Neural Conversion Research Articles

Predicting long-term deposit customers using convolutional neural network and data conversion technique

Predicting long-term deposit customers using convolutional neural network and data conversion technique

3D model for human glia conversion into subtype-specific neurons, including dopamine neurons

Two-dimensional neuronal cultures have a limited ability to recapitulate the invivo environment of the brain. Here, we introduce a three-dimensional invitro model for human glia-to-neuron conversion, surpassing the spatial and temporal constrains of two-dimensional cultures. Focused on direct conversion to induced dopamine neurons (iDANs) relevant to Parkinson disease, the model generates functionally mature iDANs in 2weeks and allows long-term survival. As proof of concept, we use single-nucleus RNA sequencing and molecular lineage tracing during iDAN generation and find that all glial subtypes generate neurons and that conversion relies on the coordinated expression of three neural conversion factors. We also show the formation of mature and functional iDANs over time. The model facilitates molecular investigations of the conversion process to enhance understanding of conversion outcomes and offers a system for invitro reprogramming studies aimed at advancing alternative therapeutic strategies in the diseased brain.

More Efficient Energy Management for Hybrid AC/DC Microgrids With Hard Constrained Neural Network-Based Conversion Loss Surrogate Models

More Efficient Energy Management for Hybrid AC/DC Microgrids With Hard Constrained Neural Network-Based Conversion Loss Surrogate Models

Development of mobile heart rate monitoring tools using artificial neural networks

Currently, there is a great need for mobile monitoring systems used to detect heart rate abnormalities in citizens working or living in hard-to-reach areas. Due to problems with transmitting data from such areas, it is desirable to carry out the entire information processing process at the study site. But, despite the limited hardware costs for the implementation of such systems, it is necessary to ensure high reliability of monitoring results. Purpose is to develop a structural diagram of a mobile heart rate analysis device, select the architecture of the neural network of the conversion and segmentation subsystem of the mobile device, determine the structure of the neural network of the conversion and segmentation subsystem, train the neural network of the conversion and segmentation subsystem, conduct software modeling of the neural network of the conversion and segmentation subsystem, implementation of a neural network conversion and segmentation subsystem based on field programmable gate arrays (FPGAs). A structural diagram of a mobile heart rate analysis device has been developed. The choice of the neural network architecture of the mobile device conversion and segmentation subsystem is justified. The structure of the neural network of the transformation and segmentation subsystem is determined. The neural network of the transformation and segmentation subsystem was trained. A software simulation of the neural network of the transformation and segmentation subsystem was carried out, confirming the high reliability of the proposed neural network structure. A neural network conversion and segmentation subsystem has been implemented in the hardware description language VHDL based on an FPGA. The developed structural diagram of the mobile device and the neural network structure of the conversion and segmentation subsystem can be used when implementing mobile heart rate monitoring means.

A Compact Neural Network-Based Conversion Loss Model With Hard Constraints for Energy Management

A Compact Neural Network-Based Conversion Loss Model With Hard Constraints for Energy Management

Identification and validation of novel engineered AAV capsid variants targeting human glia.

Direct neural conversion of endogenous non-neuronal cells, such as resident glia, into therapeutic neurons has emerged as a promising strategy for brain repair, aiming to restore lost or damaged neurons. Proof-of-concept has been obtained from animal studies, yet these models do not efficiently recapitulate the complexity of the human brain, and further refinement is necessary before clinical translation becomes viable. One important aspect is the need to achieve efficient and precise targeting of human glial cells using non-integrating viral vectors that exhibit a high degree of cell type specificity. While various naturally occurring or engineered adeno-associated virus (AAV) serotypes have been utilized to transduce glia, efficient targeting of human glial cell types remains an unsolved challenge. In this study, we employ AAV capsid library engineering to find AAV capsids that selectively target human glia in vitro and in vivo. We have identified two families of AAV capsids that induce efficient targeting of human glia both in glial spheroids and after glial progenitor cell transplantation into the rat forebrain. Furthermore, we show the robustness of this targeting by transferring the capsid peptide from the parent AAV2 serotype onto the AAV9 serotype, which facilitates future scalability for the larger human brain.

Multi-Target Neural Differentiation (MTND) Therapeutic Cocktail to Suppress Brain Tumor.

Brain tumors have been proved challenging to treat. Here we established a Multi-Target Neural Differentiation (MTND) therapeutic cocktail to achieve effective and safe treatment of brain malignancies by targeting the important hallmarks in brain cancers: poor cell differentiation and compromised cell cycle. In-vitro and in-vivo experiments confirmed the significant therapeutic effect of our MTND therapy. Significantly improved therapeutic effects over current first-line chemo-drugs have been identified in clinical cells, with great inhibition of the growth and migration of tumor cells. Further in-vivo experiments confirmed that sustained MTND treatment showed a 73% reduction of the tumor area. MTND also induced strong expression of phenotypes associated with cell cycle exit/arrest and rapid neural reprograming from clinical glioma cells to glutamatergic and GABAergic expressing cells, which are two key neuronal types involved in many human brain functions, including learning and memory. Collectively, MTND induced multi-targeted genotypic expression changes to achieve direct neural conversion of glioma cells and controlled the cell cycle/tumorigenesis development, helping control tumor cells' malignant proliferation and making it possible to treat brain malignant tumors effectively and safely. These encouraging results open avenues to developing new therapies for brain malignancies beyond cytotoxic agents, providing more effective medication recommendations with reduced toxicity.

Neural Rendering-Based 3D Scene Style Transfer Method via Semantic Understanding Using a Single Style Image

In the rapidly emerging era of untact (“contact-free”) technologies, the requirement for three-dimensional (3D) virtual environments utilized in virtual reality (VR)/augmented reality (AR) and the metaverse has seen significant growth, owing to their extensive application across various domains. Current research focuses on the automatic transfer of the style of rendering images within a 3D virtual environment using artificial intelligence, which aims to minimize human intervention. However, the prevalent studies on rendering-based 3D environment-style transfers have certain inherent limitations. First, the training of a style transfer network dedicated to 3D virtual environments demands considerable style image data. These data must align with viewpoints that closely resemble those of the virtual environment. Second, there was noticeable inconsistency within the 3D structures. Predominant studies often neglect 3D scene geometry information instead of relying solely on 2D input image features. Finally, style adaptation fails to accommodate the unique characteristics inherent in each object. To address these issues, we propose a novel approach: a neural rendering-based 3D scene-style conversion technique. This methodology employs semantic nearest-neighbor feature matching, thereby facilitating the transfer of style within a 3D scene while considering the distinctive characteristics of each object, even when employing a single style image. The neural radiance field enables the network to comprehend the geometric information of a 3D scene in relation to its viewpoint. Subsequently, it transfers style features by employing the unique features of a single style image via semantic nearest-neighbor feature matching. In an empirical context, our proposed semantic 3D scene style transfer method was applied to 3D scene style transfers for both interior and exterior environments. This application utilizes the replica, 3DFront, and Tanks and Temples datasets for testing. The results illustrate that the proposed methodology surpasses existing style transfer techniques in terms of maintaining 3D viewpoint consistency, style uniformity, and semantic coherence.

A neural network approach to the estimation of in-water attenuation to absorption ratios from PACE mission measurements

Introduction: In preparation for the upcoming PACE mission, we explore the feasibility of a neural network-based approach for the conversion of measurements of the degree of linear polarization at the top of the atmosphere as carried out by the HARP2 instrument into estimations of the ratio of attenuation to absorption in the surface layer of the ocean. Polarization has been shown to contain information on the in-water inherent optical properties including the total attenuation coefficient, in contrast with approaches solely based on remote sensing reflectance that are limited to the backscattered fraction of the scattering. In turn, these properties may be further combined with inversion algorithms to retrieve projected values for the optical and physical properties of marine particulates.Methodology: Using bio-optical models to produce synthetic data in quantities sufficient for network training purposes, and with associated polarization values derived from vector radiative transfer modeling, we produce a two-step algorithm that retrieves surface-level polarization first and attenuation-to-absorption ratios second, with each step handled by a separate neural network. The networks use multispectral inputs in terms of the degree of linear polarization from the polarimeter and the remote sensing reflectance from the Ocean Color Instrument that are anticipated to be fully available within the PACE data environment.Result and Discussion: Produce results that compare favorably with expected values, suggesting that a neural network-mediated conversion of remotely sensed polarization into in-water IOPs is viable. A simulation of the PACE orbit and of the HARP2 field of view further shows these results to be robust even over the limited number of data points expected to be available for any given point on Earth’s surface over a single PACE transit.

Human-Induced Pluripotent Stem Cell Technology: Toward the Future of Personalized Psychiatry.

The polygenic and multifactorial nature of many psychiatric disorders has hampered implementation of the personalized medicine approach in clinical practice. However, induced pluripotent stem cell (iPSC) technology has emerged as an innovative tool for patient-specific disease modeling to expand the pathophysiology knowledge and treatment perspectives in the last decade. Current technologies enable adult human somatic cell reprogramming into iPSCs to generate neural cells and direct neural cell conversion to model organisms that exhibit phenotypes close to human diseases, thereby effectively representing relevant aspects of neuropsychiatric disorders. In this regard, iPSCs reflect patient pathophysiology and pharmacological responsiveness, particularly when cultured under conditions that emulate spatial tissue organization in brain organoids. Recently, the application of iPSCs has been frequently associated with gene editing that targets the disease-causing gene to deepen the illness pathophysiology and to conduct drug screening. Moreover, gene editing has provided a unique opportunity to repair the putative causative genetic lesions in patient-derived cells. Here, we review the use of iPSC technology to model and potentially treat neuropsychiatric disorders by illustrating the key studies on a series of mental disorders, including schizophrenia, major depressive disorder, bipolar disorder, and autism spectrum disorder. Future perspectives will involve the development of organ-on-a-chip platforms that control the microenvironmental conditions so as to reflect individual pathophysiological by adjusting physiochemical parameters according to personal health data. This strategy could open new ways by which to build a disease model that considers individual variability and tailors personalized treatments.

Aggregation of cryopreserved mid-hindgut endoderm for more reliable and reproducible hPSC-derived small intestinal organoid generation.

SummaryA major technical limitation hindering the widespread adoption of human pluripotent stem cell (hPSC)-derived gastrointestinal (GI) organoid technologies is the need for de novo hPSC differentiation and dependence on spontaneous morphogenesis to produce detached spheroids. Here, we report a method for simple, reproducible, and scalable production of small intestinal organoids (HIOs) based on the aggregation of cryopreservable hPSC-derived mid-hindgut endoderm (MHE) monolayers. MHE aggregation eliminates variability in spontaneous spheroid production and generates HIOs that are comparable to those arising spontaneously. With a minor modification to the protocol, MHE can be cryopreserved, thawed, and aggregated, facilitating HIO production without de novo hPSC differentiation. Finally, aggregation can also be used to generate antral stomach organoids and colonic organoids. This improved method removes significant barriers to the implementation and successful use of hPSC-derived GI organoid technologies and provides a framework for improved dissemination and increased scalability of GI organoid production.

Mitochondria and Other Organelles in Neural Development and Their Potential as Therapeutic Targets in Neurodegenerative Diseases.

The contribution of organelles to neural development has received increasing attention. Studies have shown that organelles such as mitochondria, endoplasmic reticulum (ER), lysosomes, and endosomes play important roles in neurogenesis. Specifically, metabolic switching, reactive oxygen species production, mitochondrial dynamics, mitophagy, mitochondria-mediated apoptosis, and the interaction between mitochondria and the ER all have roles in neurogenesis. Lysosomes and endosomes can regulate neurite growth and extension. Moreover, metabolic reprogramming represents a novel strategy for generating functional neurons. Accordingly, the exploration and application of mechanisms underlying metabolic reprogramming will be beneficial for neural conversion and regenerative medicine. There is adequate evidence implicating the dysfunction of cellular organelles—especially mitochondria—in neurodegenerative disorders, and that improvement of mitochondrial function may reverse the progression of these diseases through the reinforcement of adult neurogenesis. Therefore, these organelles have potential as therapeutic targets for the treatment of neurodegenerative diseases. In this review, we discuss the function of these organelles, especially mitochondria, in neural development, focusing on their potential as therapeutic targets in neurodegenerative disorders, including Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, and amyotrophic lateral sclerosis.

Neural speech-rate conversion with multispeaker WaveNet vocoder

Speech-rate conversion technology, which can expand or compress speech waveforms while preserving the pitch of the sound, is traditionally realized by signal-processing-based approaches. To improve the synthesis quality, this paper proposes a machine-learning-based approach using neural vocoders, to perform neural speech-rate conversion. The proposed approach introduces a multispeaker WaveNet vocoder trained with a multispeaker corpus. Speech-rate conversion for many and unspecified speakers, not included in the training data, is realized by resampling acoustic features or hidden features along the time direction in inference. In experiments, the multispeaker WaveNet vocoder was trained using the JVS corpus and two types of resampling methods were compared. Conventional WSOLA and STRAIGHT were also compared as signal-processing-based baselines. The test sets included Japanese speaker corpora for the monolingual condition, and an English multispeaker corpus (CMU ARCTIC) for the cross-lingual condition. The results of the experiments demonstrate that the proposed approach with resampling of hidden features can achieve higher quality speech-rate conversion than the conventional methods, in both monolingual and cross-lingual conditions, except for speakers with low fundamental frequency in conversion of fast speech.

Cone beam CT based validation of neural network generated synthetic CTs for radiotherapy in the head region.

In the past years, many different neural network-based conversion techniques for synthesizing computed tomographys (sCTs) from MR images have been published. While the model's performance can be checked during the training against the test set, test datasets can never represent the whole population. Conversion errors can still occur for special cases, for example, for unusual anatomical situations. Therefore, the performance of sCT conversion needs to be verified on a patient specific level, especially in the absence of a planning CT (pCT). In this study, the capability of cone-beam CTs (CBCTs) for the validation of sCTs generated by a neural network was investigated. 41 patients with tumors in the head region were selected. 20 of them were used for model training and 10 for validation. Different implementations of CycleGAN (with/without identity and feature loss) were used to generate sCTs. The pixel (MAE, RMSE, PSNR) and geometric error (DICE, Sensitivity, Specificity) values were reported to identify the best model. VMAT plans were created for the remaining 11 patients on the pCTs. These plans were re-calculated on sCTs and CBCTs. An automatic density overriding method ( ) and a population-based dose calculation method ( ) were employed for CBCT-based dose calculation. The dose distributions were analysed using 3D global gamma analysis, applying a threshold of 10% with respect to the prescribed dose. Differences in DVH metrics for the PTV and the organs-at-risk were compared among the dose distributions based on pCTs, sCTs, and CBCTs. The best model was the CycleGAN without identity and feature matching loss. Including the identity loss led to a metric decrease of 10% for DICE and a metric increase of 20-60HU for MAE. Using the 2%/2mm gamma criterion and pCT as reference, the mean gamma pass rates were 99.0 0.4% for sCTs. Mean gamma pass rate values comparing pCT and CBCT were 99.0 0.8% and 99.1 0.8% for the and , respectively. The mean gamma pass rates comparing sCT and CBCT resulted in 98.4 1.6% and 99.2 0.6% for and , respectively. The differences between the gamma-pass-rates of the sCT and two CBCT-based methods were not significant. The majority of deviations of the investigated DVH metrices between sCTs and CBCTs were within 2%. The dosimetric results demonstrate good agreement between sCT, CBCT, and pCT based calculations. A properly applied CBCT conversion method can serve as a tool for quality assurance procedures in an MR only radiotherapy workflow for head patients. Dosimetric deviations of DVH metrics between sCT and CBCTs of larger than 2% should be followed up. A systematic shift of approximately 1% should be taken into account when using the approach in an MR only workflow.

Present State and Future Perspectives of Prostaglandins as a Differentiation Factor in Motor Neurons.

Spinal motor neurons have the longest axons that innervate the skeletal muscles of the central nervous system. Motor neuron diseases caused by spinal motor neuron cell death are incurable due to the unique and irreplaceable nature of their neural circuits. Understanding the mechanisms of neurogenesis, neuritogenesis, and synaptogenesis in motor neurons will allow investigators to develop new in vitro models and regenerative therapies for motor neuron diseases. In particular, small molecules can directly reprogram and convert into neural stem cells and neurons, and promote neuron-like cell differentiation. Prostaglandins are known to have a role in the differentiation and tissue regeneration of several cell types and organs. However, the involvement of prostaglandins in the differentiation of motor neurons from neural stem cells is poorly understood. The general cell line used in research on motor neuron diseases is the mouse neuroblastoma and spinal motor neuron fusion cell line NSC-34. Recently, our laboratory reported that prostaglandin E2 and prostaglandin D2 enhanced the conversion of NSC-34 cells into motor neuron-like cells with neurite outgrowth. Moreover, we found that prostaglandin E2-differentiated NSC-34 cells had physiological and electrophysiological properties of mature motor neurons. In this review article, we provide contemporary evidence on the effects of prostaglandins, particularly prostaglandin E2 and prostaglandin D2, on differentiation and neural conversion. We also discuss the potential of prostaglandins as candidates for the development of new therapeutic drugs for motor neuron diseases.

A quantitative model of cellular decision making in direct neuronal reprogramming

The direct reprogramming of adult skin fibroblasts to neurons is thought to be controlled by a small set of interacting gene regulators. Here, we investigate how the interaction dynamics between these regulating factors coordinate cellular decision making in direct neuronal reprogramming. We put forward a quantitative model of the governing gene regulatory system, supported by measurements of mRNA expression. We found that nPTB needs to feed back into the direct neural conversion network most likely via PTB in order to accurately capture quantitative gene interaction dynamics and correctly predict the outcome of various overexpression and knockdown experiments. This was experimentally validated by nPTB knockdown leading to successful neural conversion. We also proposed a novel analytical technique to dissect system behaviour and reveal the influence of individual factors on resulting gene expression. Overall, we demonstrate that computational analysis is a powerful tool for understanding the mechanisms of direct (neuronal) reprogramming, paving the way for future models that can help improve cell conversion strategies.

Transcriptional Profiling During Neural Conversion.

The processes that underlie neuronal conversion ultimately involve a reorganization of transcriptional networks to establish a neuronal cell fate. As such, transcriptional profiling is a key component toward understanding this process. In this chapter, we will discuss methods of elucidating transcriptional networks during neuronal reprogramming and considerations that should be incorporated in experimental design.

Linguistic-Coupled Age-to-Age Voice Translation to Improve Speech Recognition Performance in Real Environments

We address a low-performance problem of the elderly in automatic speech recognition (ASR) through feature adaptation agnostic to the ASR. Most of the datasets for speech recognition models consist of datasets collected from adult speakers. Consequently, the majority of commercial speech recognition systems typically tend to perform well on adult speakers. In other words, the limited diversity of speakers in the training datasets yields unreliable performance for minority (e.g., elderly) speakers due to the infeasible acquisition of training data. In response, this paper suggests a neural network-based voice conversion framework to enhance speech recognition of the minority. To this end, we propose a voice translation model including an unsupervised phonology clustering to extract linguistic information to fit the minority’s speech to a current acoustic model frame. Our proposal is a spectral feature adaptation method that can be placed in front of any commercial or open ASR system, avoiding directly modifying the speech recognizer. The experimental results and analysis demonstrate the effectiveness of our proposed method through improvement in elderly speech recognition accuracy.

Direct Reprogramming of Human Fetal- and Stem Cell-Derived Glial Progenitor Cells into Midbrain Dopaminergic Neurons.

SummaryHuman glial progenitor cells (hGPCs) are promising cellular substrates to explore for the in situ production of new neurons for brain repair. Proof of concept for direct neuronal reprogramming of glial progenitors has been obtained in mouse models in vivo, but conversion using human cells has not yet been demonstrated. Such studies have been difficult to perform since hGPCs are born late during human fetal development, with limited accessibility for in vitro culture. In this study, we show proof of concept of hGPC conversion using fetal cells and also establish a renewable and reproducible stem cell-based hGPC system for direct neural conversion in vitro. Using this system, we have identified optimal combinations of fate determinants for the efficient dopaminergic (DA) conversion of hGPCs, thereby yielding a therapeutically relevant cell type that selectively degenerates in Parkinson's disease. The induced DA neurons show a progressive, subtype-specific phenotypic maturation and acquire functional electrophysiological properties indicative of DA phenotype.

Loss-of-function Mutations of CUL3, a High Confidence Gene for Psychiatric Disorders, Lead to Aberrant Neurodevelopment In Human Induced Pluripotent Stem Cells

Both rare, high risk, loss-of-function mutations and common, low risk, genetic variants in the CUL3 gene are strongly associated with neuropsychiatric disorders. Network analyses of neuropsychiatric risk genes have shown high CUL3 expression in the prenatal human brain and an enrichment in neural precursor cells (NPCs) and cortical neurons. The role of CUL3 in human neurodevelopment however, is poorly understood. In the present study, we used CRISPR/Cas9 nickase to knockout CUL3 in human induced pluripotent stem cells (iPSCs). iPSCs were subsequently differentiated into cortical glutamatergic neurons using two different protocols and tested for structural/functional alterations. Immunocytochemical analysis and transcriptomic profiling revealed that pluripotency of heterozygous CUL3 knockout (KO) iPSCs remained unchanged compared to isogenic control iPSCs. Following small molecule-mediated differentiation into cortical glutamatergic neurons however, we detected a significant delay in transition from proliferating radial glia cells/NPCs to postmitotic neurons in CUL3 KO cultures. Notably, direct neural conversion of CUL3 KO iPSCs by lentiviral expression of Neurogenin-2 massively attenuated the neurodevelopmental delay. However, both optogenetic and electrical stimulation of induced neurons revealed decreased excitability in Cullin-3 deficient cultures, while basal synaptic transmission remained unchanged. Analysis of target gene expression pointed to alterations in FGF signaling in CUL3 KO NPCs, which is required for NPC proliferation and self-renewal, while RhoA and Notch signaling appeared unaffected. Our data provide first evidence for a major role of Cullin-3 in neuronal differentiation, and for neurodevelopmental deficits underlying neuropsychiatric disorders associated with CUL3 mutations.