Latest

Deep Learning

$This shows two heads.$

·July 3, 2026·10 min read

LLMs and Math Combine to Map Human Decision-Making

A new research framework that combines large language models (LLMs) with choice mathematics to evaluate human decision-making. By deploying LLMs to automatically interpret and code thousands of free-text participant thought justifications, the framework provides a scalable, validated methodology demonstrating that human reasoning strategies shift dynamically with a problem's structure.

This shows a hand holding a book and a robotic hand.

Featured Neuroscience

·July 2, 2026·9 min read

AI Writing Strips Mystery and Complexity From Stories

A new study introducing the CASPER framework reveals that AI writing models consistently strip away narrative mystery, relying on safe archetypes and tidy resolutions, proving that scaling up model sizes fails to improve character depth or capture genuine human contradiction.

This shows a robot hand and a human hand holding a rose.

Featured Neuroscience Psychology

·July 1, 2026·12 min read

How AI Substitutes Are Replacing Genuine Human Connection

A new study warns that widespread, unmoderated use of AI chatbots among teenagers introduces severe risks of relational displacement and maladaptive relational learning, distorting social development by replacing real-world conflict resolution with frictionless, algorithmic validation.

Featured Neurology Neuroscience

·June 16, 2026·10 min read

Retinal Photos and AI Predict Early Alzheimer’s Risk

Routine, low-cost retinal photographs can accurately predict primary biological and lifestyle risk factors for Alzheimer’s disease.

This shows a head and a selection of colorful blocks.

Featured Neuroscience Psychology

·June 2, 2026·8 min read

Stroop Test Exposes Inherent LLM Flaw

Can an advanced artificial intelligence truly exercise decision-making control, or is it merely trapped in an inescapable loop of automatic pattern mimicry? A cognitive science study utilizes the classic psychological "Stroop task" to expose a fundamental limitation in large language model (LLM) attention mechanisms.

Featured Neuroscience

·May 27, 2026·10 min read

Mapping the Glymphatic System with AI

Can we prevent neurodegenerative decline by auditing the brain’s internal cleansing system during deep sleep? A collaborative study leverages physics-informed artificial intelligence to track the fluid flow velocity of the glymphatic system from MRI data.

This shows a person, a digital brain and speech bubbles.

Featured Neuroscience

·May 27, 2026·10 min read

AI Proves Language Evolves for Learnability

A new study using AI proves that language actively evolves over generations to become structured because it makes learning easier.

This shows a depressed person holding their head.

Featured Neuroscience Psychology

·May 21, 2026·10 min read

Machine Learning Doubles Depression Remission Rate

Can an algorithmic approach to lifestyle tracking radically improve depression treatment? A milestone clinical trial demonstrates that personalizing behavioral care via machine learning can nearly double standard therapy benchmarks.

This shows two computer generated brains.

Featured Neuroscience

·May 19, 2026·12 min read

Advanced AI Passes the Turing Test for the First Time

Has the boundary between human consciousness and artificial imitation officially dissolved? A landmark study provides the first empirical proof of an AI system passing the Turing test.

Featured Neurology Neuroscience Pain

·May 19, 2026·11 min read

Tracking Facial Micromovements to Measure Pain

A new study demonstrates that tracking high-speed facial micromovement spikes provides an objective, quantifiable metric for human discomfort.

This shows a prism, math equations and a neural network.

Featured Neuroscience

·May 8, 2026·11 min read

Why Massive AI Models Actually Generalize Better

Physicists have developed a mathematical "toy model" using statistical physics to explain one of the great mysteries of deep learning: why massive neural networks learn patterns instead of just memorizing data. By applying renormalization theory, the team has shown how high-dimensional fluctuations stabilize learning, paving the way for more efficient and predictable artificial intelligence.