Does pre-speech auditory modulation reflect processes related to feedback monitoring or speech movement planning?

Previous studies have revealed that auditory processing is modulated during the planning phase immediately prior to speech onset. To date, the functional relevance of this pre-speech auditory modulation (PSAM) remains unknown. Here, we investigated whether PSAM reflects neuronal processes that are associated with preparing auditory cortex for optimized feedback monitoring as reflected in online speech corrections. Combining electroencephalographic PSAM data from a previous data set with new acoustic measures of the same participants' speech, we asked whether individual speakers' extent of PSAM is correlated with the implementation of within-vowel articulatory adjustments during /b/-vowel-/d/ word productions. Online articulatory adjustments were quantified as the extent of change in inter-trial formant variability from vowel onset to vowel midpoint (a phenomenon known as centering). This approach allowed us to also consider inter-trial variability in formant production and its possible relation to PSAM at vowel onset and midpoint separately. Results showed that inter-trial formant variability was significantly smaller at vowel midpoint than at vowel onset. PSAM was not significantly correlated with this amount of change in variability as an index of within-vowel adjustments. Surprisingly, PSAM was negatively correlated with inter-trial formant variability not only in the middle but also at the very onset of the vowels. Thus, speakers with more PSAM produced formants that were already less variable at vowel onset. Findings suggest that PSAM may reflect processes that influence speech acoustics as early as vowel onset and, thus, that are directly involved in motor command preparation (feedforward control) rather than output monitoring (feedback control).

Application of a feedforward control-based intervention for preventing hypothermia in trauma patients in a pre-hospital emergency setting.

OBJECTIVE: To investigate the efficacy of a feedforward control-based intervention strategy for preventing hypothermia among trauma patients during pre-hospital emergency care. METHODS: We conducted a retrospective analysis comparing trauma patients treated before and after implementing the intervention, with 40 cases in each group. All patients received emergency care from the Fuzhou Emergency Center on the scene. Multivariate analysis was used to explore the risk factors for hypothermia. The effective rate, incidence of adverse reactions, quality of body temperature management, medical staff's knowledge, attitudes, and behaviors regarding mild hypothermia prevention, coagulation function, treatment time at various stages, prognosis score, and treatment situation were compared between the two groups. RESULTS: The adverse reactions, intervention methods, and degree of cognitive improvement were influencing factors for hypothermia. The effective rate (92.50%) in the feedforward control group was higher than that in the non-feedforward control group (65.00%), with a lower incidence of adverse reactions (2.50%). The temperature management quality score of the feedforward control group (6.23±0.62) was higher. The feedforward control group achieved a higher quality score for temperature management (6.23±0.62) and exhibited a greater understanding of hypothermia prevention among trauma patients (P<0.05). Compared to the non-feedforward control group, the feedforward control group showed improved coagulation function, better performance in treatment time at each node, and higher prognosis scores. CONCLUSION: The intervention model based on feedforward control can effectively improve the standard of pre-hospital emergency care and prevent the incidence of hypothermia in trauma patients.

A Cost-Effective Integrated Methodology for Electromagnetic Actuation via Visual Feedback.

Electromagnetic actuation can support many fields of technology, such as robotics or biomedical applications. In this context, fully understanding the system behavior and proposing a low-cost package for feedback control is challenging. Modeling the electromagnetic force is particularly tricky because it is a nonlinear function of the actuated object's position and coil's current. Measuring in real time the position of the actuated object with the precision required for accurate motion control is also nontrivial. In this study, we propose a novel, cost-effective electromagnetic set-up to achieve position control via visual feedback. We actuated vertically and under different experimental conditions a 10 mm diameter steel ball hanging on a low-stiffness spring, demonstrating good tracking performance (the position error remained within ±0.5 mm, with a negligible phase delay in the best scenarios). The experimental results confirm the feasibility of the proposed set-up, which is characterized by minimum complexity and realized with off-the-shelf and cost-effective components. For these reasons, such a contribution helps to understand and apply electromagnetic actuation even further.

Composite ADRC Speed Control Method Based on LTDRO Feedforward Compensation.

The performance of the extended state observer (ESO) in an Active Disturbance Rejection Control (ADRC) is limited by the operational load in stepper motor control, which has high real-time requirements and may cause delays. Additionally, the complexity of parameter tuning, especially in high-order systems, further limits the ESO's performance. This paper proposes a composite ADRC (LTDRO-ADRC) based on a load torque dimensionality reduction observer (LTDRO). Firstly, the LTDRO is designed to estimate abrupt load disturbances that are difficult to compensate for using the ESO. Secondly, the transfer function under the double-closed loop is deduced. Additionally, the LTDRO uses a magnetic encoder to gather the system state and calculate the load torque. It then outputs a compensating current feedforward to the current loop input. This method reduces the delay and complexity of the ESO, improving the response speed of the ADRC speed ring and the overall response of the system to load changes. Simulation and experimental results demonstrate that it significantly enhances dynamic control performance and steady-state errors. LTDRO-ADRC can stabilize the speed again within 49 ms and 17 ms, respectively, in the face of sudden load increase and sudden load removal. At the same time, in terms of steady-state error, compared with ADRC and CADRC, they have increased by 94% and 88%, respectively. In terms of zero-speed starting motors, the response speed is increased by 58% compared to a traditional ADRC.

The influence of postural threat-induced anxiety on locomotor learning and updating.

The ability to adapt our locomotion in a feedforward (i.e., "predictive") manner is crucial for safe and efficient walking behavior. Equally important is the ability to quickly deadapt and update behavior that is no longer appropriate for the given context. It has been suggested that anxiety induced via postural threat may play a fundamental role in disrupting such deadaptation. We tested this hypothesis, using the "broken escalator" phenomenon: Fifty-six healthy young adults walked onto a stationary walkway ("BEFORE" condition, 5 trials), then onto a moving walkway akin to an airport travelator ("MOVING" condition, 10 trials), and then again onto the stationary walkway ("AFTER" condition, 5 trials). Participants completed all trials while wearing a virtual reality headset, which was used to induce postural threat-related anxiety (raised clifflike drop at the end of the walkway) during different phases of the paradigm. We found that performing the locomotor adaptation phase in a state of increased threat disrupted subsequent deadaptation during AFTER trials: These participants displayed anticipatory muscular activity as if expecting the platform to move and exhibited inappropriate anticipatory forward trunk movement that persisted during multiple AFTER trials. In contrast, postural threat induced during AFTER trials did not affect behavioral or neurophysiological outcomes. These findings highlight that actions learned in the presence of postural threat-induced anxiety are strengthened, leading to difficulties in deadapting these behaviors when no longer appropriate. Given the associations between anxiety and persistent maladaptive gait behaviors (e.g., "overly cautious" gait, functional gait disorders), the findings have implications for the understanding of such conditions.NEW & NOTEWORTHY Safe and efficient locomotion frequently requires movements to be adapted in a feedforward (i.e., "predictive") manner. These adaptations are not always correct, and thus inappropriate behavior must be quickly updated. Here we showed that increased threat disrupts this process. We found that locomotor actions learned in the presence of postural threat-induced anxiety are strengthened, subsequently impairing one's ability to update (or "deadapt") these actions when they are no longer appropriate for the current context.

Motor deficits in autism differ from that of developmental coordination disorder.

LAY ABSTRACT: A vast majority of individuals with autism spectrum disorder experience impairments in motor skills. Those are often labelled as additional developmental coordination disorder despite the lack of studies comparing both disorders. Consequently, motor skills rehabilitation programmes in autism are often not specific but rather consist in standard programmes for developmental coordination disorder. Here, we compared motor performance in three groups of children: a control group, an autism spectrum disorder group and a developmental coordination disorder group. Despite similar level of motor skills evaluated by the standard movement assessment battery for children, in a Reach-to-Displace Task, children with autism spectrum disorder and developmental coordination disorder showed specific motor control deficits. Children with autism spectrum disorder failed to anticipate the object properties, but could correct their movement as well as typically developing children. In contrast, children with developmental coordination disorder were atypically slow, but showed a spared anticipation. Our study has important clinical implications as motor skills rehabilitations are crucial to both populations. Specifically, our findings suggest that individuals with autism spectrum disorder would benefit from therapies aiming at improving their anticipation, maybe through the support of their preserved representations and use of sensory information. Conversely, individuals with developmental coordination disorder would benefit from a focus on the use of sensory information in a timely fashion.

Effect of respiratory training based on core stability training on feedforward control in patients with chronic nonspecific low back pain / 中国康复理论与实践

ObjectiveTo investigate the effect of respiratory training based on core stability training on feedforward control in patients with chronic nonspecific low back pain (CNLBP). MethodsA total of 60 patients with CNLBP in Jiaxing Second Hospital from January, 2022 to March, 2023 were randomly divided into control group (n = 30) and experimental group (n = 30). Both groups received health education, physical factor therapy and core stability training, while the experimental group received respiratory training in addition, for four weeks. Visual Analogue Scale (VAS) score, Japanese Orthopaedic Association low back pain (JOA) score and Oswestry Dysfunction Index (ODI) were compared between two groups before and after treatment, while surface electromyography was used to detect the root mean square (RMS) and integrated electromyography (iEMG) of transversus abdominis, multifidus and triceps (movement muscles), and the activation sequence and relative activation time of transversus abdominis, multifidus and triceps were calculated. ResultsAfter treatment, the scores of VAS, JOA and ODI improved significantly in both groups (|t| > 8.515, P < 0.001), and the scores were better in the experimental group than in the control group (|t| > 2.089, P < 0.05). RMS and iEMG of transversus abdominis and multifidus improved significantly after treatment in both groups (|t| > 18.831, P < 0.001), and were significantly better in the experimental group (|t| > 3.481, P < 0.05). The transversus abdominis and multifidus in both groups were activated before the movement muscles, and the relative activation time of transversus abdominis and multifidus increased in negative (|t| > 48.115, P < 0.001), the experimental group being better (|t| > 3.229, P < 0.05). ConclusionCombination of core stability training and respiratory training is beneficial in reducing the pain of patients with CNLBP, reducing the lumbar dysfunction, improving the order of muscle activation, and strengthening feedforward control.

Enhancing projection based iterative learning control: A set-membership approach.

Lithographic machine tools require both high motion accuracy and high motion flexibility. Projection based iterative learning control (P-ILC) is appealing for wafer stages to achieve two goals, simultaneously. P-ILC contains a nonparametric feedforward controller based on ILC, and a parametric feedforward controller with a projection step for feedforward tuning. In this paper, a set-membership based frequency-domain ILC algorithm (SM-F-ILC) is employed in the enhancing P-ILC scheme to improve the performance in the nonparametric feedforward control mode. SM-F-ILC can effectively compensate for repetitive errors, attenuate the nonrepetitive error accumulation and achieve fast convergence speed with model uncertainties. These superiorities also facilitate to improve the performance of P-ILC in the parametric feedforward control mode. The validity of the enhancing P-ILC scheme is demonstrated by experimental results.

An Empirical Study of Feedforward Control in Unplanned Extubation of Nasogastric Tube.

Objective: To explore the effect of feedforward control on reducing the incidence of unplanned extubation and improving the quality of catheter nursing. Methods: A total of 186 patients with nasogastric tube after gastrointestinal surgery in the eastern region of our hospital from September 2020 to September 2021 were selected as the control group; 186 patients with nasogastric tube after gastrointestinal surgery in the western region of our hospital at the same period were selected as the experimental group. The influencing factors of unplanned extubation in patients with long-term postoperative nasogastric tube were analyzed, and effective preoperative and postoperative health education was conducted. The ratio of unplanned extubation of nasogastric tube and nursing satisfaction of patients in the two groups were compared. Results: Patient constraint, perceived pressure score, anxiety score, nasal gastrointestinal canal health education feedback score and indwell tube comfort score were independent risk factors for unplanned extubation. The restraint rate and the incidence of unplanned extubation in the experimental group were lower than those in the control group after intervention, with statistical significance (P < 0.05). The nursing satisfaction of the experimental group was significantly higher than that of the control group after feedforward cognitive intervention. After intervention, serum albumin and gastric PH in the experimental group were significantly higher than those in the control group (P < 0.05). Conclusion: The safe nursing management method of feed forward control can effectively reduce the incidence of unplanned extubation in inpatients, which is worth further promoting in nursing work.

Grasping behavior does not recover after sight restoration from congenital blindness.

We investigated whether early visual input is essential for establishing the ability to use predictions in the control of actions and for perception. To successfully interact with objects, it is necessary to pre-program bodily actions such as grasping movements (feedforward control). Feedforward control requires a model for making predictions, which is typically shaped by previous sensory experience and interaction with the environment.1 Vision is the most crucial sense for establishing such predictions.2,3 We typically rely on visual estimations of the to-be-grasped object's size and weight in order to scale grip force and hand aperture accordingly.4,5,6 Size-weight expectations play a role also for perception, as evident in the size-weight illusion (SWI), in which the smaller of two equal-weight objects is misjudged to be heavier.7,8 Here, we investigated predictions for action and perception by testing the development of feedforward controlled grasping and of the SWI in young individuals surgically treated for congenital cataracts several years after birth. Surprisingly, what typically developing individuals do easily within the first years of life, namely to adeptly grasp new objects based on visually predicted properties, cataract-treated individuals did not learn after years of visual experience. Contrary, the SWI exhibited significant development. Even though the two tasks differ in substantial ways, these results may suggest a potential dissociation in using visual experience to make predictions about an object's features for perception or action. What seems a very simple task-picking up small objects-is in truth a highly complex computation that necessitates early structured visual input to develop.

A methodology to implement a closed-loop feedback-feedforward level control in a laboratory-scale flotation bank using peristaltic pumps.

This paper describes the implementation of a level control strategy in a laboratory-scale flotation system. The laboratory-scale system consists of a bank of three flotation tanks connected in series, which mimics a flotation system found in mineral processing plants. Besides the classical feedback control strategy, we have also included a feedforward strategy to better account for process disturbances. Results revealed that the level control performance significantly improves when a feedforward strategy is considered. This methodology uses peristaltic pumps for level control, which has not been extensively documented even though: (1) peristaltic pumps are commonly used in laboratory-scale systems, and (2) the control implementation is not as straightforward as those control strategies that use valves. Therefore, we believe that this paper, which describes a proven methodology that has been validated in an experimental system, can be a useful reference for many researchers in the field.•Preparation of reagents to ensure that the froth stability of the froth layer is representative of an industrial flotation froth.•Calibration of instruments - convert the electrical signal from PLCs to engineering units.•Tuning PI parameters using SIMC rules by performing step-changes in each flotation cell.

Advanced Soft Robotic System for In Situ 3D Bioprinting and Endoscopic Surgery.

Three-dimensional (3D) bioprinting technology offers great potential in the treatment of tissue and organ damage. Conventional approaches generally rely on a large form factor desktop bioprinter to create in vitro 3D living constructs before introducing them into the patient's body, which poses several drawbacks such as surface mismatches, structure damage, and high contamination along with tissue injury due to transport and large open-field surgery. In situ bioprinting inside a living body is a potentially transformational solution as the body serves as an excellent bioreactor. This work introduces a multifunctional and flexible in situ 3D bioprinter (F3DB), which features a high degree of freedom soft printing head integrated into a flexible robotic arm to deliver multilayered biomaterials to internal organs/tissues. The device has a master-slave architecture and is operated by a kinematic inversion model and learning-based controllers. The 3D printing capabilities with different patterns, surfaces, and on a colon phantom are also tested with different composite hydrogels and biomaterials. The F3DB capability to perform endoscopic surgery is further demonstrated with fresh porcine tissue. The new system is expected to bridge a gap in the field of in situ bioprinting and support the future development of advanced endoscopic surgical robots.

Finite time stable inversion in discrete frequency domain: Accuracy analysis, improvement and application to wafer stage.

Stable inversion represents a classic approach to achieve the exact inverse input for non-minimum-phase (NMP) systems. Solutions deduced based on state-space equations and transfer functions have been frequently proposed, however they are basically developed in the infinite-time horizon and are inherently time-domain computation methods. Considering the practical finite-time tracking tasks, this paper investigates the finite-time stable inversion problem. In particular, the discrete-frequency-domain solution which enables frequency-domain computation is studied. As the main contribution of the paper, the accuracy issues of the discrete-frequency-domain solution are revealed and an easy-to-use procedure is provided to improve the inversion accuracy by utilizing pre-actuation and post-actuation methods. Simulation and experiment both verify the effectiveness of the developed discrete-frequency-domain stable inversion technique.

Combined feedforward and error-based active disturbance rejection control for diesel particulate filter thermal regeneration.

In this paper, a novel control solution, which combines a feedforward control law and an error-based version of the active disturbance rejection control (ADRC) scheme, is proposed for the exhaust gas temperature control during the thermal regeneration process of a diesel particulate filter in the exhaust line of a diesel engine. Attributed to the complexity of the controlled upstream diesel oxidation catalyst (DOC), its thermodynamics is firstly captured and characterized by a set of linear models through identification modeling. Then, a novel error-based ADRC controller, in which the separated components in conventional ADRC such as the extended state observer (ESO) and the feedback compensator are restructured into a single, modularized control function block, is designed by applying the identified nominal DOC model. In order to further unburden the error-based ESO for better achievements, a combined feedforward compensator is well designed on the basis of the principle of energy balance. Thus a hybrid, 2-degree-of-freedom (2-DOF) controller is developed for better dynamic performance of the controlled DOC system. Its stability performance is also analyzed in the work. The robustness and advantages of the presented hybrid control scheme are finally validated and compared with a well-tuned regular PID-based controller by means of extensive simulation and experimental tests. The results show that the proposed hybrid controller is capable of providing more accurate and faster temperature response and is less sensitive to the variation of system parameters and external disturbances. Moreover, as the error-based ADRC in the hybrid scheme takes the reference tracking error as its direct input and is compatible with the regular PID controller in terms of input and output interfaces, it herein provides an appealing control scheme for existing applications as a substitute for the conventional PID-based controllers to achieve improved performance.

Influence of feedforward control-based health education intervention on compliance, visual function and self-perceived burden among patients with diabetic retinopathy.

Objective: To assess the influence of feedforward control-based health education intervention on the compliance, visual function and self-perceived burden (SPB) among patients with diabetic retinopathy (DR). Methods: Eighty-six DR patients were divided into feedforward control and control groups (n=43). The control group was given routine nursing intervention, based on which the feedforward control group received feedforward control-based health education intervention. The health behavior indices were compared after intervention. The correlations of QOL score with SPB score and health behavior indices were analysed using Pearson's and Spearman's coefficients. Results: After intervention, the total QOL score and scores of symptoms and visual function, physical function, social activity, and mentality and psychology were significantly improved compared with those before intervention, which were significantly higher in the feedforward control group (P<0.05). SPB score was significantly lower in the two groups after intervention than that before intervention, particularly in the feedforward control group (P<0.05). The QOL score of DR patients was significantly negatively correlated with SPB score but positively correlated with health behavior indices (P<0.05). Conclusion: The feedforward control-based health education intervention mode is beneficial for guiding DR patients to promote visual function recovery and to reduce SPB.

Synthetic Stochastic Motion Platform for Testing Single Particle Tracking Microscopes.

We describe the design and implementation of a control system for testing the performance of single particle tracking microscopes with the method of synthetic motion. Single particle tracking (SPT) has become a common and powerful tool in the study of biomolecular transport in cellular biology, providing the ability to track individual biological macromolecules in their native environment. Existing methods for testing SPT techniques rely on physical simulations and there is a clear need for experimental-based schemes for both comparing different approaches and for characterizing the accuracy and precision of techniques on particular experimental setups. Synthetic motion, that is, using an actuator such as a nanopositioning stage to drive a particle along a known ground-truth trajectory, is a means for achieving these ends. However, the resolution, accuracy, and flexibility of this method is limited by the actuator static and dynamic characteristics. In this work we apply system identification and model inverse feedforward control to increase actuator bandwidth and address some common actuator nonlinearities, develop a set of dimensionless numbers that describe system limitations, and provide a set of guidelines for the practical use of synthetic motion in SPT.

Effect of a Feedforward Control Model on Nursing Attitudes in the Operating Room.

Objective: To evaluate the effect of a feedforward control model on nursing error behaviour and nursing attitudes in the operating room. Methods: A total of 216 patients admitted to our hospital from January 2018 to December 2020 were randomly selected as the research subjects and divided into a control group and a research group, with 108 cases in each. The control group received routine surgical nursing management, while the research group received surgical nursing management with a feedforward control model. Nursing quality, error behaviour, and work attitude score, as well as the occurrence of safety hazards and patient satisfaction, were compared between the two groups of patients. Results: Compared with the control group, the nursing quality score, including nursing document quality (72.0 ± 1.8 vs 97.2 ± 2.0), quality control inspection (75.4 ± 1.9 vs 95.5 ± 1.2), disinfection and isolation (73.4 ± 2.4 vs 96.6 ± 2.0) and nursing safety management (71.1 ± 2.2 vs 98.0 ± 1.8), were higher in the research group (P < 0.05); the scores for nursing error behaviour (13.2 ± 1.0 vs 19.4 ± 1.8) and work attitudes of the research group (14.0 ± 1.2 vs 19.0 ± 2.0) were higher (P < 0.05); and the research group had a lower incidence of safety hazards (11.1% vs 0.9%, P < 0.05); finally, the research group had higher patient satisfaction (71.3% vs 93.5%, P < 0.05). Conclusion: The application of a feedforward control model for nursing management in the operating room can significantly reduce nurses' errors and improve their attitudes, continuously reduce the incidence of safety incidents, improve the quality of operating room nursing and increase the satisfaction of patients and their families with the nursing care they received.

Impaired feedforward control of movements in pianists with focal dystonia.

Learning accurate and fast movements typically accompanies the modulation of feedforward control. Nevertheless, it remains unclear how motor skill learning modulates feedforward control, such as through maladaptation of the sensorimotor system by extensive training (e.g., task-specific dystonia). Here, we examined the modulation of feedforward control through motor skill learning while focusing on the motion of piano playing at either a natural tempo or the fastest tempo. The current study compared the kinematics and keypress data among individuals in three groups: healthy and well-trained pianists (i.e., subjects with skill learning), non-musicians (i.e., subjects without skill learning), and patients with focal-hand dystonia (FHD) (i.e., subjects with maladaptation by skill learning). Compared to healthy pianists, patients with FHD showed impairment in some feedforward motion components that are relevant to classifying the two playing tempi. However, while focusing on motion components that are irrelevant to the tempo classification, patients with FHD showed movements comparable to those of healthy pianists. Furthermore, patients with FHD demonstrated significantly slower movement times than healthy pianists. Our results suggest that maladaptation by skill learning affects parts of feedforward control rather than its entirety. Nevertheless, the affected feedforward components are relevant to performing movements as fast as possible, which may underlie the speed dependence of dystonic symptoms.

Lower limb kinematics during single leg landing in three directions in individuals with chronic ankle instability.

OBJECTIVES: To compare the lower limb kinematics of participants with chronic ankle instability (CAI) and healthy participants during forward, lateral, and medial landings. DESIGN: Cross-sectional study. SETTING: Laboratory. PARTICIPANTS: Eighteen athletes with CAI and 18 control athletes. MAIN OUTCOME MEASURES: Hip, knee, and ankle joint kinematics during forward, lateral, and medial single-leg landings were compared between the groups using two-way ANOVA for discrete values and statistical parametric mapping two-sample t-tests for time-series data. RESULTS: The CAI group had significantly greater ankle dorsiflexion than the control group (P ≤ 0.013), which was observed from the pre-initial contact (IC) for lateral and medial landings and post-IC for forward landing. The CAI group showed greater knee flexion than the control group from the IC for lateral landing and post-IC for forward landing (P ≤ 0.014). No significant differences in ankle inversion kinematics were found between the CAI and control groups. Lateral landing had a greater peak inversion angle and velocity than forward and medial landings (P < 0.001). Medial landing had a greater inversion velocity than forward landing (P < 0.001). CONCLUSIONS: This study suggests that individuals with CAI show feedforward protective adaptations in the pre-landing phase for lateral and medial landings.

Feedforward Control of Piezoelectric Ceramic Actuators Based on PEA-RNN.

Multilayer perceptron (MLP) has been demonstrated to implement feedforward control of the piezoelectric actuator (PEA). To further improve the control accuracy of the neural network, reduce the training time, and explore the possibility of online model updating, a novel recurrent neural network named PEA-RNN is established in this paper. PEA-RNN is a three-input, one-output neural network, including one gated recurrent unit (GRU) layer, seven linear layers, and one residual connection in the linear layers. The experimental results show that the displacement linearity error of piezoelectric ceramics reaches 8.96 µm in the open-loop condition. After using PEA-RNN compensation, the maximum displacement error of piezoelectric ceramics is reduced to 0.465 µm at the operating frequency of 10 Hz, which proves that PEA-RNN can accurately compensate piezoelectric ceramics' dynamic hysteresis nonlinearity. At the same time, the training epochs of PEA-RNN are only 5% of the MLP, and fewer training epochs provide the possibility to realize online updates of the model in the future.