Clin Implant Dent Relat Res. 2024 Jan 29. doi: 10.1111/cid.13307. Online ahead of print.

ABSTRACT

Artificial intelligence represented by deep learning has attracted attention in the field of dental implant restoration. It is widely used in surgical image analysis, implant plan design, prosthesis shape design, and prognosis judgment. This article mainly describes the research progress of deep learning in the whole process of dental implant prosthodontics. It analyzes the limitations of current research, and looks forward to the future development direction.

PMID:38286659 | DOI:10.1111/cid.13307

Comput Biol Med. 2024 Jan 17;170:107989. doi: 10.1016/j.compbiomed.2024.107989. Online ahead of print.

ABSTRACT

Accurate segmentation of the pancreas from abdominal computed tomography (CT) images is challenging but essential for the diagnosis and treatment of pancreatic disorders such as tumours and diabetes. In this study, a dataset with 229 sets of high-resolution CT images was generated and annotated. We proposed a novel 3D segmentation model named nnTransfer (nonisomorphic transfer learning) net, which employs generative model structure for self-supervision to facilitate the network's learning of image attributes from unlabelled data. The effectiveness for pancreas segmentation of nnTransfer was assessed using the Hausdorff distance (HD) and Dice similarity coefficient (DSC) on the dataset. Additionally, a histogram analysis with local thresholding was used to achieve automated whole-volume measurement of pancreatic fat (fat volume fraction, FVF). The proposed technique performed admirably on the dataset, with DSC: 0.937 ± 0.019 and HD: 2.655 ± 1.479. The mean pancreas volume and FVF of the pancreas were 91.95 ± 23.90 cm3 and 12.67 % ± 9.84 %, respectively. The nnTransfer functioned flawlessly and autonomously, facilitating the use of the FVF to evaluate pancreatic disease, particularly in patients with diabetes.

PMID:38286105 | DOI:10.1016/j.compbiomed.2024.107989

Comput Biol Med. 2024 Jan 20;170:107983. doi: 10.1016/j.compbiomed.2024.107983. Online ahead of print.

ABSTRACT

Magnetic resonance (MR) image-guided radiotherapy is widely used in the treatment planning of malignant tumors, and MR-only radiotherapy, a representative of this technique, requires synthetic computed tomography (sCT) images for effective radiotherapy planning. Convolutional neural networks (CNN) have shown remarkable performance in generating sCT images. However, CNN-based models tend to synthesize more low-frequency components and the pixel-wise loss function usually used to optimize the model can result in blurred images. To address these problems, a frequency attention conditional generative adversarial network (FACGAN) is proposed in this paper. Specifically, a frequency cycle generative model (FCGM) is designed to enhance the inter-mapping between MR and CT and extract more rich tissue structure information. Additionally, a residual frequency channel attention (RFCA) module is proposed and incorporated into the generator to enhance its ability in perceiving the high-frequency image features. Finally, high-frequency loss (HFL) and cycle consistency high-frequency loss (CHFL) are added to the objective function to optimize the model training. The effectiveness of the proposed model is validated on pelvic and brain datasets and compared with state-of-the-art deep learning models. The results show that FACGAN produces higher-quality sCT images while retaining clearer and richer high-frequency texture information.

PMID:38286104 | DOI:10.1016/j.compbiomed.2024.107983

J Biomech. 2024 Jan 19;163:111959. doi: 10.1016/j.jbiomech.2024.111959. Online ahead of print.

ABSTRACT

Single bicycle crashes, i.e., falls and impacts not involving a collision with another road user, are a significantly underestimated road safety problem. The motions and behaviours of falling people, or fall kinematics, are often investigated in the injury biomechanics research field. Understanding the mechanics of a fall can help researchers develop better protective gear and safety measures to reduce the risk of injury. However, little is known about cyclist fall kinematics or dynamics. Therefore, in this study, a video analysis of cyclist falls is performed to investigate common kinematic forms and impact patterns. Furthermore, a pipeline involving deep learning-based human pose estimation and inverse kinematics optimisation is created for extracting human motion from real-world footage of falls to initialise forward dynamics computational human body models. A bracing active response is then optimised for using a genetic algorithm. This is then applied to a case study of a cyclist fall. The kinematic forms characterised in this study can be used to inform initial conditions for computational modelling and injury estimation in cyclist falls. Findings indicate that protective response is an important consideration in fall kinematics and dynamics, and should be included in computational modelling. Furthermore, the novel reconstruction pipeline proposed here can be applied more broadly for traumatic injury biomechanics tasks. The tool developed in this study is available at https://kevgildea.github.io/KinePose/kevgildea.github.io/KinePose/.

PMID:38286096 | DOI:10.1016/j.jbiomech.2024.111959

Phys Med Biol. 2024 Jan 29. doi: 10.1088/1361-6560/ad237f. Online ahead of print.

ABSTRACT

Quantitative susceptibility mapping (QSM) is a new imaging technique for non-invasive characterization of the composition and microstructure of in vivo tissues, and it can be reconstructed from local field measurements by solving an ill-posed inverse problem. Even for deep learning networks, it is not an easy task to establish an accurate quantitative mapping between two physical quantities of different units, i.e., field shift in Hz and susceptibility value in ppm for QSM.
Approach: In this paper, we propose a spatially adaptive regularization based three-dimensional reconstruction network SAQSM. A spatially adaptive module is specially designed and a set of them at different resolutions are inserted into the network decoder, playing a role of cross-modality based regularization constraint. Therefore, the exact information of both field and magnitude data is exploited to adjust the scale and shift of feature maps, and thus any information loss or deviation occurred in previous layers could be effectively corrected. The network encoding has a dynamic perceptual initialization, which enables the network to overcome receptive field intervals and also strengthens its ability to detect features of various sizes.
Main results: Experimental results on the brain data of healthy volunteers, clinical hemorrhage and simulated phantom with calcification demonstrate that SAQSM can achieve more accurate reconstruction with less susceptibility artifacts, while perform well on the stability and generalization even for severe lesion areas.
Significance: This proposed framework may provide a valuable paradigm to quantitative mapping or multimodal reconstruction.

PMID:38286013 | DOI:10.1088/1361-6560/ad237f

Proc Natl Acad Sci U S A. 2024 Feb 6;121(6):e2314853121. doi: 10.1073/pnas.2314853121. Epub 2024 Jan 29.

ABSTRACT

Amino acid mutations that lower a protein's thermodynamic stability are implicated in numerous diseases, and engineered proteins with enhanced stability can be important in research and medicine. Computational methods for predicting how mutations perturb protein stability are, therefore, of great interest. Despite recent advancements in protein design using deep learning, in silico prediction of stability changes has remained challenging, in part due to a lack of large, high-quality training datasets for model development. Here, we describe ThermoMPNN, a deep neural network trained to predict stability changes for protein point mutations given an initial structure. In doing so, we demonstrate the utility of a recently released megascale stability dataset for training a robust stability model. We also employ transfer learning to leverage a second, larger dataset by using learned features extracted from ProteinMPNN, a deep neural network trained to predict a protein's amino acid sequence given its three-dimensional structure. We show that our method achieves state-of-the-art performance on established benchmark datasets using a lightweight model architecture that allows for rapid, scalable predictions. Finally, we make ThermoMPNN readily available as a tool for stability prediction and design.

PMID:38285937 | DOI:10.1073/pnas.2314853121

Neuro Oncol. 2024 Jan 29:noae017. doi: 10.1093/neuonc/noae017. Online ahead of print.

ABSTRACT

BACKGROUND: The aim was to predict survival of glioblastoma at eight months after radiotherapy (a period allowing for completing a typical course of adjuvant temozolomide), by applying deep learning to the first brain MRI after radiotherapy completion.

METHODS: Retrospective and prospective data were collected from 206 consecutive glioblastoma, IDH-wildtype patients diagnosed between March 2014-February 2022 across 11 UK centers. Models were trained on 158 retrospective patients from three centers. Holdout test sets were retrospective (n=19; internal validation), and prospective (n=29; external validation from eight distinct centers).Neural network branches for T2-weighted and contrast-enhanced T1-weighted inputs were concatenated to predict survival. A non-imaging branch (demographics/MGMT/treatment data) was also combined with the imaging model. We investigated the influence of individual MR sequences; non-imaging features; and weighted dense blocks pretrained for abnormality detection.

RESULTS: The imaging model outperformed the non-imaging model in all test sets (area under the receiver-operating characteristic curve, AUC p=0.038) and performed similarly to a combined imaging/non-imaging model (p>0.05). Imaging, non-imaging, and combined models applied to amalgamated test sets gave AUCs of 0.93, 0.79, and 0.91. Initializing the imaging model with pretrained weights from 10,000s of brain MRIs improved performance considerably (amalgamated test sets without pretraining 0.64; p=0.003).

CONCLUSIONS: A deep learning model using MRI images after radiotherapy, reliably and accurately determined survival of glioblastoma. The model serves as a prognostic biomarker identifying patients who will not survive beyond a typical course of adjuvant temozolomide, thereby stratifying patients into those who might require early second-line or clinical trial treatment.

PMID:38285679 | DOI:10.1093/neuonc/noae017

IEEE Trans Neural Syst Rehabil Eng. 2024 Jan 29;PP. doi: 10.1109/TNSRE.2024.3359191. Online ahead of print.

ABSTRACT

Rapid serial visual presentation (RSVP)-based brain-computer interface (BCI) is a promising target detection technique by using electroencephalogram (EEG) signals. However, existing deep learning approaches seldom considered dependencies of multi-scale temporal features and discriminative multi-view spectral features simultaneously, which limits the representation learning ability of the model and undermine the EEG classification performance. In addition, recent transfer learning-based methods generally failed to obtain transferable cross-subject invariant representations and commonly ignore the individual-specific information, leading to the poor cross-subject transfer performance. In response to these limitations, we propose a cross-scale Transformer and triple-view attention based domain-rectified transfer learning (CST-TVA-DRTL) for the RSVP classification. Specially, we first develop a cross-scale Transformer (CST) to extract multi-scale temporal features and exploit the dependencies of different scales features. Then, a triple-view attention (TVA) is designed to capture spectral features from triple views of multi-channel time-frequency images. Finally, a domain-rectified transfer learning (DRTL) framework is proposed to simultaneously obtain transferable domain-invariant representations and untransferable domain-specific representations, then utilize domain-specific information to rectify domain-invariant representations to adapt to target data. Experimental results on two public RSVP datasets suggests that our CST-TVA-DRTL outperforms the state-of-the-art methods in the RSVP classification task. The source code of our model is publicly available in https://github.com/ljbuaa/CST_TVA_DRTL.

PMID:38285586 | DOI:10.1109/TNSRE.2024.3359191

IEEE Trans Biomed Eng. 2024 Jan 29;PP. doi: 10.1109/TBME.2024.3359752. Online ahead of print.

ABSTRACT

The Ear-ECG provides a continuous Lead I like electrocardiogram (ECG) by measuring the potential difference related to heart activity by electrodes which are embedded within earphones. However, the significant increase in wearability and comfort enabled by Ear-ECG is often accompanied by a degradation in signal quality - a common obstacle that is shared by the majority of wearable technologies. We aim to resolve this issue by introducing a Deep Matched Filter (Deep-MF) for the highly accurate detection of R-peaks in wearable ECG, thus enhancing the utility of Ear-ECG in real-world scenarios. The Deep-MF consists of an encoder stage (trained as part of an encoder-decoder module to reproduce ground truth ECG), and an R-peak classifier stage. Through its operation as a Matched Filter, the encoder section searches for matches with an ECG template pattern in the input signal, prior to filtering these matches with the subsequent convolutional layers and selecting peaks corresponding to the ground truth ECG. The so condensed latent representation of R-peak information is then fed into a simple R-peak classifier, of which the output provides precise R-peak locations. The proposed Deep Matched Filter is evaluated using leave-one-subject-out cross-validation over 36 subjects with an age range of 18-75, with the Deep-MF outperforming existing algorithms for R-peak detection in noisy ECG. The Deep-MF is benchmarked against a ground truth ECG, in the form of either chest-ECG or arm-ECG, via both R-peak recall and R-peak precision metrics. The Deep-MF achieves a median R-peak recall of 94.9% and a median precision of 91.2% across subjects when evaluated with leave-one-subject-out cross validation. Moreover, when evaluated across a range of thresholds, the Deep-MF achieves an area under the curve (AUC) value of 0.97. The interpretability of Deep-MF as a Matched Filter is further strengthened by the analysis of its response to partial initialisation with an ECG template. We demonstrate that the Deep Matched Filter algorithm not only retains the initialised ECG kernel structure during the training process, but also amplifies portions of the ECG which it deems most valuable - namely the P wave, and each aspect of the QRS complex. Overall, this Deep-Match framework serves as a valuable step forward for the real-world functionality of Ear-ECG and, through its interpretable operation, the acceptance of deep learning models in e-Health.

PMID:38285581 | DOI:10.1109/TBME.2024.3359752

IEEE Trans Neural Netw Learn Syst. 2024 Jan 29;PP. doi: 10.1109/TNNLS.2024.3354982. Online ahead of print.

ABSTRACT

Deep learning methods have achieved impressive performance in compressed video quality enhancement tasks. However, these methods rely excessively on practical experience by manually designing the network structure and do not fully exploit the potential of the feature information contained in the video sequences, i.e., not taking full advantage of the multiscale similarity of the compressed artifact information and not seriously considering the impact of the partition boundaries in the compressed video on the overall video quality. In this article, we propose a novel Mixed Difference Equation inspired Transformer (MDEformer) for compressed video quality enhancement, which provides a relatively reliable principle to guide the network design and yields a new insight into the interpretable transformer. Specifically, drawing on the graphical concept of the mixed difference equation (MDE), we utilize multiple cross-layer cross-attention aggregation (CCA) modules to establish long-range dependencies between encoders and decoders of the transformer, where partition boundary smoothing (PBS) modules are inserted as feedforward networks. The CCA module can make full use of the multiscale similarity of compression artifacts to effectively remove compression artifacts, and recover the texture and detail information of the frame. The PBS module leverages the sensitivity of smoothing convolution to partition boundaries to eliminate the impact of partition boundaries on the quality of compressed video and improve its overall quality, while not having too much impacts on non-boundary pixels. Extensive experiments on the MFQE 2.0 dataset demonstrate that the proposed MDEformer can eliminate compression artifacts for improving the quality of the compressed video, and surpasses the state-of-the-arts (SOTAs) in terms of both objective metrics and visual quality.

PMID:38285580 | DOI:10.1109/TNNLS.2024.3354982

IEEE Trans Image Process. 2024 Jan 29;PP. doi: 10.1109/TIP.2024.3357250. Online ahead of print.

ABSTRACT

Deep learning-based hyperspectral image (HSI) classification methods have recently shown excellent performance, however, there are two shortcomings that need to be addressed. One is that deep network training requires a large number of labeled images, and the other is that deep network needs to learn a large number of parameters. They are also general problems of deep networks, especially in applications that require professional techniques to acquire and label images, such as HSI and medical images. In this paper, we propose a deep network architecture (SAFDNet) based on the stochastic adaptive Fourier decomposition (SAFD) theory. SAFD has powerful unsupervised feature extraction capabilities, so the entire deep network only requires a small number of annotated images to train the classifier. In addition, we use fewer convolution kernels in the entire deep network, which greatly reduces the number of deep network parameters. SAFD is a newly developed signal processing tool with solid mathematical foundation, which is used to construct the unsupervised deep feature extraction mechanism of SAFDNet. Experimental results on three popular HSI classification datasets show that our proposed SAFDNet outperforms other compared state-of-the-art deep learning methods in HSI classification.

PMID:38285575 | DOI:10.1109/TIP.2024.3357250

Biomed Tech (Berl). 2024 Jan 30. doi: 10.1515/bmt-2022-0336. Online ahead of print.

ABSTRACT

OBJECTIVES: Brain tumor classification is amongst the most complex and challenging jobs in the computer domain. The latest advances in brain tumor detection systems (BTDS) are presented as they can inspire new researchers to deliver new architectures for effective and efficient tumor detection. Here, the data of the multi-modal brain tumor segmentation task is employed, which has been registered, skull stripped, and histogram matching is conducted with the ferrous volume of high contrast.

METHODS: This research further configures a capsule network (CapsNet) for brain tumor classification. Results of the latest deep neural network (NN) architectures for tumor detection are compared and presented. The VGG16 and CapsNet architectures yield the highest f1-score and precision values, followed by VGG19. Overall, ResNet152, MobileNet, and MobileNetV2 give us the lowest f1-score.

RESULTS: The VGG16 and CapsNet have produced outstanding results. However, VGG16 and VGG19 are more profound architecture, resulting in slower computation speed. The research then recommends the latest suitable NN for effective brain tumor detection.

CONCLUSIONS: Finally, the work concludes with future directions and potential new architectures for tumor detection.

PMID:38285486 | DOI:10.1515/bmt-2022-0336

Transl Vis Sci Technol. 2024 Jan 2;13(1):23. doi: 10.1167/tvst.13.1.23.

ABSTRACT

PURPOSE: To develop and evaluate a deep learning (DL) model to assess fundus photograph quality, and quantitatively measure its impact on automated POAG detection in independent study populations.

METHODS: Image quality ground truth was determined by manual review of 2815 fundus photographs of healthy and POAG eyes from the Diagnostic Innovations in Glaucoma Study and African Descent and Glaucoma Evaluation Study (DIGS/ADAGES), as well as 11,350 from the Ocular Hypertension Treatment Study (OHTS). Human experts assessed a photograph as high quality if of sufficient quality to determine POAG status and poor quality if not. A DL quality model was trained on photographs from DIGS/ADAGES and tested on OHTS. The effect of DL quality assessment on DL POAG detection was measured using area under the receiver operating characteristic (AUROC).

RESULTS: The DL quality model yielded an AUROC of 0.97 for differentiating between high- and low-quality photographs; qualitative human review affirmed high model performance. Diagnostic accuracy of the DL POAG model was significantly greater (P < 0.001) in good (AUROC, 0.87; 95% CI, 0.80-0.92) compared with poor quality photographs (AUROC, 0.77; 95% CI, 0.67-0.88).

CONCLUSIONS: The DL quality model was able to accurately assess fundus photograph quality. Using automated quality assessment to filter out low-quality photographs increased the accuracy of a DL POAG detection model.

TRANSLATIONAL RELEVANCE: Incorporating DL quality assessment into automated review of fundus photographs can help to decrease the burden of manual review and improve accuracy for automated DL POAG detection.

PMID:38285462 | DOI:10.1167/tvst.13.1.23

Eur Radiol. 2024 Jan 29. doi: 10.1007/s00330-024-10584-z. Online ahead of print.

ABSTRACT

BACKGROUND: Intravenous thrombolysis (IVT) before endovascular treatment (EVT) for acute ischemic stroke might induce intracerebral hemorrhages which could negatively affect patient outcomes. Measuring white matter lesions size using deep learning (DL-WML) might help safely guide IVT administration. We aimed to develop, validate, and evaluate a DL-WML volume on CT compared to the Fazekas scale (WML-Faz) as a risk factor and IVT effect modifier in patients receiving EVT directly after IVT.

METHODS: We developed a deep-learning model for WML segmentation on CT and validated with internal and external test sets. In a post hoc analysis of the MR CLEAN No-IV trial, we associated DL-WML volume and WML-Faz with symptomatic-intracerebral hemorrhage (sICH) and 90-day functional outcome according to the modified Rankin Scale (mRS). We used multiplicative interaction terms between WML measures and IVT administration to evaluate IVT treatment effect modification. Regression models were used to report unadjusted and adjusted common odds ratios (cOR/acOR).

RESULTS: In total, 516 patients from the MR CLEAN No-IV trial (male/female, 291/225; age median, 71 [IQR, 62-79]) were analyzed. Both DL-WML volume and WML-Faz are associated with sICH (DL-WML volume acOR, 1.78 [95%CI, 1.17; 2.70]; WML-Faz acOR, 1.53 95%CI [1.02; 2.31]) and mRS (DL-WML volume acOR, 0.70 [95%CI, 0.55; 0.87], WML-Faz acOR, 0.73 [95%CI 0.60; 0.88]). Only in the unadjusted IVT effect modification analysis WML-Faz was associated with more sICH if IVT was given (p = 0.046). Neither WML measure was associated with worse mRS if IVT was given.

CONCLUSION: DL-WML volume and WML-Faz had a similar relationship with functional outcome and sICH. Although more sICH might occur in patients with more severe WML-Faz receiving IVT, no worse functional outcome was observed.

CLINICAL RELEVANCE STATEMENT: White matter lesion severity on baseline CT in acute ischemic stroke patients has a similar predictive value if measured with deep learning or the Fazekas scale. Safe administration of intravenous thrombolysis using white matter lesion severity should be further studied.

KEY POINTS: White matter damage is a predisposing risk factor for intracranial hemorrhage in patients with acute ischemic stroke but remains difficult to measure on CT. White matter lesion volume on CT measured with deep learning had a similar association with symptomatic intracerebral hemorrhages and worse functional outcome as the Fazekas scale. A patient-level meta-analysis is required to study the benefit of white matter lesion severity-based selection for intravenous thrombolysis before endovascular treatment.

PMID:38285103 | DOI:10.1007/s00330-024-10584-z

Ophthalmologie. 2024 Jan 29. doi: 10.1007/s00347-024-01985-y. Online ahead of print.

ABSTRACT

Retinal optical coherence tomography (OCT) biomarkers have the potential to serve as early, noninvasive, and cost-effective markers for identifying individuals at risk for cognitive impairments and neurodegenerative diseases. They may also aid in monitoring disease progression and evaluating the effectiveness of interventions targeting cognitive decline. The association between retinal OCT biomarkers and cognitive performance has been demonstrated in several studies, and their importance in cognitive assessment is increasingly being recognized. Machine learning (ML) is a branch of artificial intelligence (AI) with an exponential number of applications in the medical field, particularly its deep learning (DL) subset, which is widely used for the analysis of medical images. These techniques efficiently deal with novel biomarkers when their outcome for the applications of interest are unclear, e.g., for the diagnosis, prognosis prediction and disease staging. However, using AI-based tools for medical purposes must be approached with caution, despite the many efforts to address the black-box nature of such approaches, especially due to the general underperformance in datasets other than those used for their development. Retinal OCT biomarkers are promising as potential indicators for decline in cognitive function. The underlying mechanisms are currently being explored to gain deeper insights into this relationship linking retinal health and cognitive function. Insights from neurovascular coupling and retinal microvascular changes play an important role. Further research is needed to establish the validity and utility of retinal OCT biomarkers as early indicators of cognitive decline and neurodegenerative diseases in routine clinical practice. Retinal OCT biomarkers could then provide a new avenue for early detection, monitoring and intervention in cognitive impairment with the potential to improve patient care and outcomes.

PMID:38285070 | DOI:10.1007/s00347-024-01985-y

Mol Cancer Res. 2024 Jan 29. doi: 10.1158/1541-7786.MCR-23-0639. Online ahead of print.

ABSTRACT

Prostate cancer (PCa) harbors several genetic alterations, the most prevalent of which is TMPRSS2:ERG gene fusion, affecting nearly half of all cases. Capitalizing on the increasing availability of whole-slide images (WSIs), this study introduces a deep learning (DL) model designed to detect TMPRSS2:ERG fusion from H&E-stained WSIs of radical prostatectomy specimens. Leveraging the TCGA prostate adenocarcinoma cohort, which comprises 436 WSIs from 393 patients, we developed a robust DL model, trained across ten different splits, each consisting of distinct training, validation, and testing sets. The model's best performance achieved an Area Under the ROC curve (AUC) of 0.84 during training, and 0.72 on the TCGA test set. This model was subsequently validated on an independent cohort comprising 314 WSIs from a different institution, in which it has a robust performance at predicting TMPRSS2:ERG fusion with an AUC of 0.73. Importantly, the model identifies highly-attended tissue regions associated with TMPRSS2:ERG fusion, characterized by higher neoplastic cell content and altered immune and stromal profiles compared to fusion-negative cases. Multivariate survival analysis revealed that these morphological features correlate with poorer survival outcomes, independent of Gleason grade and tumor stage. This study underscores the potential of DL in deducing genetic alterations from routine slides and identifying their underlying morphological features which might harbor prognostic information. Implications: Our study illuminates the potential of deep learning in effectively inferring key prostate cancer genetic alterations from the tissue morphology depicted in routinely available histology slides, offering a cost-effective method that could revolutionize diagnostic strategies in oncology.

PMID:38284821 | DOI:10.1158/1541-7786.MCR-23-0639

J Chem Inf Model. 2024 Jan 29. doi: 10.1021/acs.jcim.3c01633. Online ahead of print.

ABSTRACT

Artificial Neural Networks (ANNs) are transforming how we understand chemical mixtures, providing an expressive view of the chemical space and multiscale processes. Their hybridization with physical knowledge can bridge the gap between predictivity and understanding of the underlying processes. This overview explores recent progress in ANNs, particularly their potential in the 'recomposition' of chemical mixtures. Graph-based representations reveal patterns among mixture components, and deep learning models excel in capturing complexity and symmetries when compared to traditional Quantitative Structure-Property Relationship models. Key components, such as Hamiltonian networks and convolution operations, play a central role in representing multiscale mixtures. The integration of ANNs with Chemical Reaction Networks and Physics-Informed Neural Networks for inverse chemical kinetic problems is also examined. The combination of sensors with ANNs shows promise in optical and biomimetic applications. A common ground is identified in the context of statistical physics, where ANN-based methods iteratively adapt their models by blending their initial states with training data. The concept of mixture recomposition unveils a reciprocal inspiration between ANNs and reactive mixtures, highlighting learning behaviors influenced by the training environment.

PMID:38284618 | DOI:10.1021/acs.jcim.3c01633

J Sci Food Agric. 2024 Jan 29. doi: 10.1002/jsfa.13334. Online ahead of print.

ABSTRACT

BACKGROUND: Methylcellulose has been applied as a primary binding agent to control the quality attributes of plant-based meat analogues, however a great deal of efforts have been made to search for hydrocolloids for replacing methylcellulose due to increasing awareness of clean-labels. In this study, a machine learning framework was proposed in order to describe and predict the flow behaviors of six hydrocolloid solutions, and the predicted viscosities were correlated with the textural features of their corresponding plant-based meat analogues.

RESULTS: Different shear-thinning and Newtonian behaviors were observed depending on the type of hydrocolloids and the shear rates. Methylcellulose exhibited an increasing viscosity pattern with increasing temperatures, compared to the other hydrocolloids. The machine learning algorithms (random forest and multilayer perceptron models) showed a better viscosity fitting performance than the constitutive equations (Power-law and Cross models). In addition, three hyperparameters of the multilayer perceptron model (optimizer, learning rate, and the number of hidden layers) were tuned using the Bayesian optimization algorithm.

CONCLUSION: The optimized multilayer perceptron model exhibited the superior performance of the viscosity prediction (R2 = 0.9944 - 0.9961/RMSE = 0.0545 - 0.0708). Furthermore, the machine learning-predicted viscosities overall showed similar patterns with the textural parameters of the meat analogues. This article is protected by copyright. All rights reserved.

PMID:38284425 | DOI:10.1002/jsfa.13334

Biotechnol Bioeng. 2024 Jan 29. doi: 10.1002/bit.28646. Online ahead of print.

ABSTRACT

Advanced process control in the biopharmaceutical industry often lacks real-time measurements due to resource constraints. Raman spectroscopy and Partial Least Squares (PLS) models are often used to monitor bioprocess cultures in real-time. In spite of the ease of training, the accuracy of the PLS model is impacted if it is not used to predict quality attributes for the cell lines it is trained on. To address this issue, a deep convolutional neural network (CNN) is proposed for offline modeling of metabolites using Raman spectroscopy. By utilizing asymmetric least squares smoothing to adjust Raman spectra baselines, a generic training data set is created by amalgamating spectra from various cell lines and operating conditions. This data set, combined with their derivatives, forms a two-dimensional model input. The CNN model is developed and validated for predicting different quality variables against measurements from various continuous and fed-batch experimental runs. Validation results confirm that the deep CNN model is an accurate generic model of the process to predict real-time quality attributes, even in experimental runs not included in the training data. This model is robust and versatile, requiring no recalibration when deployed at different sites to monitor various cell lines and experimental runs.

PMID:38284180 | DOI:10.1002/bit.28646

Ophthalmol Sci. 2023 Nov 17;4(3):100428. doi: 10.1016/j.xops.2023.100428. eCollection 2024 May-Jun.

ABSTRACT

PURPOSE: Nascent geographic atrophy (nGA) refers to specific features seen on OCT B-scans, which are strongly associated with the future development of geographic atrophy (GA). This study sought to develop a deep learning model to screen OCT B-scans for nGA that warrant further manual review (an artificial intelligence [AI]-assisted approach), and to determine the extent of reduction in OCT B-scan load requiring manual review while maintaining near-perfect nGA detection performance.

DESIGN: Development and evaluation of a deep learning model.

PARTICIPANTS: One thousand eight hundred and eighty four OCT volume scans (49 B-scans per volume) without neovascular age-related macular degeneration from 280 eyes of 140 participants with bilateral large drusen at baseline, seen at 6-monthly intervals up to a 36-month period (from which 40 eyes developed nGA).

METHODS: OCT volume and B-scans were labeled for the presence of nGA. Their presence at the volume scan level provided the ground truth for training a deep learning model to identify OCT B-scans that potentially showed nGA requiring manual review. Using a threshold that provided a sensitivity of 0.99, the B-scans identified were assigned the ground truth label with the AI-assisted approach. The performance of this approach for detecting nGA across all visits, or at the visit of nGA onset, was evaluated using fivefold cross-validation.

MAIN OUTCOME MEASURES: Sensitivity for detecting nGA, and proportion of OCT B-scans requiring manual review.

RESULTS: The AI-assisted approach (utilizing outputs from the deep learning model to guide manual review) had a sensitivity of 0.97 (95% confidence interval [CI] = 0.93-1.00) and 0.95 (95% CI = 0.87-1.00) for detecting nGA across all visits and at the visit of nGA onset, respectively, when requiring manual review of only 2.7% and 1.9% of selected OCT B-scans, respectively.

CONCLUSIONS: A deep learning model could be used to enable near-perfect detection of nGA onset while reducing the number of OCT B-scans requiring manual review by over 50-fold. This AI-assisted approach shows promise for substantially reducing the current burden of manual review of OCT B-scans to detect this crucial feature that portends future development of GA.

FINANCIAL DISCLOSURES: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

PMID:38284101 | PMC:PMC10818248 | DOI:10.1016/j.xops.2023.100428