Ahmad Shalbaf

Objective, quantifiable electroencephalogram (EEG) biomarkers and machine learning systems for schizophrenia (SZ) research are limited by the scarcity of large, diverse, and high-quality public datasets. To address this gap, we present ASEEG (Atieh Schizophrenia EEG), a resting-state EEG dataset designed to support methodological development and benchmarking in SZ studies. ASEEG comprises 198 recordings from 101 subjects, including 51 individuals diagnosed with SZ according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria and 50 demographically matched healthy controls. Data were collected at a clinical setting using a standardized 19-channel EEG montage (international 10–20 system) with a sampling rate of 500 Hz. For each subject, both eyes-open and eyes-closed resting-state conditions were recorded for five minutes each, enabling analysis of complementary neural dynamics. The dataset spans a wide age range (15–65 years) and provides balanced diagnostic groups and extended recording durations. ASEEG is intended for use in biomarker extraction, machine learning model development, validation, and comparative evaluation of EEG-based approaches for SZ research.

Attention-deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder with serious long-term effects if untreated, emphasizing the need for early treatment given its neurobiological heterogeneity. This study introduces a novel explainable machine learning framework to predict neurofeedback treatment response for personalized ADHD intervention, offering transparent and clinically actionable insights. Seventy-eight features, including demographic, behavioral, and personality questionnaire data from 72 ADHD patients (aged 6-68) from the two-decades brainclinics (TDBRAIN) database, were used. First, a preliminary statistical analysis selected 20 features, comprising NEO five-factor inventory (NEO-FFI) questions, behavioral, and demographic data (age, education, sleep) for further analysis. Then, four feature reduction methods, mutual information, ReliefF, minimal-redundancy-maximal-relevance, and sequential forward selection (SFS), were utilized to select the best features. Five classifiers, random forest (RF), support vector machine, logistic regression, artificial neural network, and adaptive boosting, were used to predict neurofeedback treatment response in individuals with ADHD. Subsequently, shapley additive explanations (SHAP) values via TreeExplainer were crucial for interpretability, providing global feature importance and local explanations on model predictions. The results revealed that a hierarchical feature selection approach involved initial statistical filtering, followed by the SFS method, significantly improved the RF model's discrimination to 88.3 ± 6.8% accuracy with just seven optimal features, including NEO-FFI questions and education. Notably, five of the seven SFS features were among SHAP's top 10 most significant, demonstrating the internal consistency of the model and highlighting the features most critical to the model's prediction. Therefore, this transparent machine learning approach achieves competitively higher prediction performance than previous studies and supports personalized, trustworthy ADHD medical decisions, moving beyond black-box prediction.

Introduction: Mild cognitive impairment (MCI) is the stage that occurs before Alzheimer’s disease (AD), and there is a high risk of progression to AD. However, this progression is not guaranteed, and there is a chance of remaining at this stage. This study aimed to diagnose possible AD progression among patients with MCI ...

The global rise in fatty liver diseases is alarming. Traditional diagnostic methods include ultrasound, CT scans, MRI, and liver biopsies, the latter being the gold standard for diagnosis and treatment. Recent advancements in artificial intelligence (AI) have enhanced liver biopsy accuracy, improving treatment outcomes. This study investigates how various AI techniques aid histopathologists, gastroenterologists, and liver specialists in diagnosing and assessing liver damage due to abnormal fat accumulation. We conducted a systematic review of AI applications in evaluating fatty liver diseases, particularly through histopathological image analysis. Our search encompassed five scientific databases: PubMed Central, ACM Digital Library, IEEE Xplore, Scopus, and Google Scholar. We focused on peer‐reviewed articles, conference papers, theses, and book chapters, adhering to specific terminology. The data synthesis followed the PRISMA guidelines, comparing literature based on four key indices and their annual distribution. We evaluated 37 studies utilizing histopathological imaging for the diagnosis of non‐alcoholic fatty liver disease and non‐alcoholic steatohepatitis, including related conditions, metabolic dysfunction‐associated fatty liver disease and metabolic dysfunction‐associated steatohepatitis. The review summarized the performance of various algorithms and explored the distribution of machine learning efforts. Given the complexity of histopathological images, AI algorithms can effectively stratify liver samples affected by fat. Our findings indicate that AI's diagnostic performance closely matches traditional pathological interpretations, offering reliable results for clinical applications.This article is categorized under: Application Areas > Health Care Technologies > Machine Learning Technologies > Artificial Intelligence