Nonhuman primates, particularly rhesus macaques, have served as crucial animal models for investigating complex brain functions. While previous studies have explored neural activity features in macaques, the gyro-sulcal functional dissociation characteristics are largely unknown. In this study, we employ a deep learning model named one-dimensional convolutional neural network to differentiate resting state functional magnetic resonance imaging signals between gyri and sulci in macaque brains, and further investigate the frequency-specific dissociations between gyri and sulci inferred from the power spectral density of resting state functional magnetic resonance imaging. Experimental results based on a large cohort of 440 macaques from two independent sites demonstrate substantial frequency-specific dissociation between gyral and sulcal signals at both whole-brain and regional levels. The magnitude of gyral power spectral density is significantly larger than that of sulcal power spectral density within the range of 0.01 to 0.1 Hz, suggesting that gyri and sulci may play distinct roles as the global hubs and local processing units for functional activity transmission and interaction in macaque brains. In conclusion, our study has established one of the first power spectrum maps of gyro-sulcal functional activity dissociation in macaque brains, providing a novel perspective for systematically exploring the neural mechanism of functional dissociation in mammalian brains.

Artificial intelligence (AI) models have been applied to differential dementia detection tasks in brain images from curated, high-quality benchmark databases, but not real-world data in hospitals. We describe a deep learning model specially trained for disease detection in heterogeneous clinical images from electronic health records without focusing on confounding factors. It encodes up to 14 multimodal images, alongside age and demographics, and outputs the likelihood of vascular dementia, Alzheimer's, Lewy body dementia, Pick's disease, mild cognitive impairment, and unspecified dementia. We use data from Massachusetts General Hospital (183,018 images from 11,015 patients) for training and external data (125,493 images from 6,662 patients) for testing. Performance ranged between 0.82 and 0.94 area under the curve (AUC) on data from 1003 sites. Analysis shows that the model focused on subcortical brain structures as the basis for its decisions. By detecting biomarkers in real-world data, the presented techniques will help with clinical translation of disease detection AI. Our artificial intelligence (AI) model can detect neurodegenerative disorders in brain imaging electronic health record (EHR) data. It encodes up to 14 brain images and text information from a single patient's EHR. Attention maps show that the model focuses on subcortical brain structures. Performance ranged from 0.82 to 0.94 area under the curve (AUC) on data from 1003 external sites.

This study aimed to develop and validate a predictive model to detect osteoporosis using radiomic features and machine learning (ML) approaches from lumbar spine computed tomography (CT) images during an abdominal CT examination. A total of 509 patients who underwent both quantitative CT (QCT) and abdominal CT examinations (training group, n = 279; internal validation group, n = 120; external validation group, n = 110) were analyzed in this retrospective study from two centers. Radiomic features were extracted from the lumbar spine CT images. Seven radiomic-based ML models, including logistic regression (LR), Bernoulli, Gaussian NB, SGD, decision tree, support vector machine (SVM), and K-nearest neighbor (KNN) models, were constructed. The performance of the models was assessed using the area under the curve (AUC) of receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA). The radiomic model based on LR in the internal validation group and external validation group had excellent performance, with an AUC of 0.960 and 0.786 for differentiating osteoporosis from normal BMD and osteopenia, respectively. The radiomic model based on LR in the internal validation group and Gaussian NB model in the external validation group yielded the highest performance, with an AUC of 0.905 and 0.839 for discriminating normal BMD from osteopenia and osteoporosis, respectively. DCA in the internal validation group revealed that the LR model had greater net benefit than the other models in differentiating osteoporosis from normal BMD and osteopenia. Radiomic-based ML approaches may be used to predict osteoporosis from abdominal CT images and as a tool for opportunistic osteoporosis screening.

Given artificial intelligence's transformative effects, studying safety is important to ensure it is implemented in a beneficial way. Convolutional neural networks are used in radiology research for prediction but can be corrupted through adversarial attacks. This study investigates the effect of an adversarial attack, through poisoned data. To improve generalizability, we create a generic ResNet pneumonia classification model and then use it as an example by subjecting it to BadNets adversarial attacks. The study uses various poisoned datasets of different compositions (2%, 16.7% and 100% ratios of poisoned data) and two different test sets (a normal set of test data and one that contained poisoned images) to study the effects of BadNets. To provide a visual effect of the progressing corruption of the models, SHapley Additive exPlanations (SHAP) were used. As corruption progressed, interval analysis revealed that performance on a valid test set decreased while the model learned to predict better on a poisoned test set. SHAP visualization showed focus on the trigger. In the 16.7% poisoned model, SHAP focus did not fixate on the trigger in the normal test set. Minimal effects were seen in the 2% model. SHAP visualization showed decreasing performance was correlated with increasing focus on the trigger. Corruption could potentially be masked in the 16.7% model unless subjected specifically to poisoned data. A minimum threshold for corruption may exist. The study demonstrates insights that can be further studied in future work and with future models. It also identifies areas of potential intervention for safeguarding models against adversarial attacks.

Clinical two-dimensional (2D) MRI data has seen limited application in the early diagnosis of Parkinson's disease (PD) and multiple system atrophy (MSA) due to quality limitations, yet its diagnostic and therapeutic potential remains underexplored. This study presents a novel machine learning framework using reconstructed clinical images to accurately distinguish PD from MSA and identify disease-specific neuroimaging biomarkers. The structure constrained super-resolution network (SCSRN) algorithm was employed to reconstruct clinical 2D MRI data for 56 PD and 58 MSA patients. Features were derived from a functional template, and hierarchical SHAP-based feature selection improved model accuracy and interpretability. In the test set, the Extra Trees and logistic regression models based on the functional template demonstrated an improved accuracy rate of 95.65% and an AUC of 99%. The positive and negative impacts of various features predicting PD and MSA were clarified, with larger fourth ventricular and smaller brainstem volumes being most significant. The proposed framework provides new insights into the comprehensive utilization of clinical 2D MRI images to explore underlying neuroimaging biomarkers that can distinguish between PD and MSA, highlighting disease-specific alterations in brain morphology observed in these conditions.

The stomach is one of the main digestive organs in the GIT, essential for digestion and nutrient absorption. However, various gastrointestinal diseases, including gastritis, ulcers, and cancer, affect health and quality of life severely. The precise diagnosis of gastrointestinal (GI) tract diseases is a significant challenge in the field of healthcare, as misclassification leads to late prescriptions and negative consequences for patients. Even with the advancement in machine learning and explainable AI for medical image analysis, existing methods tend to have high false negative rates which compromise critical disease cases. This paper presents a hybrid deep learning based explainable artificial intelligence (XAI) approach to improve the accuracy of gastrointestinal disorder diagnosis, including stomach diseases, from images acquired endoscopically. Swin Transformer with DCNN (EfficientNet-B3, ResNet-50) is integrated to improve both the accuracy of diagnostics and the interpretability of the model to extract robust features. Stacked machine learning classifiers with meta-loss and XAI techniques (Grad-CAM) are combined to minimize false negatives, which helps in early and accurate medical diagnoses in GI tract disease evaluation. The proposed model successfully achieved an accuracy of 93.79% with a lower misclassification rate, which is effective for gastrointestinal tract disease classification. Class-wise performance metrics, such as precision, recall, and F1-score, show considerable improvements with false-negative rates being reduced. AI-driven GI tract disease diagnosis becomes more accessible for medical professionals through Grad-CAM because it provides visual explanations about model predictions. This study makes the prospect of using a synergistic DL with XAI open for improvement towards early diagnosis with fewer human errors and also guiding doctors handling gastrointestinal diseases.

The present study aimed to develop and validate a fusion model based on multi-phase contrast-enhanced computed tomography (CECT) radiomics features combined with clinical features to preoperatively predict the expression levels of Ki-67 in patients with gastric cancer (GC). A total of 164 patients with GC who underwent surgical treatment at our hospital between September 2015 and September 2023 were retrospectively included and were randomly divided into a training set (n=114) and a testing set (n=50). Using Pyradiomics, radiomics features were extracted from multi-phase CECT images and were combined with significant clinical features through various machine learning algorithms [support vector machine (SVM), random forest (RandomForest), K-nearest neighbors (KNN), LightGBM and XGBoost] to build a fusion model. Receiver operating characteristic, area under the curve (AUC), calibration curve and decision curve analysis (DCA) were used to evaluate, validate and compare the predictive performance and clinical utility of the model. Among the three single-phase models, for the arterial phase model, the SVM radiomics model had the highest AUC value in the training set, which was 0.697; and the RandomForest radiomics model had the highest AUC value in the testing set, which was 0.658. For the venous phase model, the SVM radiomics model had the highest AUC value in the training set, which was 0.783; and the LightGBM radiomics model had the highest AUC value in the testing set, which was 0.747. For the delayed phase model, the KNN radiomics model had the highest AUC value in the training set, which was 0.772; and the SVM radiomics model had the highest AUC in the testing set, which was 0.719. The clinical feature model had the lowest AUC values in both the training set and the testing set, which were 0.614 and 0.520, respectively. Notably, the multi-phase model and the fusion model, which were constructed by combining the clinical features and the multi-phase features, demonstrated excellent discriminative performance, with the fusion model achieving AUC values of 0.933 and 0.817 in the training and testing sets, thus outperforming other models (DeLong test, both P<0.05). The calibration curve showed that the fusion model had goodness of fit (Hosmer-Lemeshow test, >0.5 in the training and validation sets). The DCA showed that the net benefit of the fusion model in identifying high expression of Ki-67 was improved compared with that of other models. Furthermore, the fusion model achieved an AUC value of 0.805 in the external validation data from The Cancer Imaging Archive. In conclusion, the fusion model established in the present study was revealed to have excellent performance and is expected to serve as a non-invasive tool for predicting Ki-67 status and guiding clinical treatment.

Parkinson's disease (PD) is a devastating neurological ailment affecting both mobility and cognitive function, posing considerable problems to the health of the elderly across the world. The absence of a conclusive treatment underscores the requirement to investigate cutting-edge diagnostic techniques to improve patient outcomes. Machine learning (ML) has the potential to revolutionize PD detection by applying large repositories of structured data to enhance diagnostic accuracy. 133 papers published between 2021 and April 2024 were reviewed using a systematic literature review (SLR) methodology, and subsequently classified into five categories: acoustic data, biomarkers, medical imaging, movement data, and multimodal datasets. This comprehensive analysis offers valuable insights into the applications of ML in PD diagnosis. Our SLR identifies the datasets and ML algorithms used for PD diagnosis, as well as their merits, limitations, and evaluation factors. We also discuss challenges, future directions, and outstanding issues.

In this study, we explore the potential of ten quantitative (radiofrequency-based) ultrasound parameters to assess the progressive loss of collagen and proteoglycans, mimicking an osteoarthritis condition in ex-vivo bovine cartilage samples. Most analyzed metrics showed significant changes as the degradation progressed, especially with collagenase treatment. We propose for the first time a combination of these ultrasound parameters through machine learning models aimed at automatically identifying healthy and degraded cartilage samples. The random forest model showed good performance in distinguishing healthy cartilage from trypsin-treated samples, with an accuracy of 60%. The support vector machine demonstrated excellent accuracy (96%) in differentiating healthy cartilage from collagenase-degraded samples. Histological and mechanical analyses further confirmed these findings, with collagenase having a more pronounced impact on both mechanical and histological properties, compared to trypsin. These metrics were obtained using an ultrasound probe having a transmission frequency of 15 MHz, typically used for the diagnosis of musculoskeletal diseases, enabling a fully non-invasive procedure without requiring arthroscopic probes. As a perspective, the proposed quantitative ultrasound assessment has the potential to become a new standard for monitoring cartilage health, enabling the early detection of cartilage pathologies and timely interventions.

Deep learning-based diagnostic models often suffer performance drops due to distribution shifts between training (source) and test (target) domains. Collecting and labeling sufficient target domain data for model retraining represents an optimal solution, yet is limited by time and scarce resources. Active learning (AL) offers an efficient approach to reduce annotation costs while maintaining performance, but struggles to handle the challenge posed by distribution variations across different datasets. In this study, we propose a novel unsupervised Active learning framework for Domain Adaptation, named ADAptation, which efficiently selects informative samples from multi-domain data pools under limited annotation budget. As a fundamental step, our method first utilizes the distribution homogenization capabilities of diffusion models to bridge cross-dataset gaps by translating target images into source-domain style. We then introduce two key innovations: (a) a hypersphere-constrained contrastive learning network for compact feature clustering, and (b) a dual-scoring mechanism that quantifies and balances sample uncertainty and representativeness. Extensive experiments on four breast ultrasound datasets (three public and one in-house/multi-center) across five common deep classifiers demonstrate that our method surpasses existing strong AL-based competitors, validating its effectiveness and generalization for clinical domain adaptation. The code is available at the anonymized link: https://github.com/miccai25-966/ADAptation.