This review addresses the growing global burden of hereditary hearing loss, summarizes key advances in gene therapy, and aims to accelerate its clinical translation.

Infantile epileptic spasms are among the most serious forms of epilepsy in early childhood, yet diagnosing them remains a major clinical challenge.

Today, seizure identification often depends on specialists reviewing hours of clinical video recordings. The process is time-consuming, resource-intensive, and difficult to scale, particularly in regions with limited access to pediatric neurology expertise.

To address this challenge, researchers developed a video-based AI system capable of automatically recognizing seizure-related movement patterns from patient videos.

Trained on a large clinical dataset, the model achieved more than 90% detection accuracy. In external validation studies, it demonstrated higher seizure-detection sensitivity than the average performance of six clinical experts from different medical centers.

The broader significance extends beyond algorithm performance. By transforming routine video recordings into a scalable screening and monitoring tool, such systems could support earlier diagnosis, reduce clinician workload, and expand access to specialized neurological assessment.

AI is not replacing clinicians. Rather, it offers a way to augment clinical expertise and help ensure that critical seizure events are less likely to be missed.

#AIinHealthcare #Neurology #Epilepsy #ComputerVision #DigitalHealth

Sleep disorders, including obstructive sleep apnea (OSA), chronic insomnia, and sleep deprivation, are increasingly linked to dysfunction of the blood-brain barrier (BBB), a key regulator of central nervous system homeostasis. Growing evidence demonstrates that BBB injury is associated with cognitive impairment, including Alzheimer’s disease (AD), white matter injury, cerebral small vessel disease, and vascular cognitive impairment. This review summarizes evidence that disturbed sleep impairs BBB integrity and discusses the underlying mechanisms. BBB injury may increase para-cellular permeability, alter endothelial transport, and impair metabolic clearance pathways, thereby facilitating the entry of circulating toxins into the brain and amplifying glial activation, cerebrovascular dysfunction, and neurodegenerative processes. Furthermore, BBB breakdown in the hippocampus may represent an early marker of cognitive dysfunction, which is partly independent of classical amyloid-β and tau pathology. However, this finding requires further validation in larger, longitudinal cohorts. Finally, we discuss the future of therapeutic strategies aimed at both sleep restoration and BBB-related mechanisms and potential biomarkers for assessing BBB dysfunction. We identify key knowledge gaps, including the lack of validated BBB biomarkers in sleep disorders and limited evidence for reversibility of BBB injury after treatment.

This study evaluates the clinical efficacy and safety of modified deep lateral wall decompression surgery in thyroid-associated ophthalmopathy and optic neuropathy, demonstrating significant postoperative improvements in vision, exophthalmos, and intraocular pressure with a low rate of complications.

Foundation and multimodal models now match expert-level performance on many diagnostic, prognostic, and biomarker tasks in pathology. Yet only a handful of AI systems are used in routine clinical practice, and the products that have reached patients differ substantially from research prototypes. We define this mismatch as the adoption paradox of computational pathology.

We first survey the technical landscape from task-specific deep learning to large unimodal foundation models, multimodal systems, and early agentic architectures. We then examine what has actually entered the clinic, identifying four product archetypes (digital pathology platforms, population scale cytology screening, assistive detection in surgical pathology, and quantitative immunohistochemistry scoring). Using a three stage maturity model algorithmic capability (Stage 1), system integration (Stage 2), and institutional adoption (Stage 3), we analyze the structural barriers that gate each transition.

Three interconnected barriers explain most of the gap: (1) Data and infrastructure fragility [(pre-analytical variability, scanner-induced domain shift, format fragmentation, annotation scarcity, manual quality control (QC)]; (2) Workflow misalignment (cognitive rhythm of pathologists, automation bias, scenario-dependent latency); (3) Institutional trust deficits (shallow interpretability, incomplete prospective validation, unclear reimbursement, unsettled liability, and regulatory gaps for generative/adaptive systems).

We outline system-level pathways for each stage, including infrastructure first, AI, workflow-embedded intelligence, and adaptive governance. Our central claim is that the next phase of progress will depend less on architectural novelty than on the slower institutional work that turns capability into clinical benefit. The framework provides an actionable lens for regulators, developers, and healthcare organizations to diagnose why a given AI system remains a prototype and what is needed to move it into routine use.

This review provides a critical assessment of the ongoing debate surrounding altermagnetism in ruthenium dioxide (RuO₂), synthesizes a decade of experimental and theoretical efforts, offering a clear roadmap of the studies conducted on RuO₂.

Type 2 diabetes (T2D) is marked by insulin resistance together with the progressive failure of pancreatic β-cells. Increasing evidence suggests that this failure is shaped not only by β-cells themselves, but also by the local islet immune microenvironment, where resident and recruited macrophages can influence inflammation, islet architecture, and insulin secretion. Exercise is a cornerstone of diabetes management, yet how it acts within this islet microenvironment has remained unclear.

While optical computing systems (OCS) with high bandwidth, low latency, and inherent parallelism are promising accelerators for artificial intelligence, the existing OCS implementations require direct participation of physical hardware, tightly coupling development workflows to device access and limiting offline design. Now, researchers have developed a Digital Twin OCS, a system-level, measurement-driven digital surrogate that enables a hardware-decoupled and fully offline development paradigm for OCS.