency. The optimality and convergence process of both proposed algorithms are investigated by comparing their solutions with a popular non-linear solver. Simulation results show that the proposed methods outperform the benchmark in terms of computational cost, fuel consumption, and traffic efficiency. In particular, the proposed fuel-economy merging rule can save 57.1% fuel consumption on average on four different traffic volumes. Meanwhile, the proposed optimal control algorithms can reduce 2.2% travel time on average comparing to the “first-in-first-out” merging rule.

Classification of Myopathy and Normal Electromyogram (EMG) Data with a New Deep Learning Architecture

Electromyograms (EMG) are recorded movements of nerves and muscles that help diagnose muscles and nerve-related disorders. It is frequently used in the diagnosis of neuromuscular diseases such as myopathy, which causes many changes in EMG signal properties. The most useful auxiliary test in the diagnosis of myopathy is EMG. Therefore, it has become imperative to identify computer-assisted anomalies with full accuracy and to develop an efficient classifier. In this study, a new machine learning method with a deep learning architecture that can score normal and myopathy EMG from the EMGLAB database is proposed. Using the discrete wavelet transform Coiflets 5 (Coif 5) wavelet, the EMG signals are decomposed into subbands and various statistical features are obtained from the wavelet coefficients. The success rates of the decision tree C4.5 algorithm, which is one of the traditional learning architectures, and the Long Short-term Memory (LSTM) algorithm, which is one of the deep learning architectures, were compared. Unlike the studies in the literature, with the LSTM algorithm, a 100% success rate was achieved with the proposed model. In addition, a real-time approach is presented by analyzing the test data classification time of the model.

An Intelligent Detection Method for Approach Distances of Large Construction Equipment in Substations

The safe approach distance detection of large construction equipment in substations is important to ensure the safety and stability of the power system, as well as to prevent equipment damage, power outages and other accidents. The current method is unable to intelligently distinguish construction equipment from power equipment and realize real-time safety approach distance detection. Therefore, this paper constructs a safety approach distance detection system for large-scale construction equipment in substations based on stereo vision and target detection, and realizes real-time high-precision safe approach distance detection between large-scale construction equipment and electric power equipment. Firstly, the system distinguishes construction equipment from power equipment using a GhostNet-based substation construction target detection model. Secondly, the system obtains spatial information regarding the operation scene using a lightweight stereo matching model based on channel attention, then calculates the spatial surface center of the target based on the spatial information and detection results, and finally calculates the safety approach distance between construction equipment and power equipment. Compared with MobileNetv3-YOLOv4, the map and the recall rate of the proposed method are improved by 13.1% and 29.0%; compared with the AnyNet stereo matching method, the proposed method decreases the end point error and 3 pixels error by 34.2% and 25.8%. The actual data show that the detection speed of the proposed method is 19.35 frames per second, and the mean absolute error is 0.942 m and the mean relative error is 3.802%. This method can accurately measure the safe approach distance in real time in real scenarios to guarantee the safety of personnel and equipment.

Real-time optimization of urban channel gate control based on a segmentation hydraulic model

With the urban water resources becoming increasingly scarce, the optimal control engineering has emerged as a promising approach to improve the efficiency of water use in the environment. A hydraulic model is capable of accurately modeling and predicting the complex hydrodynamic processes occurring within a channel. However, its optimization and simulation time are often prolonged by the complexity of the channel system, resulting in poor real-time performance. This study presents a segmented hydraulic real-time optimization approach that combines rule-based simulation (RS) with real-time optimization (RTO). The aim of the proposed method is to reduce hydraulic model complexity and improve optimization time by dividing the full hydraulic model (FHM) into optimized segmented hydraulic model (SHMO) and non-optimized segmented hydraulic model (SHMN). The approach presents two main improvements: (1) a segmentation point recognition method based on RS is used to obtain SHMO from the FHM; and (2) a segmented optimization framework is employed to enable RTO based on SHMO. We demonstrate the effectiveness of the approach using a case study of China's Qing River. The results indicate that FHM can be successfully divided into SHMO and SHMN with similar simulation effect (R > 0.88 and RMSE < 0.1) by using the segmentation point recognition method, and the segmented hydraulic real-time optimization approach can reduce optimization time (average 68%) of hydraulics model. The case study indicated that the proposed method is a computationally efficient and feasible approach for real-time regulation of urban channel gate control based on hydraulic model.

The application of infrared thermography technology in flap: A perspective from bibliometric and visual analysis.

The application of infrared thermography technology (IRT) in flap has become a major focus of research, as it provides a non-invasive, real-time, and quantitative approach for monitoring flap perfusion. In this regard, we conducted a comprehensive visualization and scientometric analysis to systematically summarize and discuss the current state of research in this field. We systematically reviewed publications on the application of IRT in flap procedures from 1999 to 2022, using the Web of Science Core Collection (WoSCC). Through scientometric analysis, we examined annual trends, affiliations, countries, journals, authors, and their relationships, providing insights into current hotspots and future developments in this area. We analysed 522 English studies and found a steady increase in annual publications. The United States and Germany had the highest publication rates, with Beth Israel Deaconess Medical Center and Shanghai Jiaotong University being leading institutions. Notably, Lee BT and Alex Keller emerged as influential authors in this field. Compared to existing techniques, infrared-based technology offers significant advantages for non-invasive monitoring of flap perfusion, including simplicity of operation and objective results. Future trends should focus on interdisciplinary collaborations to develop new infrared devices and achieve intelligent image processing, enabling broader application in various clinical scenarios. This bibliometric study summarizes the progress and landscape of research on 'the Application of infrared thermography technology in flap' over the past two decades, providing valuable insights and serving as a reliable reference to drive further advancements and spark researchers' interest in this field.

Road Condition Monitoring Using Vehicle Built-in Cameras and GPS Sensors: A Deep Learning Approach

Road authorities worldwide can leverage the advances in vehicle technology by continuously monitoring their roads’ conditions to minimize road maintenance costs. The existing methods for carrying out road condition surveys involve manual observations using standard survey forms, performed by qualified personnel. These methods are expensive, time-consuming, infrequent, and can hardly provide real-time information. Some automated approaches also exist but are very expensive since they require special vehicles equipped with computing devices and sensors for data collection and processing. This research aims to leverage the advances in vehicle technology in providing a cheap and real-time approach to carry out road condition monitoring (RCM). This study developed a deep learning model using the You Only Look Once, Version 5 (YOLOv5) algorithm that was trained to capture and categorize flexible pavement distresses (FPD) and reached 95% precision, 93.4% recall, and 97.2% mean Average Precision. Using vehicle built-in cameras and GPS sensors, these distresses were detected, images were captured, and locations were recorded. This was validated on campus roads and parking lots using a car featured with a built-in camera and GPS. The vehicles’ built-in technologies provided a more cost-effective and efficient road condition monitoring approach that could also provide real-time road conditions.

A novel triplex real-time PCR method for the simultaneous authentication of meats and antlers from sika deer (Cervus nippon) and red deer (Cervus elaphus)

Deer products have been used as a traditional Chinese medical and edible food for over 2000 years. Due to the high price, the meats and antlers are often adulterated with low-value materials. The present study developed a triplex real-time PCR assay to simultaneously identify the DNA from sika deer and red deer in deer products. The triplex real-time PCR approach showed high specificity for 10 animal species, and the method simultaneously amplified an endogenous control to eliminate false-negative results. The absolute limits of detection for raw meat, meat product, and antlers were 0.00005–0.000025 ng, 0.00005–0.000025 ng, and 0.0001–0.000025 ng, respectively, for sika deer and 0.00005–0.000025 ng, 0.0005–0.00025 ng, and 0.0025–0.00025 ng, respectively, for red deer. This approach could determine 0.1% (w/w) adulteration. In conclusion, the established triplex real-time PCR assay displayed high specificity and sensitivity, and applicability and can be used for authentic animal products from sika deer and red deer. Data will be made available on request.

Convexification approaches for regional route guidance and demand management with generalized MFDs

Traffic congestion is one of the main concerns in big cities with many adverse socioeconomic effects. A promising solution is to simultaneously regulate the admission of vehicles (i.e., demand management) and redistribute traffic flows within the network (i.e., route guidance). In this work, we integrate demand management with route guidance within a Model Predictive Control framework using regional traffic dynamics with generalized Macroscopic Fundamental Diagram (MFD) shapes. Dealing with generalized MFD shapes is challenging due to the resulting nonlinear and non-convex optimization formulation. To tackle this challenge, we develop two real-time solution approaches: (i) a successive convexification approach that constructs convex bounding sets for all nonlinear terms, and (ii) a linear approximation approach that solves the problem using triangular macroscopic fundamental diagram approximations. The proposed approaches offer a trade-off between execution speed and solution quality, as the linear approximation approach runs faster while the successive convexification approach yields better quality and accuracy solutions. Macroscopic simulation results illustrate the efficiency of the successive convexification and linear approximation approaches yielding an optimality gap of less than 3.5% and 10% in all considered cases, respectively. Furthermore, both approaches outperform a state-of-practice nonlinear solver in terms of solution quality and execution time. Finally, substantial gains are also obtained regarding travel time and traffic flow efficiency in a realistic microsimulation environment.

A digital twin architecture for real-time and offline high granularity analysis in smart buildings

Smart buildings aim to create a safe, comfortable and sustainable environment for the occupants while increasing the energy performance of the building to reduce the environmental footprint and operational cost. Towards this direction, the use of model-based digital twins able to simulate key aspects of the building is important. Despite the utilization of various digital twin approaches for smart buildings in recent years, most of these approaches mainly focus on one aspect of the building, such as indoor environmental quality or energy consumption. This observation necessitates the development of a virtual but realistic and non-invasive environment able to simulate all the key domains of the building, including indoor air quality, thermal comfort, and energy consumption. Towards this direction, this work proposes an architecture for smart buildings investigations, that emulates the thermal comfort, air-quality, and electricity consumption of the building, while considering the activities and behavior of occupants. For simulating the building operations, ordinary differential equations and power balance modeling are incorporated to facilitate real-time or faster than real-time digital twin approaches, while computational fluid dynamics modeling is used complementarily, to enable higher granularity simulations for offline analysis. Both models have been validated in a real building achieving less than 1% average zone temperature error between the simulation models and the actual building sensor measurements. Furthermore, a user-friendly software platform has been developed and integrated with the proposed architecture, where multiple important features and characteristics have been included to enable various investigations. In particular, the developed platform enables physical and simulated building integration, model calibration using field measurements, experimental configuration and management, execution of novel what-if investigations, and evaluation of novel algorithms using hardware-in-the-loop configurations. The proposed architecture is applied in a physical building, allowing the demonstration of several investigations and the execution of interesting use-cases.

Skeletal muscle-specific DJ-1 ablation-induced atrogenes expression and mitochondrial dysfunction contributing to muscular atrophy.

DJ-1 is a causative gene for Parkinson's disease. DJ-1-deficient mice develop gait-associated progressive behavioural abnormalities and hypoactive forearm grip strength. However, underlying activity mechanisms are not fully explored. Western blotting and quantitative real-time polymerase chain reaction approaches were adopted to analyse DJ-1 expression in skeletal muscle from aged humans or mice and compared with young subjects. Skeletal muscle-specific-DJ-1 knockout (MDKO) mice were generated, followed by an assessment of the physical activity phenotypes (grip strength, maximal load capacity, and hanging, rotarod, and exercise capacity tests) of the MDKO and control mice on the chow diet. Muscular atrophy phenotypes (cross-sectional area and fibre types) were determined by imaging and quantitative real-time polymerase chain reaction. Mitochondrial function and skeletal muscle morphology were evaluated by oxygen consumption rate and electron microscopy, respectively. Tail suspension was applied to address disuse atrophy. RNA-seq analysis was performed to indicate molecular changes in muscles with DJ-1 ablation. Dual-luciferase reporter assays were employed to identify the promoter region of Trim63 and Fbxo32 genes, which were indirectly regulated by DJ-1 via the FoxO1 pathway. Cytoplasmic and nuclear fractions of DJ-1-deleted muscle cells were analysed by western blotting. Compound 23 was administered into the gastrocnemius muscle to mimic the of DJ-1 deletion effects. DJ-1 expression decreased in atrophied muscles of aged human (young men, n=2; old with aged men, n=2; young women, n=2; old with aged women, n=2) and immobilization mice (n=6, P<0.01). MDKO mice exhibited no body weight difference compared with control mice on the chow diet (Flox, n=8; MDKO, n=9). DJ-1-deficient muscles were slightly dystrophic (Flox, n=7; MDKO, n=8; P<0.05), with impaired physical activities and oxidative capacity (n=8, P<0.01). In disuse-atrophic conditions, MDKO mice showed smaller cross-sectional area (n=5, P<0.01) and more central nuclei than control mice (Flox, n=7; MDKO, n=6; P<0.05), without alteration in muscle fibre types (Flox, n=6; MDKO, n=7). Biochemical analysis indicated that reduced mitochondrial function and upregulated of atrogenes induced these changes. Furthermore, RNA-seq analysis revealed enhanced activity of the FoxO1 signalling pathway in DJ-1-ablated muscles, which was responsible for the induction of atrogenes. Finally, compound 23 (an inhibitor of DJ-1) could mimic the effects of DJ-1 ablation in vivo. Our results illuminate the crucial of skeletal muscle DJ-1 in the regulation of catabolic signals from mechanical stimulation, providing a therapeutic target for muscle wasting diseases.

Multi-Round versus Real-Time Delphi survey approach for achieving consensus in the COHESION core outcome set: a randomised trial

BackgroundDelphi surveys are commonly used to prioritise critical outcomes in core outcome set (COS) development. This trial aims to compare a three-round (Multi-Round) Delphi (MRD) with a Real-Time Delphi (RTD) in the prioritisation of outcomes for inclusion in a COS for neonatal encephalopathy treatments and explore whether ‘feedback’, ‘iteration’, and ‘initial condition’ effects may occur in the two survey methods.MethodsWe recruited 269 participants (parents/caregivers, healthcare providers and researchers/academics) of which 222 were randomised to either the MRD or the RTD. We investigated the outcomes prioritised in each survey and the ‘feedback’, ‘iteration’, and ‘initial condition’ effects to identify differences between the two survey methods.ResultsIn the RTD, n = 92 participants (83%) fully completed the survey. In the MRD, n = 60 participants (54%) completed all three rounds. Of the 92 outcomes presented, 26 (28%) were prioritised differently between the RTD and MRD. Significantly fewer participants amended their scores when shown stakeholder responses in the RTD compared to the MRD (‘feedback effect’). The ‘iteration effect’ analysis found most experts appeared satisfied with their initial ratings in the RTD and did not amend their scores following stakeholder response feedback. Where they did amend their scores, ratings were amended substantially, suggesting greater convergence. Variance in scores reduced with subsequent rounds of the MRD (‘iteration effect’). Whilst most participants did not change their initial scores in the RTD, of those that did, later recruits tended to align their final score more closely to the group mean final score than earlier recruits (an ‘initial condition’ effect).ConclusionThe feedback effect differed between the two Delphi methods but the magnitude of this difference was small and likely due to the large number of observations rather than because of a meaningfully large difference. It did not appear to be advantageous to require participants to engage in three rounds of a survey due to the low change in scores. Larger drop-out through successive rounds in the MRD, together with a lesser convergence of scores and longer time to completion, indicate considerable benefits of the RTD approach.Trial registrationNCT04471103. Registered on 14 July 2020.

Microneedle coupled epidermal sensor for multiplexed electrochemical detection of kidney disease biomarkers

Early diagnosis of chronic kidney disease (CKD) and constant monitoring to guide optimal intervention is critical to prevent renal failure and other critical diseases. However, the conventional blood tests in hospital are time-consuming and have poor patient compliance. Herein, we demonstrate a real-time, minimally invasive, and self-administrable approach to detect kidney biomarkers in the skin interstitial fluid (ISF) using a polymeric microneedle coupled electrochemical sensor array (MNESA). Microneedles can readily penetrate stratum corneum and quickly extract ISF onto the sensors. Four biomarkers are simultaneously detected to avoid false positive and provide an accurate assessment of kidney functions. Using an artificial skin model, it is shown that MNSEA gives specific and sensitive responses to these kidney biomarkers in physiologically relevant ranges (phosphate: 0.3–1.8 mM, 3.62 μA/mM; uric acid: 50–550 μM, 4.19 nA/μM; creatinine: 50–550 μM, 12.58 nA/μM; urea: 1–16 mM, 44.6 mV/decade). Using a mouse model, we demonstrate that this approach is as reliable as the commercial assays and is feasible to readily monitor the progression of CDK.

Theta-phase-specific modulation of dentate gyrus memory neurons.

The theta rhythm, a quasi-periodic 4-10 Hz oscillation, is observed during memory processing in the hippocampus, with different phases of theta hypothesized to separate independent streams of information related to the encoding and recall of memories. At the cellular level, the discovery of hippocampal memory cells (engram neurons), as well as the modulation of memory recall through optogenetic activation of these cells, has provided evidence that certain memories are stored, in part, in a sparse ensemble of neurons in the hippocampus. In previous research, however, engram reactivation has been carried out using open-loop stimulation at fixed frequencies; the relationship between engram neuron reactivation and ongoing network oscillations has not been taken into consideration. To address this concern, we implemented a closed-loop reactivation of engram neurons that enabled phase-specific stimulation relative to theta oscillations in the local field potential in CA1. Using this real-time approach, we tested the impact of activating dentate gyrus engram neurons during the peak (encoding phase) and trough (recall phase) of theta oscillations. Consistent with previously hypothesized functions of theta oscillations in memory function, we show that stimulating dentate gyrus engram neurons at the trough of theta is more effective in eliciting behavioral recall than either fixed-frequency stimulation or stimulation at the peak of theta. Moreover, phase-specific trough stimulation is accompanied by an increase in the coupling between gamma and theta oscillations in CA1 hippocampus. Our results provide a causal link between phase-specific activation of engram cells and the behavioral expression of memory.