Machine Learning-Based Aggression Detection in Children with ADHD Using Sensor-Based Physical Activity Monitoring

Predicting cognitive frailty in community-dwelling older adults: a machine learning approach based on multidomain risk factors

Cognitive frailty (CF), a clinical syndrome involving both physical frailty (PF) and impaired cognition (IC), is associated with adverse health outcomes in older adults. This study aimed to identify key predictors of CF and develop a machine learning-based model for CF risk assessment using data from 2404 community-dwelling older adults in the Korean Frailty and Aging Cohort Study (2016-2017). PF was evaluated using Fried frailty phenotype, while IC was assessed using Mini-Mental State Examination (MMSE). Participants exhibiting at least one frailty phenotype and MMSE score ≤ 24 were classified as having CF. A comprehensive analysis encompassing sociodemographic, clinical, and health status characteristics was conducted. A machine learning approach incorporating recursive feature elimination and bootstrapping was employed to develop the prediction model. Among the diverse CF-associated characteristics, the machine learning-based model identified six optimal features (key predictors): motor capacity, education level, physical function limitation, nutritional status, balance confidence, and activities of daily living. The model demonstrated robust predictive performance, achieving an area under the curve of 84.34%, with high sensitivity, specificity, and accuracy. These findings underscore the importance of comprehensive health assessments for early CF detection and demonstrate the potential of predictive modeling in facilitating personalized interventions for at-risk older adults.

Automated ADHD detection using dual-modal sensory data and machine learning

This study explores using dual-modal sensory data and machine learning to objectively identify Attention-Deficit/Hyperactivity Disorder (ADHD), a neurodevelopmental disorder traditionally diagnosed through subjective clinical evaluations. Six machine learning algorithms, including Logistic Regression (LR), Random Forest (RF), XGBoost (XGB), LightGBM (LGBM), Neural Network (NN), and Support Vector Machine (SVM), were evaluated using both activity and heart rate variability (HRV) data collected from 103 participants. The results show that both activity and HRV data performed similarly when analyzed individually. However, when the two datasets were combined, the highest F1-score increased by 12 % compared to the activity data and 23 % compared to the HRV data. This combination leverages the complementary strengths of both data, representing a key contribution of our work. With the combined data, the SVM model performed best, achieving an F1-Score of 0.87 and a Matthews Correlation Coefficient of 0.77. This study highlights the significant potential of interdisciplinary collaboration and the use of diverse data sources to advance ADHD detection through cutting-edge machine learning techniques.

Advanced machine learning techniques reveal multidimensional EEG abnormalities in children with ADHD: a framework for automatic diagnosis

Introduction

Attention deficit hyperactivity disorder (ADHD) is a prevalent neurodevelopmental disorder that affects attention, impulse control, and multitasking abilities in children and adults. Understanding electroencephalography (EEG) characteristics of children with ADHD can provide new diagnostic tools and personalized treatment plans. This study aims to explore potentially promising EEG features using advanced machine learning techniques and feature selection technique (i.e., SHapley Additive exPlanations (SHAP) algorithm) to reveal brain function abnormalities between pediatric children with ADHD and healthy controls (HC) in a data-driven manner.

Methods

Multidimensional EEG characteristics were extracted from multiple domain (including power spectral density (PSD), fuzzy entropy (FuzEn), and functional connectivity features of mutual information (MI)) using a publicly-available dataset. Then, four widely-employed machine learning algorithms (including random forest (RF), XGBoost, CatBoost, and LightGBM) were used for classification calculations, and the SHAP algorithm was then used to assess the importance of the contributing features to interpret the model's decision process.

Results

The results showed that the highest classification accuracy of 99.58% for pediatric ADHD detection was obtained with the CatBoost model based on the optimal feature subset of 206 features (PSD/FuzEn/MI = 53/5/148). According to the optimal feature subset statistics, there is an increase in the power of theta, alpha, and beta rhythms, an elevated power ratio between theta and beta (theta/beta ratio, TBR), and reorganization of whole-brain functional connectivity across all frequency bands in children with ADHD, primarily characterized by enhanced functional connectivity.

Discussion

We showed that EEG features was effectively extracted by machine learning methods, which could play a critical role in classification between pediatric ADHD and HC. These findings provide strong evidence for revealing the electrophysiological mechanisms through multidimensional EEG characteristics and move a step forward towards future automatic diagnosis of ADHD.

Wearable sensors in paediatric neurology

Wearable sensors have the potential to transform diagnosis, monitoring, and management of children who have neurological conditions. Traditional methods for assessing neurological disorders rely on clinical scales and subjective measures. The snapshot of the disease progression at a particular time point, lack of cooperation by the children during assessments, and susceptibility to bias limit the utility of these measures. Wearable sensors, which capture data continuously in natural settings, offer a non-invasive and objective alternative to traditional methods. This review examines the role of wearable sensors in various paediatric neurological conditions, including cerebral palsy, epilepsy, autism spectrum disorder, attention-deficit/hyperactivity disorder, as well as Rett syndrome, Down syndrome, Angelman syndrome, Prader-Willi syndrome, neuromuscular disorders such as Duchenne muscular dystrophy and spinal muscular atrophy, ataxia, Gaucher disease, headaches, and sleep disorders. The review highlights their application in tracking motor function, seizure activity, and daily movement patterns to gain insights into disease progression and therapeutic response. Although challenges related to population size, compliance, ethics, and regulatory approval remain, wearable technology promises to improve clinical trials and outcomes for patients in paediatric neurology.

Putting the usability of wearable technology in forensic psychiatry to the test: a randomized crossover trial

Introduction

Forensic psychiatric patients receive treatment to address their violent and aggressive behavior with the aim of facilitating their safe reintegration into society. On average, these treatments are effective, but the magnitude of effect sizes tends to be small, even when considering more recent advancements in digital mental health innovations. Recent research indicates that wearable technology has positive effects on the physical and mental health of the general population, and may thus also be of use in forensic psychiatry, both for patients and staff members. Several applications and use cases of wearable technology hold promise, particularly for patients with mild intellectual disability or borderline intellectual functioning, as these devices are thought to be user-friendly and provide continuous daily feedback.

Method

In the current randomized crossover trial, we addressed several limitations from previous research and compared the (continuous) usability and acceptance of four selected wearable devices. Each device was worn for one week by staff members and patients, amounting to a total of four weeks. Two of the devices were general purpose fitness trackers, while the other two devices used custom made applications designed for bio-cueing and for providing insights into physiological reactivity to daily stressors and events.

Results

Our findings indicated significant differences in usability, acceptance and continuous use between devices. The highest usability scores were obtained for the two fitness trackers (Fitbit and Garmin) compared to the two devices employing custom made applications (Sense-IT and E4 dashboard). The results showed similar outcomes for patients and staff members.

Discussion

None of the devices obtained usability scores that would justify recommendation for future use considering international standards; a finding that raises concerns about the adaptation and uptake of wearable technology in the context of forensic psychiatry. We suggest that improvements in gamification and motivational aspects of wearable technology might be helpful to tackle several challenges related to wearable technology.

Improving the Concrete Crack Detection Process via a Hybrid Visual Transformer Algorithm

Inspections of concrete bridges across the United States represent a significant commitment of resources, given their biannual mandate for many structures. With a notable number of aging bridges, there is an imperative need to enhance the efficiency of these inspections. This study harnessed the power of computer vision to streamline the inspection process. Our experiment examined the efficacy of a state-of-the-art Visual Transformer (ViT) model combined with distinct image enhancement detector algorithms. We benchmarked against a deep learning Convolutional Neural Network (CNN) model. These models were applied to over 20,000 high-quality images from the Concrete Images for Classification dataset. Traditional crack detection methods often fall short due to their heavy reliance on time and resources. This research pioneers bridge inspection by integrating ViT with diverse image enhancement detectors, significantly improving concrete crack detection accuracy. Notably, a custom-built CNN achieves over 99% accuracy with substantially lower training time than ViT, making it an efficient solution for enhancing safety and resource conservation in infrastructure management. These advancements enhance safety by enabling reliable detection and timely maintenance, but they also align with Industry 4.0 objectives, automating manual inspections, reducing costs, and advancing technological integration in public infrastructure management.

Harnessing Speech-Derived Digital Biomarkers to Detect and Quantify Cognitive Decline Severity in Older Adults

INTRODUCTION

Current cognitive assessments suffer from floor/ceiling and practice effects, poor psychometric performance in mild cases, and repeated assessment effects. This study explores the use of digital speech analysis as an alternative tool for determining cognitive impairment. The study specifically focuses on identifying the digital speech biomarkers associated with cognitive impairment and its severity.

METHODS

We recruited older adults with varying cognitive health. Their speech data, recorded via a wearable microphone during the reading aloud of a standard passage, were processed to derive digital biomarkers such as timing, pitch, and loudness. Cohen's d effect size highlighted group differences, and correlations were drawn to the Montreal Cognitive Assessment (MoCA). A stepwise approach using a Random Forest model was implemented to distinguish cognitive states using speech data and predict MoCA scores based on highly correlated features.

RESULTS

The study comprised 59 participants, with 36 demonstrating cognitive impairment and 23 serving as cognitively intact controls. Among all assessed parameters, similarity, as determined by Dynamic Time Warping (DTW), exhibited the most substantial positive correlation (rho = 0.529, p < 0.001), while timing parameters, specifically the ratio of extra words, revealed the strongest negative correlation (rho = -0.441, p < 0.001) with MoCA scores. Optimal discriminative performance was achieved with a combination of four speech parameters: total pause time, speech-to-pause ratio, similarity via DTW, and intelligibility via DTW. Precision and balanced accuracy scores were found to be 88.1 ± 1.2% and 76.3 ± 1.3%, respectively.

DISCUSSION

Our research proposes that reading-derived speech data facilitates the differentiation between cognitively impaired individuals and cognitively intact, age-matched older adults. Specifically, parameters based on timing and similarity within speech data provide an effective gauge of cognitive impairment severity. These results suggest speech analysis as a viable digital biomarker for early detection and monitoring of cognitive impairment, offering novel approaches in dementia care.

Breast Cancer Detection with an Ensemble of Deep Learning Networks Using a Consensus-Adaptive Weighting Method

Breast cancer's high mortality rate is often linked to late diagnosis, with mammograms as key but sometimes limited tools in early detection. To enhance diagnostic accuracy and speed, this study introduces a novel computer-aided detection (CAD) ensemble system. This system incorporates advanced deep learning networks-EfficientNet, Xception, MobileNetV2, InceptionV3, and Resnet50-integrated via our innovative consensus-adaptive weighting (CAW) method. This method permits the dynamic adjustment of multiple deep networks, bolstering the system's detection capabilities. Our approach also addresses a major challenge in pixel-level data annotation of faster R-CNNs, highlighted in a prominent previous study. Evaluations on various datasets, including the cropped DDSM (Digital Database for Screening Mammography), DDSM, and INbreast, demonstrated the system's superior performance. In particular, our CAD system showed marked improvement on the cropped DDSM dataset, enhancing detection rates by approximately 1.59% and achieving an accuracy of 95.48%. This innovative system represents a significant advancement in early breast cancer detection, offering the potential for more precise and timely diagnosis, ultimately fostering improved patient outcomes.