AI Model Detects Fatty Liver Disease from Chest X Ray Images

ENG: Fatty liver disease, a condition caused by the buildup of fat in the liver, affects roughly one in four people worldwide and can lead to cirrhosis or liver cancer if left untreated. While ultrasounds, CT scans, and MRIs are the standard diagnostic tools, they require specialized equipment and are costly. Chest X-rays, by comparison, are inexpensive, widely available, and involve lower radiation exposure. Although they are mainly used to assess the lungs and heart, they also capture part of the liver, which makes them a potential tool for detecting fatty liver disease.

Credit: Shutterstock
Read More

New Genetic Models Offer Insight into Autism Spectrum Disorder

ENG: A research team from Kobe University, led by neuroscientist Takumi Toru, has developed a groundbreaking resource for studying the genetic causes of autism spectrum disorder (ASD). By combining traditional methods of manipulating mouse embryonic stem cells with the precision of the CRISPR gene editing system, the team created a bank of 63 embryonic stem cell lines carrying mutations most strongly associated with ASD. These cell lines were shown to differentiate into a wide variety of cell types and even form adult mice with the targeted genetic alterations, establishing them as reliable models for investigating how specific mutations impact cellular and behavioral traits related to autism.

Credit: Nomura Jun
Read More

Why Vision Language Models Fail to Understand Negation

ENG: A new study by researchers at MIT reveals a critical weakness in vision-language models: their inability to process negation. This flaw can lead to serious consequences, especially in high-stakes environments like healthcare. For example, a radiologist searching for similar X-rays using a model might receive misleading results if the model ignores key phrases such as “no enlarged heart.” Without recognizing negation, the model may retrieve cases that include both tissue swelling and an enlarged heart, potentially pointing to incorrect diagnoses and treatment paths.

Credit: MIT
Read More

Unlocking Early Memories in the Infant Brain

ENG: Though adults typically can’t recall specific experiences from their earliest years, new research from Yale challenges the long-held belief that this is due to the immaturity of the hippocampus, the part of the brain responsible for memory formation. In a recent study, researchers observed that infants as young as four months old showed signs of memory encoding. When these infants were shown new images, those with higher hippocampal activity during the initial viewing were more likely to recognize the images later. This indicates that the brain is capable of forming episodic memories tied to specific events, even in early infancy.

Credit: 160/90
Read More