Neanderthal Bones Reveal Violence and Compassion Together
A Neanderthal who died about 75,000 years ago may have been killed by a wooden spear, yet survived an earlier injury with help from companions. Researchers studying remains from Shanidar Cave found a 15-millimeter circular hole in a shoulder blade. Its shape and trajectory suggest a sharpened wooden weapon entered the chest from above, possibly during an intentional attack, although a hunting accident cannot be excluded. Damaged front teeth reveal that the individual, known as Shanidar A, survived severe facial trauma for at least a year, potentially requiring assistance with eating. The remains were found among several Neanderthals placed near a prominent rock inside the cave. Published in Scientific Reports, the discovery illuminates both violence and social care in prehistoric communities. (Cambridge University)
Prenatal Depression Therapy May Shape Infant Brains
Treating depression during pregnancy may influence a baby’s developing brain, according to a randomized clinical trial reported in Nature Communications. Pregnant participants with elevated depressive symptoms were assigned either eight weekly sessions of interpersonal psychotherapy or enhanced usual care, including educational support. Researchers later examined their infants’ hippocampi, brain regions involved in memory and emotion. Babies in the psychotherapy group showed a roughly 5% difference in hippocampal development compared with the comparison group. The brain measurements were associated with less negative emotionality at six months. Unlike earlier observational studies, the randomized design provides stronger evidence that a prenatal mental-health intervention can affect early neurodevelopment. The findings do not establish lasting benefits, but support integrating evidence-based depression treatment into prenatal care. (Medical Xpress)
Astronomers Identify the First Known Green Star
Astronomers have identified what may be the first genuinely green star, overturning the familiar assumption that stars appear only reddish, white or blue. The object, WD J1330+6435, is a white dwarf about 280 light-years away whose atmosphere contains sodium, calcium and lithium from rocky planetary debris. Under the remnant’s extraordinary atmospheric pressure, sodium absorbs a broad swath of yellow, orange and red light, leaving a chartreuse appearance in models of human color vision. Researchers at the University of Warwick reconstructed stellar colors from spectra rather than direct photographs. Their survey of 151 nearby white dwarfs also uncovered unusually pink and yellow examples. Published in RAS Techniques and Instruments, the work shows how stellar chemistry, not just temperature, can determine a star’s apparent color. (University of Warwick)
Bacterial Virus Defense Hints at Telomerase Origins
A bacterial defense mechanism has revealed a surprising evolutionary connection to telomerase, the enzyme that helps protect the ends of human chromosomes. Researchers at Columbia University discovered DRT10, an antiviral system that uses RNA as a template to manufacture repetitive DNA rather than cutting invading genetic material as CRISPR does. Structural and evolutionary analyses suggest that DRT10-related enzymes and telomerase may share an ancient origin. In laboratory experiments, scientists also used DRT10 to synthesize human telomeric DNA repeats inside human cells. The finding, published in Nature, could eventually inspire new approaches to disorders involving defective telomere maintenance or improve the persistence of cancer-fighting immune cells. However, no treatment has been demonstrated, and researchers still need to establish whether the technology can safely lengthen functional telomeres. (CUIMC)
CERN Plans AI Model for Particle Physics
CERN and French artificial-intelligence developer Mistral AI (official website) have announced plans for a foundation model trained not on text, but on billions of simulated particle collisions. Their proposed Large Physics Model, dubbed Flare, would help reconstruct collisions inside the Large Hadron Collider, identify particles and flag unexpected signals that might point toward new physics. Currently, physicists rely on chains of specialized algorithms to transform raw detector measurements into interpretable events. A single broadly trained model could potentially handle several of those tasks and help engineers design future detectors. CERN already stores roughly one exabyte of collider data, with collision rates expected to rise dramatically after its planned upgrade. Announced October 15, the partnership is a development project, not a completed scientific breakthrough; no delivery timetable has been disclosed. (CERN)
Anesthesia Leaves More Cognition Intact Than Expected
Even when people are rendered unconscious by general anesthesia, the hippocampus may continue performing surprisingly sophisticated cognitive work. Baylor College of Medicine researchers recorded individual neurons from epilepsy patients under anesthesia using high-density Neuropixels probes. The cells distinguished unexpected tones, with discrimination improving during the experiment, suggesting that neural plasticity can occur without conscious awareness. When patients heard stories, hippocampal activity also encoded grammatical and semantic information and even anticipated features of upcoming words. The findings do not imply that anesthetized patients consciously understood what they heard. Instead, they suggest that capabilities often associated with conscious cognition—including learning, prediction and language analysis—can operate outside awareness. The results complicate attempts to define exactly what consciousness contributes to information processing in the human brain. (ScienceDaily)
Fear May Blur Memories of Danger and Safety
Fear may make memories of dangerous and safe environments more similar rather than helping the brain distinguish between them, according to research from McGill University and the University of Toronto Scarborough. Researchers used an augmented-reality system to switch mice between environments associated with danger and safety while miniature microscopes monitored hundreds of hippocampal neurons. Surprisingly, representations of the two environments increasingly overlapped. Researchers suggest this apparent loss of precision may actually be adaptive: combining related memories could allow animals to retrieve threat information more rapidly when danger appears. Problems may arise when this “better safe than sorry” mechanism becomes excessively sensitive. The findings offer a new framework for understanding why people with PTSD or anxiety disorders can experience powerful fear responses when encountering otherwise harmless situations that resemble previous threats. (McGill University)
People Still Draw a Line Around AI Consciousness
Humanlike behavior alone apparently does not convince people that artificial intelligence is conscious. Researchers studying 1,092 participants across five experiments compared judgments about humans and AI systems behaving identically under the same circumstances. Greater responsiveness caused both humans and machines to be rated as more mentally capable, but consciousness produced a striking exception: even the most responsive AI was judged less conscious than the least responsive human. When researchers substituted the word “awareness,” however, the divide largely disappeared. The findings suggest people may use an essentialist conception of consciousness—something humans inherently possess—while treating awareness more functionally, according to observable behavior. Researchers argue the distinction matters increasingly as companies and media describe AI systems using language suggesting intentions, plans, feelings or other seemingly human mental states. (ScienceDaily)
Breathwork Produces Psychedelic-Like Changes in Consciousness
Intensive breathing can produce profound changes in subjective experience alongside measurable shifts in brain activity, according to research presented at the European College of Neuropsychopharmacology Congress. Thirty experienced breathwork practitioners performed high-ventilation breathing while undergoing MRI scans. Cerebral blood flow fell roughly 45 percent, while participants reported altered perception, vivid imagery, diminished boundaries of self and feelings of unity resembling some psychedelic experiences. Importantly, the intensity of those experiences appeared to track changes in communication within the default mode network more closely than reduced blood flow itself. That network is strongly associated with self-referential thought. The researchers describe breathwork as a potentially useful drug-free model for investigating altered consciousness but stress that the study was small, uncontrolled and involved experienced practitioners, making larger randomized studies necessary. (Eureka Alert)
Brain Cells Encode Meaning Through Distributed Patterns
The human hippocampus appears to represent word meanings using distributed neural patterns with intriguing similarities to representations found in large language models. Researchers recorded individual hippocampal neurons from ten epilepsy patients as they listened to narrative speech. Neural activity encoded semantic information even after researchers accounted for sounds and grammatical features. Rather than individual neurons representing single concepts, meanings emerged across populations of cells, with similar words producing related activity patterns. Researchers then compared those patterns with mathematical embeddings created by language models. Neural responses correlated with semantic relationships captured by the models and aligned particularly closely with GPT-2 representations. Responses also changed according to context and words with multiple meanings. The findings provide a new window into how the hippocampus participates in understanding continuously changing natural language. (Medical Xpress)