Brain is a complex network with an anatomical and functional organization. The differences in brain organization of those with attention-deficit/hyperactivity disorder (ADHD) are still not well understood. Here, we study brain organization in ADHD subjects using a complex network derived from resting-state functional magnetic resonance imaging (fMRI) data of ADHD and normal subjects. Our results reveal that the brain networks of ADHD subjects are reorganized compared to those without ADHD in global and local brain functional networks. We find that the ADHD subjects show decreasing brain network integration and increasing brain network segregation. More interestingly, we find similarities of brain topology properties between local and global brain networks. Our finding indicates that cognitive dysfunction in ADHD is probably associated with disrupted global and local brain network topological properties. Our results can help us understand the pathophysiological mechanism of ADHD and serve as a sensitive and specific biomarker of ADHD.

In this paper, we first investigate why typical two-stage methods are not as fast as single-stage, fast detectors like YOLO [26,27] and SSD [22]. We find that Faster R-CNN [28] and R-FCN [17] perform an intensive computation after or before RoI warping. Faster R-CNN involves two fully connected layers for RoI recognition, while R-FCN produces a large score maps. Thus, the speed of these networks is slow due to the heavy-head design in the architecture. Even if we significantly reduce the base model, the computation cost cannot be largely decreased accordingly.We propose a new two-stage detector, Light-Head R-CNN, to address the shortcoming in current two-stage approaches. In our design, we make the head of network as light as possible, by using a thin feature map and a cheap R-CNN subnet (pooling and single fully-connected layer). Our ResNet-101 based light-head R-CNN outperforms state-of-art object detectors on COCO while keeping time efficiency. More importantly, simply replacing the backbone with a tiny network (e.g, Xception), our Light-Head R-CNN gets 30.7 mmAP at 102 FPS on COCO, significantly outperforming the single-stage, fast detectors like YOLO [26,27] and SSD [22] on both speed and accuracy. Code will be made publicly available.

Background Accurate and robust pathological image analysis for colorectal cancer (CRC) diagnosis is time-consuming and knowledge-intensive, but is essential for CRC patients’ treatment. The current heavy workload of pathologists in clinics/hospitals may easily lead to unconscious misdiagnosis of CRC based on daily image analyses. Methods Based on a state-of-the-art transfer-learned deep convolutional neural network in artificial intelligence (AI), we proposed a novel patch aggregation strategy for clinic CRC diagnosis using weakly labeled pathological whole-slide image (WSI) patches. This approach was trained and validated using an unprecedented and enormously large number of 170,099 patches, > 14,680 WSIs, from > 9631 subjects that covered diverse and representative clinical cases from multi-independent-sources across China, the USA, and Germany. Results Our innovative AI tool consistently and nearly perfectly agreed with (average Kappa statistic 0.896) and even often better than most of the experienced expert pathologists when tested in diagnosing CRC WSIs from multicenters. The average area under the receiver operating characteristics curve (AUC) of AI was greater than that of the pathologists (0.988 vs 0.970) and achieved the best performance among the application of other AI methods to CRC diagnosis. Our AI-generated heatmap highlights the image regions of cancer tissue/cells. Conclusions This first-ever generalizable AI system can handle large amounts of WSIs consistently and robustly without potential bias due to fatigue commonly experienced by clinical pathologists. It will drastically alleviate the heavy clinical burden of daily pathology diagnosis and improve the treatment for CRC patients. This tool is generalizable to other cancer diagnosis based on image recognition.

The concealed information test (CIT) has drawn much attention and has been widely investigated in recent years. In this study, a novel CIT method based on denoised P3 and machine learning was proposed to improve the accuracy of lie detection. Thirty participants were chosen as the guilty and innocent participants to perform the paradigms of 3 types of stimuli. The electroencephalogram (EEG) signals were recorded and separated into many single trials. In order to enhance the signal noise ratio (SNR) of P3 components, the independent component analysis (ICA) method was adopted to separate non-P3 components (i.e., artifacts) from every single trial. In order to automatically identify the P3 independent components (ICs), a new method based on topography template was proposed to automatically identify the P3 ICs. Then the P3 waveforms with high SNR were reconstructed on Pz electrodes. Second, the 3 groups of features based on time,frequency, and wavelets were extracted from the reconstructed P3 waveforms. Finally, 2 classes of feature samples were used to train a support vector machine (SVM) classifier because it has higher performance compared with several other classifiers. Meanwhile, the optimal number of P3 ICs and some other parameter values in the classifiers were determined by the cross-validation procedures. The presented method achieved a balance test accuracy of 84.29% on detecting P3 components for the guilty and innocent participants. The presented method improves the efficiency of CIT in comparison with previous reported methods.

Xerostomia induced by radiotherapy is a common toxicity for head and neck carcinoma patients. In this study, the deformable image registration of planning computed tomography (CT) and weekly cone‐beam CT (CBCT) was used to override the Hounsfield unit value of CBCT, and the modified CBCT was introduced to estimate the radiation dose delivered during the course of treatment. Herein, the beams from each patient's treatment plan were applied to the modified CBCT to construct the weekly delivered dose. Then, weekly doses were summed together to obtain the accumulated dose. A total of 42 parotid glands (PGs) of 21 nasopharyngeal carcinoma patients were analyzed. Doses delivered to the parotid glands significantly increased compared with the planning doses. V20, V30, V40, Dmean, and D50 increased by 11.3%, 28.6%, 44.4%, 9.5%, and 8.4% respectively. Of the 21 patients included in the study, eight developed xerostomia and the remaining 13 did not. Both planning and delivered PG Dmean for all patients exceeded tolerance (26 Gy). Among the 21 patients, the planning dose and delivered dose of Dmean were 30.6 Gy and 33.6 Gy, respectively, for patients with xerostomia, and 26.3 Gy and 28.0 Gy, respectively, for patients without xerostomia. The D50 of the planning and delivered dose for patients was below tolerance (30 Gy). The results demonstrated that the p‐value of V20, V30, D50, and Dmean difference of the delivery dose between patients with xerostomia and patients without xerostomia was less than 0.05. However, for the planning dose, the significant dosimetric difference between the two groups only existed in D50 and Dmean. Xerostomia is closely related to V20, V30, D50, and Dmean.

Background Identifying cognitive impairment early enough could support timely intervention that may hinder or delay the trajectory of cognitive impairment, thus increasing the chances for successful cognitive aging. Objective We aimed to build a prediction model based on machine learning for cognitive impairment among Chinese community-dwelling elderly people with normal cognition. Methods A prospective cohort of 6718 older people from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) register, followed between 2008 and 2011, was used to develop and validate the prediction model. Participants were included if they were aged 60 years or above, were community-dwelling elderly people, and had a cognitive Mini-Mental State Examination (MMSE) score ≥18. They were excluded if they were diagnosed with a severe disease (eg, cancer and dementia) or were living in institutions. Cognitive impairment was identified using the Chinese version of the MMSE. Several machine learning algorithms (random forest, XGBoost, naïve Bayes, and logistic regression) were used to assess the 3-year risk of developing cognitive impairment. Optimal cutoffs and adjusted parameters were explored in validation data, and the model was further evaluated in test data. A nomogram was established to vividly present the prediction model. Results The mean age of the participants was 80.4 years (SD 10.3 years), and 50.85% (3416/6718) were female. During a 3-year follow-up, 991 (14.8%) participants were identified with cognitive impairment. Among 45 features, the following four features were finally selected to develop the model: age, instrumental activities of daily living, marital status, and baseline cognitive function. The concordance index of the model constructed by logistic regression was 0.814 (95% CI 0.781-0.846). Older people with normal cognitive functioning having a nomogram score of less than 170 were considered to have a low 3-year risk of cognitive impairment, and those with a score of 170 or greater were considered to have a high 3-year risk of cognitive impairment. Conclusions This simple and feasible cognitive impairment prediction model could identify community-dwelling elderly people at the greatest 3-year risk for cognitive impairment, which could help community nurses in the early identification of dementia.

Our results suggest that the choice of DTI acquisition protocol and post-processing methods can influence the accurate estimation and reproducibility of DTI measurements and should be considered carefully for clinical applications.

Machine-assisted pathological recognition has been focused on supervised learning (SL) that suffers from a significant annotation bottleneck. We propose a semi-supervised learning (SSL) method based on the mean teacher architecture using 13,111 whole slide images of colorectal cancer from 8803 subjects from 13 independent centers. SSL (~3150 labeled, ~40,950 unlabeled; ~6300 labeled, ~37,800 unlabeled patches) performs significantly better than the SL. No significant difference is found between SSL (~6300 labeled, ~37,800 unlabeled) and SL (~44,100 labeled) at patch-level diagnoses (area under the curve (AUC): 0.980 ± 0.014 vs. 0.987 ± 0.008, P value = 0.134) and patient-level diagnoses (AUC: 0.974 ± 0.013 vs. 0.980 ± 0.010, P value = 0.117), which is close to human pathologists (average AUC: 0.969). The evaluation on 15,000 lung and 294,912 lymph node images also confirm SSL can achieve similar performance as that of SL with massive annotations. SSL dramatically reduces the annotations, which has great potential to effectively build expert-level pathological artificial intelligence platforms in practice.