A single neuron rivals an AI network, and ovaries don't retire

A mouse study published in Nature Communications found that right after birth, hippocampal networks are densely wired and hyperconnected, with neurons that fire from a single incoming signal [9]. That is the opposite of how mature circuits work, where neurons typically require multiple inputs before they fire. As the brain develops, those haphazard connections are aggressively pruned into something sparser and more precise. The pruning starts soon after birth and continues into adolescence. Study co-author Peter Jonas, a neuroscientist at the Institute of Science and Technology Austria, was surprised by both the timing and the force of those early connections: "You might think that early in development, you have poor synapses and weak synapses, but we found the opposite." The researchers call this the tabula plena, a full slate, rather than the tabula rasa, the blank slate neuroscience had long assumed.

The theory fits primates, dolphins, wolves, and bats well. It does not fit octopuses at all [78]. Cephalopods, which include octopuses, squids, and cuttlefish, have large brains relative to their bodies and display genuine behavioral sophistication. Most are also solitary, often cannibalistic, and die shortly after reproducing, with essentially no parenting or long-term social structure. A new paper in iScience, co-authored by anthropologist Kiran Basava and economist Michael Muthukrishna of the London School of Economics, proposes instead that brains evolved to store and process information acquired through learning, whether social or solo. The cultural brain hypothesis holds that what matters is not the size of your group but the demands placed on you to acquire, store, and use learned information. Octopuses, which adapt rapidly through individual trial and error, are the test case that social-brain theory cannot explain and this alternative can.

A study from Hebrew University upends that view at the cellular level [95]. Researchers developed a method in which an artificial neural network is tasked with perfectly replicating the input-output behavior of a single human cortical neuron. The more computational layers the AI needs to mimic the biological cell, the more complex that cell is. For a human cortical neuron, the answer came back startling: the equivalent of an entire multi-layered deep neural network. Other mammals' neurons were measurably simpler by the same test. The physical explanation points to the dendritic tree, the branching arms through which neurons receive incoming signals. Human cortical dendrites are denser, more richly branched, and electrically distinct in ways that let individual cells perform operations previously assigned to whole circuits. In direct tests, a single human cortical neuron could discriminate between two different complex images independently, a feat the team verified experimentally [95].

The compliance rate surprised almost everyone: 83 percent of participants maintained ketosis through the one-month primary phase, rising to 94 percent in a voluntary four-month extension, with no severe side effects reported. When researchers statistically controlled for weight loss, the psychiatric improvements held, which matters because it means the diet's effects on the brain are not simply a byproduct of losing weight. Participants who completed the four-month extension showed sustained reductions in clinical depression scores and core schizophrenia symptoms, alongside measurable gains in daily cognitive performance. Higher blood ketone levels correlated directly with sharper drops in blood glucose and in depression. The working hypothesis is metabolic: antipsychotic medications often disrupt the brain's ability to use glucose, starving neural circuits. Ketones, produced when the body runs on fat instead of carbohydrates, offer a different fuel source that bypasses those disrupted pathways and may restore energy to circuits standard medications cannot reach.

Researchers tracked ovarian tissue from reproductive youth through post-reproductive old age and found that once the follicle reserve was exhausted, the organ did not switch off. Its molecular identity changed entirely. Immune cells, including T cells and macrophages, flooded the tissue. Giant multinucleated cells appeared. Transcriptomic analysis showed the organ's gene expression looking less like a reproductive gland and more like an active inflammatory immune tissue, one producing signals that travel beyond the ovary itself. The researchers wrote that post-reproductive ovaries "completely lost their reproductive cellular identity and instead transitioned to an immune-like inflammatory organ." Those circulating signals may contribute to the low-grade systemic inflammation that is increasingly linked to aging and age-related disease across multiple organ systems.

No one had ever recorded it happening in real time until April 2024 [80]. Marine geophysicist Jean-Yves Royer and colleagues from the French National Center of Scientific Research spent years constructing an underwater observatory at the Southeast Indian Ridge, between Australia and Antarctica, with five hydrophones positioned specifically to catch one of these events. In late April 2024, the ridge split open. In less than two hours, an estimated 150 million cubic meters of magma tore through the crust as vast sheet-like intrusions called dikes, triggering earthquakes, reawakening dormant faults, and draining the magma reservoir below. The seafloor above it collapsed. The team measured several meters of displacement in both directions. Royer called it "a once-in-forty-year event." The results appear in Nature.