EGGROLL is a novel optimization algorithm that utilizes low-rank perturbations to significantly enhance the scalability and efficiency of Evolution Strategies for training large neural networks, achieving up to a hundredfold improvement in training throughput while maintaining competitive performance.
Schizophrenia stems from abnormal brain development, which can begin even before birth. Yet symptoms typically don’t appear until later in life.
“For a long time, the brain is able to compensate for developmental errors and maintain relatively normal function. But at some point, it’s like a chain snapping — the brain can no longer compensate, and that’s when symptoms emerge. Until that point, however, prevention should be possible,” says one of the study’s first authors.
They investigated when this turning point occurs. By tracking brain development from the fetal stage to adulthood, they found that dramatic changes happen late in the brain’s development. Up until the transition from childhood to adolescence, molecular and functional changes in the brain were rather minor, likely explaining lack of symptoms before adolescence.
The researchers have worked with mice carrying a specific genetic mutation known as “15q13.3 microdeletion syndrome.” In humans, this syndrome is associated with epilepsy, schizophrenia, autism, and other neurodevelopmental disorders.
“We know that sleep is often disrupted in people with psychiatric disorders, so we chose to use sleep as a behavioral marker—something we could observe. We examined both the mice’s behavior and the activity of a specific type of brain cell. Our findings show that one particular cell type (γ-aminobutyric acid (GABAergic) projecting neurons) is significantly affected in the test animals compared to healthy mice,” explains the author.
These GABAergic rare brain cells are often overlooked because they make up only a tiny fraction of the brain’s total cell population. Nevertheless, they play a crucial role in regulating many brain functions.
The new study not only demonstrates a link between this specific type of brain cell and sleep — it also shows that the mice’s sleep patterns began to resemble those of healthy mice when researchers reduced the activity of the cell type in question.
Unlike most alien planetary invasion methods in the Dark Forest universe, Pluribus acts as a cosmic Trojan Horse, an interstellar gift engineered to disarm an entire civilization the moment it’s opened. Sent to Earth by an alien beacon from a relatively nearby star system, Pluribus hides behind the appearance of progress. Even if the extraterrestrial senders turn out to be benevolent, their initiative still aligns with the Dark Forest Hypothesis.
Chapters:
00:00 Pluribus Signal as a Weapon.
03:01 Galactic Disarmament.
04:10 Humanity’s Defense.
06:03 The Dark Forest Hypothesis.
Footage:
Produced in part with SpaceEngine PRO © Cosmographic Software LLC.
Some elements in this video are also made with the help of artificial intelligence.
Music:
Atlantis by Audionautix is licensed under a Creative Commons Attribution 4.0 license. https://creativecommons.org/licenses/.… to Find Stellardrone’s Music: https://stellardrone.bandcamp.com Feedback or inquiries: [email protected].
Higher levels of fine particulate matter air pollution was associated with increased dementia severity and increased Alzheimer disease neuropathologic change.
Importance Exposure to fine particulate matter air pollution (PM2.5) may increase risk for dementia. It is unknown whether this association is mediated by dementia-related neuropathologic change found at autopsy.
Objective To examine associations between PM2.5 exposure, dementia severity, and dementia-associated neuropathologic change.
Design, Setting, and Participants This cohort study used data associated with autopsy cases collected from 1999 to 2022 at the Center for Neurodegenerative Disease Research Brain Bank at the University of Pennsylvania. Data were analyzed from January to June 2025. Participants included 602 cases with common forms of dementia and/or movement disorders and older controls after excluding 429 cases with missing data on neuropathologic measures, demographic factors, APOE genotype, or residential address.
Researchers at Karolinska Institutet in Sweden have identified a brain circuit that can drive repetitive and compulsive behaviors in mice, even when natural rewards such as food or social contact are available. The study has been published in the journal Science Advances and may contribute to increased knowledge about obsessive-compulsive disorder and addiction.
Both animals and humans can become stuck in certain behaviors, but exactly how this is regulated in the brain has been unknown. Now, researchers have been able to show that a specific nerve circuit in the brain can put behaviors into a kind of “repeat mode,” where mice continue to perform the same actions over and over again, even when there is no longer any reward.
The researchers investigated a neural circuit that runs from the nucleus accumbens, part of the brain’s reward system, to a region in the hypothalamus, which in turn is connected to the lateral habenula, an area that processes unpleasant experiences. By activating this circuit using optogenetics, a method in which nerve cells are controlled by light, the researchers were able to induce a negative state in mice that led to repetitive behaviors such as digging and sniffing—even when food or other rewards were available.
All of modern mathematics is built on the foundation of set theory, the study of how to organize abstract collections of objects. But in general, research mathematicians don’t need to think about it when they’re solving their problems. They can take it for granted that sets behave the way they’d expect, and carry on with their work.
Descriptive set theorists are an exception. This small community of mathematicians never stopped studying the fundamental nature of sets — particularly the strange infinite ones that other mathematicians ignore.
Their field just got a lot less lonely. In 2023, a mathematician named Anton Bernshteyn (opens a new tab) published a deep and surprising connection (opens a new tab) between the remote mathematical frontier of descriptive set theory and modern computer science.
#artificialintelligence
Agentic AI is a form of artificial intelligence that does more than just generate; it will act, reason somewhat, collaborate, and execute on its own. Agentic AI transforms its role from a limited tool to that of a collaborative coworker.
This shift affects various sectors, including cybersecurity, national defense, healthcare, key infrastructure, finance, supply chains, and corporate automation. Additionally, it accelerates the integration of robotics, neuromorphic systems, sensor-driven edge computing, and artificial intelligence.
Systems with the ability to plan and pursue goals characterize Agentic AI. IIt combines APIs and tools, engages with dynamic environments, makes decisions, uses reasoning, and continues to learn and adapt.
These images have been selected to showcase the art that neuroscience research can create.
As described by the authors: The cacophony voltage-gated calcium channel serves as the primary conduit for the calcium that triggers neurotransmitter release at countless synapses across the fruit fly (Drosophila) nervous system. To support this role at different synapse types, alternate splicing confers different biophysical properties upon cacophony. However, conventional techniques that might discriminate splice isoforms, such as antibodies, toxins, and pharmacological agents, are poorly suited for identifying splice isoforms across multiple neurons in a living nervous system.
This image demonstrates the transgenic expression of a bichromatic fluorescent exon reporter in most neurons of the fly brain. Green fluorescent protein (GFP) fluorescence was particularly bright relative to red fluorescent protein (TagRFP) in the α, β, and γ lobes of the mushroom body (MB), indicating a bias towards the inclusion of exon 11 at the expense of exon 10. Differences were also evident between neurons of the optic lobes.