Assessment of multi-modal magnetic resonance imaging for glioma based on a deep learning reconstruction approach with the denoising method.

BACKGROUND: Deep learning reconstruction (DLR) with denoising has been reported as potentially improving the image quality of magnetic resonance imaging (MRI). Multi-modal MRI is a critical non-invasive method for tumor detection, surgery planning, and prognosis assessment; however, the DLR on multi-modal glioma imaging has not been assessed. PURPOSE: To assess multi-modal MRI for glioma based on the DLR method. MATERIAL AND METHODS: We assessed multi-modal images of 107 glioma patients (49 preoperative and 58 postoperative). All the images were reconstructed with both DLR and conventional reconstruction methods, encompassing T1-weighted (T1W), contrast-enhanced T1W (CE-T1), T2-weighted (T2W), and T2 fluid-attenuated inversion recovery (T2-FLAIR). The image quality was evaluated using signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and edge sharpness. Visual assessment and diagnostic assessment were performed blindly by neuroradiologists. RESULTS: In contrast with conventionally reconstructed images, (residual) tumor SNR for all modalities and tumor to white/gray matter CNR from DLR images were higher in T1W, T2W, and T2-FLAIR sequences. The visual assessment of DLR images demonstrated the superior visualization of tumor in T2W, edema in T2-FLAIR, enhanced tumor and necrosis part in CE-T1, and fewer artifacts in all modalities. Improved diagnostic efficiency and confidence were observed for preoperative cases with DLR images. CONCLUSION: DLR of multi-modal MRI reconstruction prototype for glioma has demonstrated significant improvements in image quality. Moreover, it increased diagnostic efficiency and confidence of glioma.

Dominating Alzheimer's disease diagnosis with deep learning on sMRI and DTI-MD.

Background: Alzheimer's disease (AD) is a progressive and irreversible neurodegenerative disorder that has become one of the major health concerns for the elderly. Computer-aided AD diagnosis can assist doctors in quickly and accurately determining patients' severity and affected regions. Methods: In this paper, we propose a method called MADNet for computer-aided AD diagnosis using multimodal datasets. The method selects ResNet-10 as the backbone network, with dual-branch parallel extraction of discriminative features for AD classification. It incorporates long-range dependencies modeling using attention scores in the decision-making layer and fuses the features based on their importance across modalities. To validate the effectiveness of our proposed multimodal classification method, we construct a multimodal dataset based on the publicly available ADNI dataset and a collected XWNI dataset, which includes examples of AD, Mild Cognitive Impairment (MCI), and Cognitively Normal (CN). Results: On this dataset, we conduct binary classification experiments of AD vs. CN and MCI vs. CN, and demonstrate that our proposed method outperforms other traditional single-modal deep learning models. Furthermore, this conclusion also confirms the necessity of using multimodal sMRI and DTI data for computer-aided AD diagnosis, as these two modalities complement and convey information to each other. We visualize the feature maps extracted by MADNet using Grad-CAM, generating heatmaps that guide doctors' attention to important regions in patients' sMRI, which play a crucial role in the development of AD, establishing trust between human experts and machine learning models. Conclusion: We propose a simple yet effective multimodal deep convolutional neural network model MADNet that outperforms traditional deep learning methods that use a single-modality dataset for AD diagnosis.

Wide Dynamic Range Digital Aberration Measurement and Fast Anterior-Segment OCT Imaging †.

Ocular aberrometry with a wide dynamic range for assessing vision performance and anterior segment imaging that provides anatomical details of the eye are both essential for vision research and clinical applications. Defocus error is a major limitation of digital wavefront aberrometry (DWA), as the blurring of the detected point spread function (PSF) significantly reduces the signal-to-noise ratio (SNR) beyond the ±3 D range. With the aid of Badal-like precompensation of defocus, the dynamic defocus range of the captured aberrated PSFs can be effectively extended. We demonstrate a dual-modality MHz VCSEL-based swept-source OCT (SS-OCT) system with easy switching between DWA and OCT imaging modes. The system is capable of measuring aberrations with defocus dynamic range of 20 D as well as providing fast anatomical imaging of the anterior segment at an A-scan rate of 1.6 MHz.

Interaction of spatial attention and the associated reward value of audiovisual objects.

Reward value and selective attention both enhance the representation of sensory stimuli at the earliest stages of processing. It is still debated whether and how reward-driven and attentional mechanisms interact to influence perception. Here we ask whether the interaction between reward value and selective attention depends on the sensory modality through which the reward information is conveyed. Human participants first learned the reward value of uni-modal visual and auditory stimuli during a conditioning phase. Subsequently, they performed a target detection task on bimodal stimuli containing a previously rewarded stimulus in one, both, or neither of the modalities. Additionally, participants were required to focus their attention on one side and only report targets on the attended side. Our results showed a strong modulation of visual and auditory event-related potentials (ERPs) by spatial attention. We found no main effect of reward value but importantly we found an interaction effect as the strength of attentional modulation of the ERPs was significantly affected by the reward value. When reward effects were examined separately with respect to each modality, auditory value-driven modulation of attention was found to dominate the ERP effects whereas visual reward value on its own led to no effect, likely due to its interference with the target processing. These results inspire a two-stage model where first the salience of a high reward stimulus is enhanced on a local priority map specific to each sensory modality, and at a second stage reward value and top-down attentional mechanisms are integrated across sensory modalities to affect perception.

Evolving immunotherapeutic solutions for triple-negative breast carcinoma.

Triple-negative breast carcinoma (TNBC) remains a formidable clinical hurdle owing to its high aggressiveness and scant therapeutic options. Nonetheless, the evolving landscape of immunotherapeutic strategies opens up promising avenues for tackling this hurdle. This review discusses the advancing immunotherapy for TNBC, accentuating personalized interventions due to tumor microenvironment (TME) diversity. Immune checkpoint inhibitors (ICIs) hold pivotal significance, both as single-agent therapies and when administered alongside cytotoxic agents. Moreover, the concurrent inhibition of multiple immune checkpoints represents a potent approach to augment the efficacy of cancer immunotherapy. Synergistic effects have been observed when ICIs are combined with targeted treatments like PARP inhibitors, anti-angiogenics, and ADCs (antibody-drug conjugates). Emerging tactics include tumor vaccines, cellular immunotherapy, and oncolytic viruses, leveraging the immune system's ability for selective malignant cell destruction. This review offers an in-depth examination of the diverse landscape of immunotherapy development for TNBC, furnishing meticulous insights into various advancements within this field. In addition, immunotherapeutic interventions offer hope for TNBC, needing further research for optimization.

The Lung Cancer Surgical Study Group of the Japan Clinical Oncology Group: outstanding contribution and entering a new phase.

The Lung Cancer Surgical Study Group (LCSSG) of the Japan Clinical Oncology Group (JCOG) was organized in 1986 and initially included 26 collaborative institutions, which has increased to 52 institutions currently. JCOG-LCSSG includes thoracic surgeons, medical oncologists, pathologists, and radiotherapists. In the early period, the JCOG-LCSSG mainly focused on combined modality therapies for lung cancer. Since the 2000s, the JCOG-LCSSG has investigated adequate modes of surgical resection for small-sized and peripheral non-small cell lung cancer and based on the radiological findings of whole tumor size and ground-glass opacity. Trials, such as JCOG0802, JCOG0804, and JCOG1211, have shown the appropriateness of sublobar resection, which has significantly influenced routine clinical practice. With the introduction of targeted therapy and immunotherapy, treatment strategies for lung cancer have changed significantly. Additionally, with the increasing aging population and medical costs, tailored medicine is strongly recommended to address medical issues. To ensure comprehensive treatment, strategies, including surgical and nonsurgical approaches, should be developed. Currently, the JCOG-LCSSG has conducted numerous clinical trials to adjust the diversity of lung cancer treatment strategies. This review highlights recent advancements in the surgical field, current status, and future direction of the JCOG-LCSSG.

Bodily-tactile early intervention: a pilot study of the role of maternal touch and emotional availability in interactions between three children with visual impairment and additional disabilities and their mothers.

Background: Children with visual impairment and additional disabilities (VIAD) have difficulty accessing the visual information related to their parents' facial expressions and gestures. Similarly, it may be hard for parents to detect their children's subtle expressions. These challenges in accessibility may compromise emotional availability (EA) in parent-child interactions. The systematic use of the bodily-tactile modality for expressive and receptive communicative functions may function as a strategy to compensate for a child's lack of vision. This multiple-case study explored the effects of a bodily-tactile early intervention for three mothers and their one-year-old children with VIAD. Methods: Video data from baseline, intervention, and follow-up sessions were analyzed using a bodily-tactile coding procedure and EA Scales. Results: During the intervention, all mothers began to use a more bodily-tactile modality in early play routines and in different communicative functions. They increased their use of anticipatory cues, noticing responses, and tactile signs. Moreover, the children were more emotionally available to their mothers during the intervention and follow-up compared to the baseline. Conclusion: The results indicated that, during a short intervention, mothers could adopt a systematic use of the bodily-tactile modality in interactions with their children with VIAD. The results also suggest that, when mothers increased flexibility in communication channels, it was positively linked to their children's EA.

Robust Unsupervised Super-Resolution of Infant MRI via Dual-Modal Deep Image Prior.

Magnetic resonance imaging (MRI) is commonly used for studying infant brain development. However, due to the lengthy image acquisition time and limited subject compliance, high-quality infant MRI can be challenging. Without imposing additional burden on image acquisition, image super-resolution (SR) can be used to enhance image quality post-acquisition. Most SR techniques are supervised and trained on multiple aligned low-resolution (LR) and high-resolution (HR) image pairs, which in practice are not usually available. Unlike supervised approaches, Deep Image Prior (DIP) can be employed for unsupervised single-image SR, utilizing solely the input LR image for de novo optimization to produce an HR image. However, determining when to stop early in DIP training is non-trivial and presents a challenge to fully automating the SR process. To address this issue, we constrain the low-frequency k-space of the SR image to be similar to that of the LR image. We further improve performance by designing a dual-modal framework that leverages shared anatomical information between T1-weighted and T2-weighted images. We evaluated our model, dual-modal DIP (dmDIP), on infant MRI data acquired from birth to one year of age, demonstrating that enhanced image quality can be obtained with substantially reduced sensitivity to early stopping.

An IL-7 fusion protein targeting EDA fibronectin upregulates TCF1 on CD8+ T-cells, preferentially accumulates to neoplastic lesions, and boosts PD-1 blockade.

BACKGROUND: Anti-PD-1 antibodies have revolutionized cancer immunotherapy due to their ability to induce long-lasting complete remissions in a proportion of patients. Current research efforts are attempting to identify biomarkers and suitable combination partners to predict or further improve the activity of immune checkpoint inhibitors. Antibody-cytokine fusions are a class of pharmaceuticals that showed the potential to boost the anticancer properties of other immunotherapies. Extradomain A-fibronectin (EDA-FN), which is expressed in most solid and hematological tumors but is virtually undetectable in healthy adult tissues, is an attractive target for the delivery of cytokine at the site of the disease. METHODS: In this work, we describe the generation and characterization of a novel interleukin-7-based fusion protein targeting EDA-FN termed F8(scDb)-IL7. The product consists of the F8 antibody specific to the alternatively spliced EDA of FN in the single-chain diabody (scDb) format fused to human IL-7. RESULTS: F8(scDb)-IL7 efficiently stimulates human peripheral blood mononuclear cells in vitro. Moreover, the product significantly increases the expression of T Cell Factor 1 (TCF-1) on CD8+T cells compared with an IL2-fusion protein. TCF-1 has emerged as a pivotal transcription factor that influences the durability and potency of immune responses against tumors. In preclinical cancer models, F8(scDb)-IL7 demonstrates potent single-agent activity and eradicates sarcoma lesions when combined with anti-PD-1. CONCLUSIONS: Our results provide the rationale to explore the combination of F8(scDb)-IL7 with anti-PD-1 antibodies for the treatment of patients with cancer.

Caregiver Report of Adverse Childhood Events: Comparison of Self-Administered and Telephone Questionnaires.

Adverse childhood experiences (ACEs) are traumatic experiences that increase people's susceptibility to adverse physical health, mental health, and social consequences in adulthood. Screening for ACEs in primary care settings is complicated by a lack of consensus on appropriate methods for identifying exposure to ACEs. It is unclear whether self-report methods could increase disclosure of ACEs as compared to interview-based methods. This study compares data on the prevalence of ACEs from two publicly available surveys conducted on the same population of children's caregivers: the 2019 Ohio subsample of the web/mail-based National Survey of Children's Health and the telephone-based 2019 Ohio Medicaid Assessment Survey. We find higher disclosure of caregiver-reported child exposure to ACEs in the telephone interview survey, highlighting the importance of the role of verbal communication in developing a safe and trusting relationship in the disclosure of trauma.

Improved diagnostic yield of peripheral pulmonary malignant lesions with emphysema using a combination of radial endobronchial ultrasonography and rapid on-site evaluation.

BACKGROUND: This is a retrospective cohort study from a single center of Chest Medical District of Nanjing Brain Hospital Affiliated to Nanjing Medical University, Jiangsu Province, China. It was aim to evaluate the diagnostic value of radial endobronchial ultrasound (R-EBUS) combination with rapid on-site evaluation (ROSE) guided transbronchial lung biopsy (TBLB) for peripheral pulmonary lesions in patients with emphysema. METHODS: All 170 patients who underwent PPLs with emphysema received an R-EBUS examination with or without the ROSE procedure, and the diagnostic yield, safety, and possible factors influencing diagnosis were analyzed between the two groups by the SPSS 25.0 software. RESULTS: The pooled and benign diagnostic yields were not different in the two groups (P = 0.224, 0.924), but the diagnostic yield of malignant PPLs was significantly higher in the group with ROSE than the group without ROSE (P = 0.042). The sensitivity of ROSE was 79.10%, the specificity, 91.67%, the positive predictive value, 98.15%, and the negative predictive value, 84.62%. The diagnostic accuracy, was 95.52%. In the group of R-EBUS + ROSE, the procedural time and the number of times of biopsy or brushing were both significantly reduced (all P<0.05). The incidence of pneumothorax (1.20%) and bleeding (10.84%) in the group of R-EBUS + ROSE were also less than those in the group of R-EBUS (P<0.05). The lesion's diameter ≥ 2 cm, the distance between the pleura and the lesion ≥ 2 cm, the positive air bronchograms sign, the location of the ultrasound probe within the lesion, and the even echo with clear margin feature of lesion ultrasonic image, these factors are possibly relevant to a higher diagnostic yield. The diagnostic yield of PPLs those were adjacent to emphysema were lower than those PPLs which were away from emphysema (P = 0.048) in the group without ROSE, however, in the group of R-EBUS + ROSE, there was no such difference whether the lesion is adjacent to emphysema or not (P = 0.236). CONCLUSION: Our study found that the combination of R-EBUS and ROSE during bronchoscopy procedure was a safe and effective modality to improve diagnostic yield of PPLs with emphysema, especially for malignant PPLs. The distance between the pleura and the lesion ≥ 2 cm, the positive air bronchograms sign, the location of the ultrasound probe within the lesion, and the even echo with clear margin feature of lesion ultrasonic image, these factors possibly indicated a higher diagnostic yield. Those lesions' position is adjacent to emphysema may reduce diagnostic yield but ROSE may make up for this deficiency.

An immunofluorescence-guided segmentation model in H&E images is enabled by tissue artifact correction by CycleGAN.

Despite recent advances, the adoption of computer vision methods into clinical and commercial applications has been hampered by the limited availability of accurate ground truth tissue annotations required to train robust supervised models. Generating such ground truth can be accelerated by annotating tissue molecularly using immunofluorescence staining (IF) and mapping these annotations to a post-IF H&E (terminal H&E). Mapping the annotations between the IF and the terminal H&E increases both the scale and accuracy by which ground truth could be generated. However, discrepancies between terminal H&E and conventional H&E caused by IF tissue processing have limited this implementation. We sought to overcome this challenge and achieve compatibility between these parallel modalities using synthetic image generation, in which a cycle-consistent generative adversarial network (CycleGAN) was applied to transfer the appearance of conventional H&E such that it emulates the terminal H&E. These synthetic emulations allowed us to train a deep learning (DL) model for the segmentation of epithelium in the terminal H&E that could be validated against the IF staining of epithelial-based cytokeratins. The combination of this segmentation model with the CycleGAN stain transfer model enabled performative epithelium segmentation in conventional H&E images. The approach demonstrates that the training of accurate segmentation models for the breadth of conventional H&E data can be executed free of human-expert annotations by leveraging molecular annotation strategies such as IF, so long as the tissue impacts of the molecular annotation protocol are captured by generative models that can be deployed prior to the segmentation process.

Enhancing prompt perception in dementia: a comparative study of mixed reality cue modalities.

Introduction: Dementia impacts millions worldwide and is challenging individuals' ability to engage in daily activities. Active living is crucial in mitigating dementia's neurodegenerative effects, yet people with dementia often struggle to initiate and complete tasks independently. Technologies offer promising solutions to engage people with dementia in activities of active living and improving their quality of life through prompting and cueing. It is anticipated that developments in sensor and wearable technologies will result in mixed reality technology becoming more accessible in everyday homes, making them more deployable. The possibility of mixed reality technologies to be programmed for different applications, and to adapt them to different levels of impairments, behaviours and contexts, will make them more scalable. Objective: The study aimed to develop a better understanding of modalities of prompts that people with dementia perceive successfully and correctly in mixed reality environments. It investigated interactions of people with dementia with different types of visual (graphics, animation, etc.) and sound (human voice, tones, etc.) prompts in mixed reality technologies. Methods: We used the Research through Design (RtD) method in this study. This paper describes the findings from the user research carried out in the study. We conducted observation studies with twenty-two people with dementia playing games on off-the-shelf mixed reality technologies, including both Augmented Reality (HoloLens, ArKit on iPhone) and Augmented Virtuality (Xbox Kinect and Osmo) technologies. The interactions with the technologies during the gameplay were video recorded for thematic analysis in Noldus Observer XT (version 16.0) for successful and correct perception of prompts. Results: A comparison of the probability estimates of correct perception of the prompts by people with dementia suggests that human voice, graphic symbols and text are the most prominently perceived modalities of prompts. Feedback prompts for every action performed by people with dementia on the technology are critical for successful perception and should always be provided in the design. Conclusion: The study has resulted in recommendations and guidelines for designers to design prompts for people with dementia in mixed-reality environments. The work lays the foundation for considering mixed reality technologies as assistive tools for people with dementia, fostering discussions on their accessibility and inclusive design in technology development.

Impact of sensory modality and tempo in motor timing.

Background: Accurate motor timing requires the coordinated control of actions in response to external stimuli. Over the past few years, several studies have investigated the effect of sensory input on motor timing; however, the evidence remains conflicting. The purpose of this study was to examine the impact of sensory modality and tempo on the accuracy of timed movements and explore strategies for enhancing motor timing. Methods: Participants (n = 30) performed synchronization and adaptation circle drawing tasks in virtual reality. In Experiment 1, participants synchronized circle drawing with repeated stimuli based on sensory modalities (auditory, visual, tactile, audio-visual, audio-tactile, and visual-tactile) and tempos (20, 30, and 60 bpm). In Experiment 2, we examined timing adaptation in circle drawing tasks under conditions of unexpected tempo changes, whether increased or decreased. Results: A significant interaction effect between modality and tempo was observed in the comparison of timing accuracy. Tactile stimuli exhibited significantly higher timing accuracy at 60 bpm, whereas auditory stimuli demonstrated a peak accuracy at 30 bpm. The analysis revealed a significantly larger timing error when adapting to changes in the tempo-down condition compared with the tempo-up condition. Discussion: Through Experiment 1, we found that sensory modality impacts motor timing differently depending on the tempo, with tactile modality being effective at a faster tempo and auditory modality being beneficial at a moderate tempo. Additionally, Experiment 2 revealed that adapting to changes by correcting timing errors is more challenging with decreasing tempo than with increasing tempo. Our findings suggest that motor timing is intricately influenced by sensory modality and tempo variation. Therefore, to enhance the motor timing, a comprehensive understanding of these factors and their applications is imperative.

Effect of Takeover Request Time and Warning Modality on Trust in L3 Automated Driving.

OBJECTIVE: This study investigated the effects of four takeover request (TOR) times and seven warning modalities on performance and trust in automated driving on a mildly congested urban road scenario, as well as the relationship between takeover performance and trust. BACKGROUND: Takeover is crucial in L3 automated driving, where human-machine codriving is employed. Establishing trust in takeover scenarios among drivers can enhance the acceptance of autonomous vehicles, thereby promoting their widespread adoption. METHOD: Using a driving simulator, data from 28 participants, including collision counts, takeover time (ToT), electrodermal activity (EDA) data, and self-reported trust scores, were collected and analyzed primarily using Generalized Linear Mixed Models (GLMM). RESULTS: Collisions during the takeover undermined participants' trust in the autonomous driving system. As TOR time increased, participants' trust improved, and the longer TOR time did not lead to participant confusion. There was no significant relationship between warning modality and trust. Furthermore, the combination of three warning modalities did not exhibit a notable advantage over the combination of two modalities. CONCLUSION: The study examined the effects of TOR time and warning modality on trust, as well as preliminarily explored the potential association between takeover performance, including collisions and ToT, and trust in autonomous driving takeovers. APPLICATION: Researchers and designers of automotive interactions were given referenceable TOR time and warning modality by this study, which extended the autonomous driving takeover scenarios. These findings contributed to boosting drivers' confidence in transferring control to the automated system.

Cross-Modality Person Re-Identification Method with Joint-Modality Generation and Feature Enhancement.

In order to minimize the disparity between visible and infrared modalities and enhance pedestrian feature representation, a cross-modality person re-identification method is proposed, which integrates modality generation and feature enhancement. Specifically, a lightweight network is used for dimension reduction and augmentation of visible images, and intermediate modalities are generated to bridge the gap between visible images and infrared images. The Convolutional Block Attention Module is embedded into the ResNet50 backbone network to selectively emphasize key features sequentially from both channel and spatial dimensions. Additionally, the Gradient Centralization algorithm is introduced into the Stochastic Gradient Descent optimizer to accelerate convergence speed and improve generalization capability of the network model. Experimental results on SYSU-MM01 and RegDB datasets demonstrate that our improved network model achieves significant performance gains, with an increase in Rank-1 accuracy of 7.12% and 6.34%, as well as an improvement in mAP of 4.00% and 6.05%, respectively.

Cross-Modality Medical Image Segmentation via Enhanced Feature Alignment and Cross Pseudo Supervision Learning.

Given the diversity of medical images, traditional image segmentation models face the issue of domain shift. Unsupervised domain adaptation (UDA) methods have emerged as a pivotal strategy for cross modality analysis. These methods typically utilize generative adversarial networks (GANs) for both image-level and feature-level domain adaptation through the transformation and reconstruction of images, assuming the features between domains are well-aligned. However, this assumption falters with significant gaps between different medical image modalities, such as MRI and CT. These gaps hinder the effective training of segmentation networks with cross-modality images and can lead to misleading training guidance and instability. To address these challenges, this paper introduces a novel approach comprising a cross-modality feature alignment sub-network and a cross pseudo supervised dual-stream segmentation sub-network. These components work together to bridge domain discrepancies more effectively and ensure a stable training environment. The feature alignment sub-network is designed for the bidirectional alignment of features between the source and target domains, incorporating a self-attention module to aid in learning structurally consistent and relevant information. The segmentation sub-network leverages an enhanced cross-pseudo-supervised loss to harmonize the output of the two segmentation networks, assessing pseudo-distances between domains to improve the pseudo-label quality and thus enhancing the overall learning efficiency of the framework. This method's success is demonstrated by notable advancements in segmentation precision across target domains for abdomen and brain tasks.

Evidence modeling for reliability learning and interpretable decision-making under multi-modality medical image segmentation.

Reliability learning and interpretable decision-making are crucial for multi-modality medical image segmentation. Although many works have attempted multi-modality medical image segmentation, they rarely explore how much reliability is provided by each modality for segmentation. Moreover, the existing approach of decision-making such as the softmax function lacks the interpretability for multi-modality fusion. In this study, we proposed a novel approach named contextual discounted evidential network (CDE-Net) for reliability learning and interpretable decision-making under multi-modality medical image segmentation. Specifically, the CDE-Net first models the semantic evidence by uncertainty measurement using the proposed evidential decision-making module. Then, it leverages the contextual discounted fusion layer to learn the reliability provided by each modality. Finally, a multi-level loss function is deployed for the optimization of evidence modeling and reliability learning. Moreover, this study elaborates on the framework interpretability by discussing the consistency between pixel attribution maps and the learned reliability coefficients. Extensive experiments are conducted on both multi-modality brain and liver datasets. The CDE-Net gains high performance with an average Dice score of 0.914 for brain tumor segmentation and 0.913 for liver tumor segmentation, which proves CDE-Net has great potential to facilitate the interpretation of artificial intelligence-based multi-modality medical image fusion.

The Double-Edged Sword of Online Learning for Ethnoracial Differences in Adolescent Mental Health During Late Period of the COVID-19 Pandemic in the United States: National Survey.

BACKGROUND: Despite several theories suggesting online learning during the COVID-19 pandemic would aggravate ethnoracial disparities in mental health among adolescents, extant findings suggest no ethnoracial differences in mental health or that those from minoritized ethnoracial groups reported better mental health than their White counterparts. OBJECTIVE: This study aimed to identify why findings from prior studies appear to not support that ethnoracial disparities in mental health were aggravated by testing 2 pathways. In pathway 1 pathway, online learning was associated with reporting fewer confidants, which in turn was associated with poorer mental health. In pathway 2, online learning was associated with reporting better sleep, which in turn was associated with better mental health. METHODS: We analyzed survey data from a US sample (N=540) of 13- to 17-year-olds to estimate how school modality was associated with mental health via the 2 pathways. The sample was recruited from the AmeriSpeak Teen Panel during spring of 2021, with an oversample of Black and Latino respondents. Ethnoracial categories were Black, Latino, White, and other. Mental health was measured with the 4-item Patient Health Questionnaire, which assesses self-reported frequency of experiencing symptoms consistent with anxiety and depression. School modality was recorded as either fully online or with some in-person component (fully in-person or hybrid). We recorded self-reports of the number of confidants and quality of sleep. Covariates included additional demographics and access to high-speed internet. We estimated bivariate associations between ethnoracial group membership and both school modality and mental health. To test the pathways, we estimated a path model. RESULTS: Black and Latino respondents were more likely to report being in fully online learning than their White counterparts (P<.001). Respondents in fully online learning reported fewer confidants than those with any in-person learning component (ß=-.403; P=.001), and reporting fewer confidants was associated with an increased likelihood of reporting symptoms consistent with anxiety (ß=-.121; P=.01) and depression (ß=-.197; P<.001). Fully online learning respondents also reported fewer concerns of insufficient sleep than their in-person learning counterparts (ß=-.162; P=.006), and reporting fewer concerns was associated with a decreased likelihood of reporting symptoms consistent with anxiety (ß=.601; P<.001) and depression (ß=.588; P<.001). Because of these countervailing pathways, the total effect of membership in a minoritized ethnoracial group on mental health was nonsignificant. CONCLUSIONS: The findings compel more nuanced discussions about the consequences of online learning and theorizing about the pandemic's impact on minoritized ethnoracial groups. While online learning may be a detriment to social connections, it appears to benefit sleep. Interventions should foster social connections in online learning and improve sleep, such as implementing policies to enable later start times for classes. Future research should incorporate administrative data about school modality, rather than relying on self-reports.