Autism spectrum disorder (ASD) necessitates prompt diagnostic scrutiny to enable immediate, targeted interventions. This study unveils an advanced convolutional-neural-network (CNN) algorithm that was meticulously engineered to examine resting-state functional magnetic resonance imaging (fMRI) for early ASD detection in pediatric cohorts. The CNN architecture amalgamates convolutional, pooling, batch-normalization, dropout, and fully connected layers, optimized for high-dimensional data interpretation. Rigorous preprocessing yielded 22,176 two-dimensional echo planar samples from 126 subjects (56 ASD, 70 controls) who were sourced from the Autism Brain Imaging Data Exchange (ABIDE I) repository. The model, trained on 17,740 samples across 50 epochs, demonstrated unparalleled diagnostic metrics—accuracy of 99.39%, recall of 98.80%, precision of 99.85%, and an F1 score of 99.32%—and thereby eclipsed extant computational methodologies. Feature map analyses substantiated the model’s hierarchical feature extraction capabilities. This research elucidates a deep learning framework for computer-assisted ASD screening via fMRI, with transformative implications for early diagnosis and intervention.

Existing research typically focuses on only one domain of cognition with regard to fairness-theory of mind or executive function. However, children with High-functioning autism spectrum disorder (HF-ASD) are cognitively impaired in both domains. Moreover, little is known about fairness characteristics in children with HF-ASD in relation to both domains of cognition. Thirty children with HF-ASD as well as 39 children with typical development (TD) were evaluated in this study. We investigated the development of children's fairness characteristics as a responder in a mini ultimatum game (UG). The different 'brain types,' i.e., with or without HF-ASD, were evaluated using the Empathy Questionnaire-Systemizing Questionnaire (E/SC-Q). Furthermore, we explored the relationship between fairness and brain types using Pearson correlation analyses. Children in the HF-ASD group were more likely to accept unfair offers than were children in the TD group (χ² = 17.513, p = .025). In the HF-ASD group, the acceptance rate of unfair offers was correlated with the discrepancy score (r = 0.363, p = .048), while there were no significant correlations in the TD group. In HF-ASD group, compared with Type S, acceptance rate of unfair offer was significant higher in Extreme Type S 'brain type' (F = 28.584, p < .001). While dividing TD participants by 'brain type', there was no significant difference in acceptance rate of unfair offer among five difference 'brain types' (F = 1.131, p = .358). Stepwise regression revealed that Extreme Type S positively predicted acceptance of unfair offers (F [1, 68] = 8.695, p < .001). Our findings show that children with HF-ASD were more likely to accept an unfair offer; in particular, the more unbalanced the development of empathy and systemizing was, the more significant the unfairness preference observed. Extreme Type S positively predicted the acceptance of unfair offers by children with HF-ASD. World Health Organization class I registered international clinical trial platform, ChiCTR-ROC-17012877.

Objective. There is a need to develop optimized, evidence-based parent training programs tailored for preschoolers with attention deficit hyperactivity disorder (ADHD). The objective of this study was to explore a behavioral management training program aimed at the parents of preschool children with ADHD, which directly analyzes parent-child interaction from the perspective of system theory, and the intervention effect on ADHD in preschool children. Methods. A multicenter randomized controlled study was conducted using system-based group therapy with 62 parents of preschool children with ADHD aged four to six years. ADHD symptoms, behavioral and emotional problems, and social functioning were compared with 61 control children whose parents did not receive training by applying the ADHD Rating Scale (ADHD-RS), Strengths and Difficulties Questionnaire (SDQ), and Questionnaire-Children with Difficulties (QCD) at the time of subject entry and at two and six months of entry, respectively. Results. The results of the ADHD-RS assessment showed that children in the intervention group had significantly lower factor scores for attention deficit, hyperactivity, and impulsivity than the children in the control group after parental training and at follow-up ( ). Total scores on the SDQ scale, as well as character problems, hyperactivity, and peer interaction scores, significantly decreased with statistically significant differences (all ), and emotional symptoms and prosocial behavior did not notable decline ( ). Compared with the control group, the total scores of the QCD scale and the scores of each factor in the intervention group remained significantly higher at the follow-up ( ). Conclusion. After continuous intervention for eight weeks, parents were able to help the children with preschool ADHD to improve their ADHD symptoms and emotional behavioral and social functioning significantly, and the efficacy was maintained at the four-month follow-up; the systemic-based parent training in behavior management (PTBM) is applicable to the treatment of preschool ADHD and is worth promoting.

Attention deficit hyperactivity disorder (ADHD) is a neuro-developmental disorder that affects approximately 5–10% of school-aged children worldwide. Early diagnosis and intervention are essential to improve the quality of life of patients and their families. In this study, we propose ConvMixer-ECA, a novel deep learning architecture that combines ConvMixer with efficient channel attention (ECA) blocks for the accurate diagnosis of ADHD using electroencephalogram (EEG) signals. The model was trained and evaluated using EEG recordings from 60 healthy children and 61 children with ADHD. A series of experiments were conducted to evaluate the performance of the ConvMixer-ECA. The results showed that the ConvMixer-ECA performed well in ADHD recognition with 94.52% accuracy. The incorporation of attentional mechanisms, in particular ECA, improved the performance of ConvMixer; it outperformed other attention-based variants. In addition, ConvMixer-ECA outperformed state-of-the-art deep learning models including EEGNet, CNN, RNN, LSTM, and GRU. t-SNE visualization of the output of this model layer validated the effectiveness of ConvMixer-ECA in capturing the underlying patterns and features that separate ADHD from typically developing individuals through hierarchical feature learning. These outcomes demonstrate the potential of ConvMixer-ECA as a valuable tool to assist clinicians in the early diagnosis and intervention of ADHD in children.

Structural health monitoring in resource-constrained environments demands crack segmentation models that match the accuracy of heavyweight convolutional networks while conforming to the power, memory, and latency limits of watt-level edge devices. This study presents a lightweight dual-attention network, which is a four-stage U-Net compressed to one-quarter of the channel depth and augmented—exclusively at the deepest layer—with a compact dual-attention block that couples channel excitation with spatial self-attention. The added mechanism increases computation by only 19%, limits the weight budget to 7.4 MB, and remains fully compatible with post-training INT8 quantization. On a pixel-labelled concrete crack benchmark, the proposed network achieves an intersection over union of 0.827 and an F1 score of 0.905, thus outperforming CrackTree, Hybrid 2020, MobileNetV3, and ESPNetv2. While refined weight initialization and Dice-augmented loss provide slight improvements, ablation experiments show that the dual-attention module is the main factor influencing accuracy. With 110 frames per second on a 10 W Jetson Nano and 220 frames per second on a 5 W Coral TPU achieved without observable accuracy loss, hardware-in-the-loop tests validate real-time viability. Thus, the proposed network offers cutting-edge crack segmentation at the kiloflop scale, thus facilitating ongoing, on-device civil infrastructure inspection.

This research introduces a novel approach for detecting defects in concrete structures. It utilizes the Gramian Angular Difference Field (GADF) in combination with a Convolutional Neural Network (CNN) enhanced by depthwise separable convolutions and attention mechanisms. The key contribution of this work is the use of GADF to transform one-dimensional impact-echo signals into two-dimensional images, thereby improving feature extraction and computational efficiency for analysis by the CNN. This advancement offers a new perspective in non-destructive testing technologies for concrete infrastructure. Comprehensive evaluation on a varied dataset of concrete structural defects reveals that our GADF-CNN model achieves an impressive test accuracy of 98.24 %, surpassing conventional models like VGG16, ResNet18, DenseNet, and ResNeXt50, and excelling in precision, recall, and F1-score metrics. Ultimately, this study enhances the integration of sophisticated image transformation techniques with deep learning, contributing to safer and more durable concrete infrastructure, and represents a noteworthy development in the field.