Urgent.News

What's breaking now, across thousands of outlets.

AI

Local neural wiring may set the brain's range of activity patterns

The human brain contains billions of neurons, specialized nerve cells that receive, transmit and process information through electrical and chemical signals. Human thoughts, sensory perceptions and behaviors are known to emerge not from the activity of individual neurons but from collective patterns of activity distributed across different neuron populations.

Local neural wiring may set the brain's range of activity patterns

Neurons in the human brain communicate via electrical and chemical signals, giving rise to thoughts, perceptions, and behaviors through collective patterns of activity. Researchers measured the dimensionality, or number of independent activity patterns, of neural activity in the mouse visual cortex. Their findings, published in Nature Neuroscience, indicate that local neural wiring plays a significant role in shaping the range of activity patterns the brain can explore.

Using Neuropixels probes, the researchers analyzed high-density recordings to identify how synaptic motifs in recurrent networks control the dimensionality of neural activity across different brain areas. They discovered that these synaptic motifs, which are prevalent in both mouse and human brains, influence the number of distinct activity patterns that neurons can exhibit simultaneously.

This suggests that local circuitry in the cortex systematically controls the brain's degrees of freedom. The researchers also observed that cortical activity transitions among states with varying dimensionalities over time. Their work sheds light on the biological factors that control neural activity and could inform studies of brain disorders and the development of brain-inspired computational models.

Written by urgent.news from Medical Xpress's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at medicalxpress.com →

More in AI

Token Math for AI Coding: When a Free Server Beats Self-Hosting

The decision between a free hosted AI coding server and a self-hosted stack is rarely about price. It is about three measurable variables: token burn per task, latency tolerance, and privacy surface.

  • Focus on token consumption, latency tolerance, and privacy surface.
  • Use token budget script for initial estimate.
  • Compare projected token burn against free allowance.

The Data Boundary Problem: Using a Free Server Without Leaking Your Prompts

A free server is a data boundary decision, not a cost decision. Every prompt you send to a managed endpoint leaves your network.

  • Data boundary problem occurs with free AI servers, exposing sensitive info.
  • MonkeyCode's free server simplifies usage but doesn't solve data boundary issue.
  • Privacy gate uses local proxy to sanitize prompts and log requests.

Opinion: AI Patch Acceptance Is a Vanity Metric — Revert Rate Is the Truth

Every AI code review metric you track measures the hour before merge, and that is precisely the hour when the least information exists.

  • Pre-merge metrics like AI patch acceptance are misleading indicators of patch quality.
  • Revert rate accurately reflects a patch's performance after deployment.
  • Classifying reverts into categories guides appropriate response actions.

More from Friday 4 September →