PLoS One. 2024 Jul 8;19(7):e0302497. doi: 10.1371/journal.pone.0302497. eCollection 2024.

ABSTRACT

This paper presents a deep-learning-based method to detect recreational vessels. The method takes advantage of existing underwater acoustic measurements from an Estuarine Soundscape Observatory Network based in the estuaries of South Carolina (SC), USA. The detection method is a two-step searching method, called Deep Scanning (DS), which includes a time-domain energy analysis and a frequency-domain spectrum analysis. In the time domain, acoustic signals with higher energy, measured by sound pressure level (SPL), are labeled for the potential existence of moving vessels. In the frequency domain, the labeled acoustic signals are examined against a predefined training dataset using a neural network. This research builds training data using diverse vessel sound features obtained from real measurements, with a duration between 5.0 seconds and 7.5 seconds and a frequency between 800 Hz to 10,000 Hz. The proposed method was then evaluated using all acoustic data in the years 2017, 2018, and 2021, respectively; a total of approximately 171,262 2-minute.wav files at three deployed locations in May River, SC. The DS detections were compared to human-observed detections for each audio file and results showed the method was able to classify the existence of vessels, with an average accuracy of around 99.0%.

PMID:38976700 | DOI:10.1371/journal.pone.0302497

IEEE Trans Neural Netw Learn Syst. 2024 Jul 8;PP. doi: 10.1109/TNNLS.2024.3418527. Online ahead of print.

ABSTRACT

Learning with little data is challenging but often inevitable in various application scenarios where the labeled data are limited and costly. Recently, few-shot learning (FSL) gained increasing attention because of its generalizability of prior knowledge to new tasks that contain only a few samples. However, for data-intensive models such as vision transformer (ViT), current fine-tuning-based FSL approaches are inefficient in knowledge generalization and, thus, degenerate the downstream task performances. In this article, we propose a novel mask-guided ViT (MG-ViT) to achieve an effective and efficient FSL on the ViT model. The key idea is to apply a mask on image patches to screen out the task-irrelevant ones and to guide the ViT focusing on task-relevant and discriminative patches during FSL. Particularly, MG-ViT only introduces an additional mask operation and a residual connection, enabling the inheritance of parameters from pretrained ViT without any other cost. To optimally select representative few-shot samples, we also include an active learning-based sample selection method to further improve the generalizability of MG-ViT-based FSL. We evaluate the proposed MG-ViT on classification, object detection, and segmentation tasks using gradient-weighted class activation mapping (Grad-CAM) to generate masks. The experimental results show that the MG-ViT model significantly improves the performance and efficiency compared with general fine-tuning-based ViT and ResNet models, providing novel insights and a concrete approach toward generalizing data-intensive and large-scale deep learning models for FSL.

PMID:38976473 | DOI:10.1109/TNNLS.2024.3418527

IEEE Trans Neural Syst Rehabil Eng. 2024 Jul 8;PP. doi: 10.1109/TNSRE.2024.3424669. Online ahead of print.

ABSTRACT

The process of reconstructing underlying cortical and subcortical electrical activities from Electroencephalography (EEG) or Magnetoencephalography (MEG) recordings is called Electrophysiological Source Imaging (ESI). Given the complementarity between EEG and MEG in measuring radial and tangential cortical sources, combined EEG/MEG is considered beneficial in improving the reconstruction performance of ESI algorithms. Traditional algorithms mainly emphasize incorporating predesigned neurophysiological priors to solve the ESI problem. Deep learning frameworks aim to directly learn the mapping from scalp EEG/MEG measurements to the underlying brain source activities in a data-driven manner, demonstrating superior performance compared to traditional methods. However, most of the existing deep learning approaches for the ESI problem are performed on a single modality of EEG or MEG, meaning the complementarity of these two modalities has not been fully utilized. How to fuse the EEG and MEG in a more principled manner under the deep learning paradigm remains a challenging question. This study develops a Multi-Modal Deep Fusion (MMDF) framework using Attention Neural Networks (ANN) to fully leverage the complementary information between EEG and MEG for solving the ESI inverse problem, which is termed as MMDF-ANN. Specifically, our proposed brain source imaging approach consists of four phases, including feature extraction, weight generation, deep feature fusion, and source mapping. Our experimental results on both synthetic dataset and real dataset demonstrated that using a fusion of EEG and MEG can significantly improve the source localization accuracy compared to using a single-modality of EEG or MEG. Compared to the benchmark algorithms, MMDF-ANN demonstrated good stability when reconstructing sources with extended activation areas and situations of EEG/MEG measurements with a low signal-to-noise ratio.

PMID:38976470 | DOI:10.1109/TNSRE.2024.3424669

IEEE Trans Nanobioscience. 2024 Jul 8;PP. doi: 10.1109/TNB.2024.3424576. Online ahead of print.

ABSTRACT

Traditional DNA storage technologies rely on passive filtering methods for error correction during synthesis and sequencing, which result in redundancy and inadequate error correction. Addressing this, the Low Quality Sequence Filter (LQSF) was introduced, an innovative method employing deep learning models to predict high-risk sequences. The LQSF approach leverages a classification model trained on error-prone sequences, enabling efficient pre-sequencing filtration of low-quality sequences and reducing time and resources in subsequent stages. Analysis has demonstrated a clear distinction between high and low-quality sequences, confirming the efficacy of the LQSF method. Extensive training and testing were conducted across various neural networks and test sets. The results showed all models achieving an AUC value above 0.91 on ROC curves and over 0.95 on PR curves across different datasets. Notably, models such as Alexnet, VGG16, and VGG19 achieved a perfect AUC of 1.0 on the Original dataset, highlighting their precision in classification. Further validation using Illumina sequencing data substantiated a strong correlation between model scores and sequence error-proneness, emphasizing the model's applicability. The LQSF method marks a significant advancement in DNA storage technology, introducing active sequence filtering at the encoding stage. This pioneering approach holds substantial promise for future DNA storage research and applications.

PMID:38976468 | DOI:10.1109/TNB.2024.3424576

IEEE Trans Med Imaging. 2024 Jul 8;PP. doi: 10.1109/TMI.2024.3424471. Online ahead of print.

ABSTRACT

Medical image segmentation has been significantly advanced with the rapid development of deep learning (DL) techniques. Existing DL-based segmentation models are typically discriminative; i.e., they aim to learn a mapping from the input image to segmentation masks. However, these discriminative methods neglect the underlying data distribution and intrinsic class characteristics, suffering from unstable feature space. In this work, we propose to complement discriminative segmentation methods with the knowledge of underlying data distribution from generative models. To that end, we propose a novel hybrid diffusion framework for medical image segmentation, termed HiDiff, which can synergize the strengths of existing discriminative segmentation models and new generative diffusion models. HiDiff comprises two key components: discriminative segmentor and diffusion refiner. First, we utilize any conventional trained segmentation models as discriminative segmentor, which can provide a segmentation mask prior for diffusion refiner. Second, we propose a novel binary Bernoulli diffusion model (BBDM) as the diffusion refiner, which can effectively, efficiently, and interactively refine the segmentation mask by modeling the underlying data distribution. Third, we train the segmentor and BBDM in an alternate-collaborative manner to mutually boost each other. Extensive experimental results on abdomen organ, brain tumor, polyps, and retinal vessels segmentation datasets, covering four widely-used modalities, demonstrate the superior performance of HiDiff over existing medical segmentation algorithms, including the state-of-the-art transformer- and diffusion-based ones. In addition, HiDiff excels at segmenting small objects and generalizing to new datasets. Source codes are made available at https://github.com/takimailto/HiDiff.

PMID:38976467 | DOI:10.1109/TMI.2024.3424471

IEEE Trans Ultrason Ferroelectr Freq Control. 2024 Jul 8;PP. doi: 10.1109/TUFFC.2024.3424549. Online ahead of print.

ABSTRACT

Color Doppler echocardiography enables visualization of blood flow within the heart. However, the limited frame rate impedes the quantitative assessment of blood velocity throughout the cardiac cycle, thereby compromising a comprehensive analysis of ventricular filling. Concurrently, deep learning is demonstrating promising outcomes in post-processing of echocardiographic data for various applications. This work explores the use of deep learning models for intracardiac Doppler velocity estimation from a reduced number of filtered I/Q signals. We used a supervised learning approach by simulating patient-based cardiac color Doppler acquisitions and proposed data augmentation strategies to enlarge the training dataset. We implemented architectures based on convolutional neural networks. In particular, we focused on comparing the U-Net model and the recent ConvNeXt models, alongside assessing real-valued versus complex-valued representations. We found that both models outperformed the state-of-the-art autocorrelator method, effectively mitigating aliasing and noise. We did not observe significant differences between the use of real and complex data. Finally, we validated the models on in vitro and in vivo experiments. All models produced quantitatively comparable results to the baseline and were more robust to noise. ConvNeXt emerged as the sole model to achieve high-quality results on in vivo aliased samples. These results demonstrate the interest of supervised deep learning methods for Doppler velocity estimation from a reduced number of acquisitions.

PMID:38976463 | DOI:10.1109/TUFFC.2024.3424549

IEEE Trans Ultrason Ferroelectr Freq Control. 2024 Jul 8;PP. doi: 10.1109/TUFFC.2024.3424955. Online ahead of print.

ABSTRACT

Ultrasound Localization Microscopy (ULM), an emerging medical imaging technique, effectively resolves the classical trade-off between resolution and penetration inherent in traditional ultrasound imaging, opening up new avenues for noninvasive observation of the microvascular system. However, traditional microbubble tracking methods encounter various practical challenges. These methods typically entail multiple processing stages, including intricate steps like pairwise correlation and trajectory optimization, rendering real-time applications unfeasible. Furthermore, existing deep learning-based tracking techniques neglect the temporal aspects of microbubble motion, leading to ineffective modeling of their dynamic behavior. To address these limitations, this study introduces a novel approach called the Gated Recurrent Unit (GRU)-based Multitasking Temporal Neural Network (GRU-MT). GRU-MT is designed to simultaneously handle microbubble trajectory tracking and trajectory optimization tasks. Additionally, we enhance the nonlinear motion model initially proposed by Piepenbrock et al. to better encapsulate the nonlinear motion characteristics of microbubbles, thereby improving trajectory tracking accuracy. In this study, we perform a series of experiments involving network layer substitutions to systematically evaluate the performance of various temporal neural networks, including Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), GRU, Transformer, and its bidirectional counterparts, on the microbubble trajectory tracking task. Concurrently, the proposed method undergoes qualitative and quantitative comparisons with traditional microbubble tracking techniques. The experimental results demonstrate that GRU-MT exhibits superior nonlinear modeling capabilities and robustness, both in simulation and in vivo dataset. Additionally, it achieves reduced trajectory tracking errors in shorter time intervals, underscoring its potential for efficient microbubble trajectory tracking. Model code is open-sourced at https://github.com/zyt-Lib/GRU-MT.

PMID:38976462 | DOI:10.1109/TUFFC.2024.3424955

IEEE Trans Cybern. 2024 Jul 8;PP. doi: 10.1109/TCYB.2024.3410844. Online ahead of print.

ABSTRACT

Recently, deep learning-based electroencephalogram (EEG) analysis and decoding have attracted widespread attention for monitoring the clinical condition of users and identifying their intention/emotion. Nevertheless, the existing methods generally model EEG signals with limited viewpoints or restricted concerns about the characteristics of the EEG signals, and thus represent complex spectro-/spatiotemporal patterns and suffer from high variability. In this work, we propose the novel EEG-oriented self-supervised learning methods and a novel deep architecture to learn rich representation, including information about the diverse spectral characteristics of neural oscillations, the spatial properties of electrode sensor distribution, and the temporal patterns of both the global and local viewpoints. Along with the proposed self-supervision strategies and deep architectures, we devise a feature normalization strategy to resolve the intra-/inter-subject variability problem. We demonstrate the validity of our proposed deep learning framework on the four publicly available datasets by conducting comparisons with the state of the art baselines. It is also noteworthy that we exploit the same network architecture for the four different EEG paradigms and outperform the comparison methods, thereby meeting the challenge of the task-dependent network architecture engineering in EEG-based applications.

PMID:38976458 | DOI:10.1109/TCYB.2024.3410844

Vis Comput Ind Biomed Art. 2024 Jul 8;7(1):17. doi: 10.1186/s42492-024-00168-5.

ABSTRACT

Pneumonia is a serious disease that can be fatal, particularly among children and the elderly. The accuracy of pneumonia diagnosis can be improved by combining artificial-intelligence technology with X-ray imaging. This study proposes X-ODFCANet, which addresses the issues of low accuracy and excessive parameters in existing deep-learning-based pneumonia-classification methods. This network incorporates a feature coordination attention module and an omni-dimensional dynamic convolution (ODConv) module, leveraging the residual module for feature extraction from X-ray images. The feature coordination attention module utilizes two one-dimensional feature encoding processes to aggregate feature information from different spatial directions. Additionally, the ODConv module extracts and fuses feature information in four dimensions: the spatial dimension of the convolution kernel, input and output channel quantities, and convolution kernel quantity. The experimental results demonstrate that the proposed method can effectively improve the accuracy of pneumonia classification, which is 3.77% higher than that of ResNet18. The model parameters are 4.45M, which was reduced by approximately 2.5 times. The code is available at https://github.com/limuni/X-ODFCANET .

PMID:38976189 | DOI:10.1186/s42492-024-00168-5

Int J Comput Assist Radiol Surg. 2024 Jul 8. doi: 10.1007/s11548-024-03219-7. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVE: Accurate IVD segmentation is crucial for diagnosing and treating spinal conditions. Traditional deep learning methods depend on extensive, annotated datasets, which are hard to acquire. This research proposes an intensity-based self-supervised domain adaptation, using unlabeled multi-domain data to reduce reliance on large annotated datasets.

METHODS: The study introduces an innovative method using intensity-based self-supervised learning for IVD segmentation in MRI scans. This approach is particularly suited for IVD segmentations due to its ability to effectively capture the subtle intensity variations that are characteristic of spinal structures. The model, a dual-task system, simultaneously segments IVDs and predicts intensity transformations. This intensity-focused method has the advantages of being easy to train and computationally light, making it highly practical in diverse clinical settings. Trained on unlabeled data from multiple domains, the model learns domain-invariant features, adeptly handling intensity variations across different MRI devices and protocols.

RESULTS: Testing on three public datasets showed that this model outperforms baseline models trained on single-domain data. It handles domain shifts and achieves higher accuracy in IVD segmentation.

CONCLUSIONS: This study demonstrates the potential of intensity-based self-supervised domain adaptation for IVD segmentation. It suggests new directions for research in enhancing generalizability across datasets with domain shifts, which can be applied to other medical imaging fields.

PMID:38976178 | DOI:10.1007/s11548-024-03219-7

Neuroinformatics. 2024 Jul 8. doi: 10.1007/s12021-024-09675-5. Online ahead of print.

ABSTRACT

The brain is an intricate system that controls a variety of functions. It consists of a vast number of cells that exhibit diverse characteristics. To understand brain function in health and disease, it is crucial to classify neurons accurately. Recent advancements in machine learning have provided a way to classify neurons based on their electrophysiological activity. This paper presents a deep-learning framework that classifies neurons solely on this basis. The framework uses data from the Allen Cell Types database, which contains a survey of biological features derived from single-cell recordings from mice and humans. The shared information from both sources is used to classify neurons into their broad types with the help of a joint model. An accurate domain-adaptive model, integrating electrophysiological data from both mice and humans, is implemented. Furthermore, data from mouse neurons, which also includes labels of transgenic mouse lines, is further classified into subtypes using an interpretable neural network model. The framework provides state-of-the-art results in terms of accuracy and precision while also providing explanations for the predictions.

PMID:38976152 | DOI:10.1007/s12021-024-09675-5

Clin Cancer Res. 2024 Jul 8. doi: 10.1158/1078-0432.CCR-23-3842. Online ahead of print.

ABSTRACT

PURPOSE: Recent artificial intelligence (AI) algorithms aided intraoperative decision-making via stimulated Raman histology (SRH) during craniotomy. This study assesses deep-learning algorithms for rapid intraoperative diagnosis from SRH images in small stereotactic-guided brain biopsies. It defines a minimum tissue sample size threshold to ensure diagnostic accuracy.

EXPERIMENTAL DESIGN: A prospective single-center study examined 121 SRH images from 84 patients with unclear intracranial lesions undergoing stereotactic brain biopsy. Unprocessed, label-free samples were imaged with a portable fiber-laser Raman scattering microscope. Three deep-learning models were tested to (I) identify tumorous/non-tumorous tissue as qualitative biopsy control, (II) subclassify into high-grade glioma (CNS WHO grade 4), diffuse low-grade glioma (CNS WHO grade 2-3), metastases, lymphoma, or gliosis, and (III) molecularly subtype IDH- and 1p/19q-status of adult-type diffuse gliomas. Model predictions were evaluated against frozen section analysis and final neuropathological diagnoses.

RESULTS: The first model identified tumorous/non-tumorous tissue with 91.7% accuracy. Sample size on slides impacted accuracy in brain tumor subclassification (81.6%, κ=0.72 frozen section; 73.9%, κ=0.61 second model), with SRH being smaller than H&E (4.1±2.5mm² vs 16.7±8.2mm², p<0.001). SRH images with over 140 high-quality patches and a mean squeezed sample of 5.26mm² yielded 89.5% accuracy in subclassification and 93.9% in molecular subtyping of adult-type diffuse gliomas.

CONCLUSIONS: AI-based SRH image analysis is non-inferior to frozen section analysis in detecting and subclassifying brain tumors during small stereotactic-guided biopsies once a critical squeezed sample size is reached. Beyond frozen section analysis, it enables valid molecular glioma subtyping, allowing faster treatment decisions in the future. Refinement is needed for long-term application.

PMID:38976016 | DOI:10.1158/1078-0432.CCR-23-3842

Brief Bioinform. 2024 May 23;25(4):bbae321. doi: 10.1093/bib/bbae321.

ABSTRACT

The process of drug discovery is widely known to be lengthy and resource-intensive. Artificial Intelligence approaches bring hope for accelerating the identification of molecules with the necessary properties for drug development. Drug-likeness assessment is crucial for the virtual screening of candidate drugs. However, traditional methods like Quantitative Estimation of Drug-likeness (QED) struggle to distinguish between drug and non-drug molecules accurately. Additionally, some deep learning-based binary classification models heavily rely on selecting training negative sets. To address these challenges, we introduce a novel unsupervised learning framework called DrugMetric, an innovative framework for quantitatively assessing drug-likeness based on the chemical space distance. DrugMetric blends the powerful learning ability of variational autoencoders with the discriminative ability of the Gaussian Mixture Model. This synergy enables DrugMetric to identify significant differences in drug-likeness across different datasets effectively. Moreover, DrugMetric incorporates principles of ensemble learning to enhance its predictive capabilities. Upon testing over a variety of tasks and datasets, DrugMetric consistently showcases superior scoring and classification performance. It excels in quantifying drug-likeness and accurately distinguishing candidate drugs from non-drugs, surpassing traditional methods including QED. This work highlights DrugMetric as a practical tool for drug-likeness scoring, facilitating the acceleration of virtual drug screening, and has potential applications in other biochemical fields.

PMID:38975893 | DOI:10.1093/bib/bbae321

Network. 2024 Jul 8:1-39. doi: 10.1080/0954898X.2024.2373127. Online ahead of print.

ABSTRACT

Early detection of lung cancer is necessary to prevent deaths caused by lung cancer. But, the identification of cancer in lungs using Computed Tomography (CT) scan based on some deep learning algorithms does not provide accurate results. A novel adaptive deep learning is developed with heuristic improvement. The proposed framework constitutes three sections as (a) Image acquisition, (b) Segmentation of Lung nodule, and (c) Classifying lung cancer. The raw CT images are congregated through standard data sources. It is then followed by nodule segmentation process, which is conducted by Adaptive Multi-Scale Dilated Trans-Unet3+. For increasing the segmentation accuracy, the parameters in this model is optimized by proposing Modified Transfer Operator-based Archimedes Optimization (MTO-AO). At the end, the segmented images are subjected to classification procedure, namely, Advanced Dilated Ensemble Convolutional Neural Networks (ADECNN), in which it is constructed with Inception, ResNet and MobileNet, where the hyper parameters is tuned by MTO-AO. From the three networks, the final result is estimated by high ranking-based classification. Hence, the performance is investigated using multiple measures and compared among different approaches. Thus, the findings of model demonstrate to prove the system's efficiency of detecting cancer and help the patient to get the appropriate treatment.

PMID:38975771 | DOI:10.1080/0954898X.2024.2373127

Spine (Phila Pa 1976). 2024 Jul 8. doi: 10.1097/BRS.0000000000005093. Online ahead of print.

ABSTRACT

STUDY DESIGN: Diagnostics.

OBJECTIVES: Based on deep learning semantic segmentation model, we sought to assess pelvic tilt by area ratio of the lesser pelvic and the obturator foramen in anteroposterior (AP) radiographs.

BACKGROUND: Pelvic tilt is a critical factor in hip and spinal surgery, commonly evaluated by medical professionals through sagittal pelvic radiographs. The inherent pelvic asymmetry, as well as potential obstructions from clothing and musculature in roentgenography, may result in ghosting and blurring artifacts, thereby complicating precise measurement.

METHODS: PT directly affects the area ratio of the lesser pelvis to the obturator foramen in AP radiographs. An exponential regression analysis of simulated radiographs from ten male and ten female pelvises in specific tilt positions derived a formula correlating this area ratio with PT. Two blinded investigators evaluated this formula using 161 simulated AP pelvic radiographs. A deep learning semantic segmentation model was then fine-tuned to automatically calculate the area ratio, enabling intelligent PT evaluation. This model and the regression function were integrated for automated PT measurement and tested on a dataset of 231 clinical cases.

RESULTS: We observed no disparity between males and females in the aforementioned area ratio. The test results from two blinded investigators analyzing 161 simulated radiographs revealed a mean absolute error of 0.19° (SD±4.71°), with a correlation coefficient between them reaching 0.96. Additionally, the mean absolute error obtained from testing 231 clinical AP radiographs using the fine-tuned semantic segmentation model mentioned earlier is -0.58° (SD±5.97°).

CONCLUSION: We found that using deep learning neural networks enabled a more accurate and robust automatic measurement of PT through the area ratio of the lesser pelvis and obturator foramen.

PMID:38975768 | DOI:10.1097/BRS.0000000000005093

Nano Lett. 2024 Jul 8. doi: 10.1021/acs.nanolett.4c02860. Online ahead of print.

ABSTRACT

The detection of hepatitis B surface antigen (HBsAg) is critical in diagnosing hepatitis B virus (HBV) infection. However, existing clinical detection technologies inevitably cause certain inaccuracies, leading to delayed or unwarranted treatment. Here, we introduce a label-free plasmonic biosensing method based on the thickness-sensitive plasmonic coupling, combined with supervised deep learning (DL) using neural networks. The strategy of utilizing neural networks to process output data can reduce the limit of detection (LOD) of the sensor and significantly improve the accuracy (from 93.1%-97.4% to 99%-99.6%). Compared with widely used emerging clinical technologies, our platform achieves accurate decisions with higher sensitivity in a short assay time (∼30 min). The integration of DL models considerably simplifies the readout procedure, resulting in a substantial decrease in processing time. Our findings offer a promising avenue for developing high-precision molecular detection tools for point-of-care (POC) applications.

PMID:38975746 | DOI:10.1021/acs.nanolett.4c02860

Spine (Phila Pa 1976). 2024 Jul 8. doi: 10.1097/BRS.0000000000005088. Online ahead of print.

ABSTRACT

STUDY DESIGN: A retrospective analysis.

OBJECTIVE: This research sought to develop a predictive model for surgical outcomes in patients with cervical ossification of the posterior longitudinal ligament (OPLL) using deep learning and machine learning (ML) techniques.

SUMMARY OF BACKGROUND DATA: Determining surgical outcomes assists surgeons in communicating prognosis to patients and setting their expectations. Deep learning and ML are computational models that identify patterns from large datasets and make predictions.

METHODS: Of the 482 patients, 288 patients were included in the analysis. A minimal clinically important difference (MCID) was defined as gain in Japanese Orthopaedic Association (JOA) score of 2.5 points or more. The predictive model for MCID achievement at 1 year post-surgery was constructed using patient background, clinical symptoms, and preoperative imaging features (x-ray, CT, MRI) analyzed via LightGBM and deep learning with RadImagenet.

RESULTS: The median preoperative JOA score was 11.0 (IQR: 9.0-12.0), which significantly improved to 14.0 (IQR: 12.0-15.0) at 1 year after surgery (P < 0.001, Wilcoxon signed-rank test). The average improvement rate of the JOA score was 44.7%, and 60.1% of patients achieved the MCID. Our model exhibited an area under the receiver operating characteristic curve of 0.81 and the accuracy of 71.9% in predicting MCID at 1 year. Preoperative JOA score and certain preoperative imaging features were identified as the most significant factors in the predictive models.

CONCLUSION: A predictive ML and deep learning model for surgical outcomes in OPLL patients is feasible, suggesting promising applications in spinal surgery.

LEVEL OF EVIDENCE: 4.

PMID:38975742 | DOI:10.1097/BRS.0000000000005088

Chronobiol Int. 2024 Jul 8:1-12. doi: 10.1080/07420528.2024.2371143. Online ahead of print.

ABSTRACT

Most organisms synchronize to an approximately 24-hour (circadian) rhythm. This study introduces a novel deep learning-powered video tracking method to assess the stability, fragmentation, robustness and synchronization of activity rhythms in Xyrichtys novacula. Experimental X. novacula were distributed into three groups and monitored for synchronization to a 14/10 hours of light/dark to assess acclimation to laboratory conditions. Group GP7 acclimated for 1 week and was tested from days 7 to 14, GP14 acclimated for 14 days and was tested from days 14 to 21 and GP21 acclimated for 21 days and was tested from days 21 to 28. Telemetry data from individuals in the wild depicted their natural behavior. Wild fish displayed a robust and minimally fragmented rhythm, entrained to the natural photoperiod. Under laboratory conditions, differences in activity levels were observed between light and dark phases. However, no differences were observed in activity rhythm metrics among laboratory groups related to acclimation period. Notably, longer acclimation (GP14 and GP21) led to a larger proportion of individuals displaying rhythm synchronization with the imposed photoperiod. Our work introduces a novel approach for monitoring biological rhythms in laboratory conditions, employing a specifically engineered video tracking system based on deep learning, adaptable for other species.

PMID:38975732 | DOI:10.1080/07420528.2024.2371143

Lin Chuang Er Bi Yan Hou Tou Jing Wai Ke Za Zhi. 2024 Jul;38(7):624-630. doi: 10.13201/j.issn.2096-7993.2024.07.013.

ABSTRACT

Objective:To build a VGG-based computer-aided diagnostic model for chronic sinusitis and evaluate its efficacy. Methods:①A total of 5 000 frames of diagnosed sinus CT images were collected. The normal group consisted of 1 000 frames（250 frames each of maxillary sinus, frontal sinus, septal sinus, and pterygoid sinus）, while the abnormal group consisted of 4 000 frames（1 000 frames each of maxillary sinusitis, frontal sinusitis, septal sinusitis, and pterygoid sinusitis）. ②The models were trained and simulated to obtain five classification models for the normal group, the pteroid sinusitis group, the frontal sinusitis group, the septal sinusitis group and the maxillary sinusitis group, respectively. The classification efficacy of the models was evaluated objectively in six dimensions: accuracy, precision, sensitivity, specificity, interpretation time and area under the ROC curve（AUC）. ③Two hundred randomly selected images were read by the model with three groups of physicians（low, middle and high seniority） to constitute a comparative experiment. The efficacy of the model was objectively evaluated using the aforementioned evaluation indexes in conjunction with clinical analysis. Results:①Simulation experiment: The overall recognition accuracy of the model is 83.94%, with a precision of 89.52%, sensitivity of 83.94%, specificity of 95.99%, and the average interpretation time of each frame is 0.2 s. The AUC for sphenoid sinusitis was 0.865（95%CI 0.849-0.881）, for frontal sinusitis was 0.924（0.991-0.936）, for ethmoidoid sinusitis was 0.895（0.880-0.909）, and for maxillary sinusitis was 0.974（0.967-0.982）. ②Comparison experiment: In terms of recognition accuracy, the model was 84.52%, while the low-seniority physicians group was 78.50%, the middle-seniority physicians group was 80.50%, and the seniority physicians group was 83.50%; In terms of recognition accuracy, the model was 85.67%, the low seniority physicians group was 79.72%, the middle seniority physicians group was 82.67%, and the high seniority physicians group was 83.66%. In terms of recognition sensitivity, the model was 84.52%, the low seniority group was 78.50%, the middle seniority group was 80.50%, and the high seniority group was 83.50%. In terms of recognition specificity, the model was 96.58%, the low-seniority physicians group was 94.63%, the middle-seniority physicians group was 95.13%, and the seniority physicians group was 95.88%. In terms of time consumption, the average image per frame of the model is 0.20 s, the average image per frame of the low-seniority physicians group is 2.35 s, the average image per frame of the middle-seniority physicians group is 1.98 s, and the average image per frame of the senior physicians group is 2.19 s. Conclusion:This study demonstrates the potential of a deep learning-based artificial intelligence diagnostic model for chronic sinusitis to classify and diagnose chronic sinusitis; the deep learning-based artificial intelligence diagnosis model for chronic sinusitis has good classification performance and high diagnostic efficacy.

PMID:38973043 | DOI:10.13201/j.issn.2096-7993.2024.07.013

Ultrasound Med Biol. 2024 Jul 7:S0301-5629(24)00233-3. doi: 10.1016/j.ultrasmedbio.2024.05.027. Online ahead of print.

ABSTRACT

OBJECTIVE: Bone diseases deteriorate the microstructure of bone tissue. Optical-resolution photoacoustic microscopy (OR-PAM) enables high spatial resolution of imaging bone tissues. However, the spatiotemporal trade-off limits the application of OR-PAM. The purpose of this study was to improve the quality of OR-PAM images without sacrificing temporal resolution.

METHODS: In this study, we proposed the Photoacoustic Dense Attention U-Net (PADA U-Net) model, which was used for reconstructing full-scanning images from under-sampled images. Thereby, this approach breaks the trade-off between imaging speed and spatial resolution.

RESULTS: The proposed method was validated on resolution test targets and bovine cancellous bone samples to demonstrate the capability of PADA U-Net in recovering full-scanning images from under-sampled OR-PAM images. With a down-sampling ratio of [4, 1], compared to bilinear interpolation, the Peak Signal-to-Noise Ratio and Structural Similarity Index Measure values (averaged over the test set of bovine cancellous bone) of the PADA U-Net were improved by 2.325 dB and 0.117, respectively.

CONCLUSION: The results demonstrate that the PADA U-Net model reconstructed the OR-PAM images well with different levels of sparsity. Our proposed method can further facilitate early diagnosis and treatment of bone diseases using OR-PAM.

PMID:38972792 | DOI:10.1016/j.ultrasmedbio.2024.05.027