Articles tagged with Neural Modeling
A recent study published in eNeuro resolves a long-standing issue in decision-making by showing how the brain optimizes speed and accuracy. Researchers found that the brain adjusts signal-to-noise ratio to balance speed and accuracy, shedding light on the neural mechanisms underlying this crucial aspect of human behavior.
A new study published in PNAS found that the brain's ventromedial prefrontal cortex plays a key role in forming expectations underlying effort-based choices. The study used fMRI to model neural computations for effort and reward, revealing a clear role for three brain regions: the vmPFC, dACC, and aI.
The Adversarial Threshold Neural Computer (ATNC) model, a proof-of-concept, combines Generative Adversarial Networks (GANs) with Reinforcement Learning (RL) to generate novel small organic molecules. The GAN-RL architecture demonstrated the ability to produce valid and unique molecular structures, paving the way for future drug discovery.
Researchers are using artificial networks to enhance their understanding of the human brain's complex systems. By parsing out contextual clues in image recognition, they can gain insights into how humans perceive and process information.
Four NYU faculty members have been awarded Sloan Research Fellowships for their outstanding contributions to science and engineering. Miranda Holmes-Cerfon develops mathematical tools to study mesoscale materials, while Tianning Diao works on sustainable synthetic methods for pharmaceutical products.
Machine learning scientists at Disney Research developed a dynamic word embeddings model that uncovers how the meanings of words change over time. The model, which integrates neural networks and statistics used in rocket control systems, detects semantic change throughout history by analyzing semantic vector spaces.
A team of researchers from the University of Texas at Austin has developed novel approaches to information retrieval that leverage artificial intelligence, crowdsourcing, and supercomputing. Their method combines input from multiple annotators to determine the best overall annotation for a given text, improving accuracy in extracting d...
Researchers at Columbia University School of Engineering and Applied Science have made a breakthrough in auditory attention decoding (AAD) methods, bringing cognitively controlled hearing aids closer to reality. The new system can automatically separate individual speakers from a mixture, determine which speaker the user is listening t...
Researchers have developed a method for designing energy-efficient neural networks, reducing power consumption by up to 73% compared to standard implementations. The new approach uses an analytic tool to evaluate and prune low-weight connections, resulting in more efficient networks with fewer connections.
A new computational model reveals a winner-take-all operation is enabled by a configuration of inhibitory neurons. The model, developed by MIT researchers, uses theoretical computer science to prove that a specific arrangement of inhibitory neurons is necessary for the operation.
This review article discusses the neural control mechanisms behind fly larval locomotion, a complex motor behavior shared by both vertebrate and invertebrate species. Key findings from Drosophila models reveal the crucial role of genes in regulating locomotor rhythm and pattern generation.
A new model of brain dynamics, extending balanced network theory, provides deep predictions linking brain circuits to activity. The model accurately explains experimental findings on neuronal variability in living animals, enabling potential discovery of neural signatures associated with learning or disease.
A new mathematical model has successfully reproduced complex brain activity during deep sleep, revealing its role in memory consolidation. The thalamocortical coupling is crucial for this process, according to the researchers.
A team of researchers has developed an open software platform to simulate a complete model of the adult fruit fly brain for the first time. The platform will enable global collaboration and accelerate model development, providing insights into diseases such as Alzheimer's and motor neurone disease.
A team of scientists developed a new system using stem cells to model features of the developing human brain that could be targeted by toxic chemicals or drugs. The approach, described in PNAS, uses machine learning to build a predictive model from RNA sequencing data collected from neural tissue constructs exposed to different chemicals.
A study published in PLOS Computational Biology found that an artificial deep neural network performs as well as the primate brain at object recognition. This achievement could pave the way for significant advancements in artificial intelligence and our understanding of primate visual processing.
A recent study found that prolonged electrical stimulation does not cause damage to the sacral nerve roots of rabbits. The research suggests that neurons in the rabbit sacral nerve roots can tolerate long-term electrical stimulation, which could have implications for the treatment of spinal cord injuries.
Researchers at Princeton University suggest that the brain breaks down experiences into events using subconscious mental categories. These categories are formed based on temporal relationships between people, objects, and actions.
Pitt researchers find neurons wire and fire together in clusters to dominate brain activity, with weak stimuli triggering imbalance. The model shows how clustered connections influence stochastic brain dynamics, shedding light on the brain's default spontaneous state.
A new X-ray scan has provided the highest resolution and most accurate image of an early human ancestor's brain case. The study challenges previous models of brain evolution by revealing a surprisingly small brain volume and a mix of characteristics that resemble both humans and chimpanzees.
A computational model developed by RUB scientists simulates brain activity through neighbor interactions, facilitating faster processing of moving objects. The model, published in PLoS Computational Biology, uses a neural field approach to describe the complex dynamics of cortical neurons.
Researchers at Caltech identified two learning strategies associated with unique neural signatures, model-free and model-based learning. Model-free learning relies on trial-and-error comparisons, while model-based learning generates cognitive maps of environments, enabling adaptability in novel situations.
Researchers developed a computer program that replicates complex seashell patterns using simple principles from brain function. The 'neural net' model shows how mollusks sense and build their shells based on previous layers, much like the human brain projects memories.
A study published in PLOS ONE demonstrates that the spontaneous activity of small neuronal networks in the cortex consists of highly structured patterns rather than random noise. These patterns are shaped by network connectivity and can be used to inform researchers about the underlying anatomy.
Researchers found that language comprehension involves continuous processing and competition between representations, contradicting traditional discrete models. The study used mouse movements to track students' processing of words, revealing curved trajectories indicating gradual ambiguity resolution.
Scientists have identified a positive feedback loop in rat whiskers that helps gather sensory information, potentially leading to new treatments for blepharospasm. The discovery could also deepen understanding of the human eyeblink circuit and its connection to neural circuits.
Researchers found a speed limit to neural network synchronization, set by network connectivity. The analysis revealed that even strong interactions cannot achieve faster synchronization than an upper limit. This could severely limit the speed of information processing in the brain.
Research suggests that neural connections in the brain can be changed in response to environmental cues, supporting the idea that intelligence is not fixed. This model provides a new understanding of human intelligence and has implications for education and learning outcomes.
Scientists create detailed computer simulations of brain neurons and their assemblies to understand how the brain computes. They're building VLSI chips that accurately model retinas and can produce output spikes matching real retinae, with potential applications in retinal implants.
A new computer model has shown early success in predicting toxic blue-green algae blooms up to four weeks before they occur. The model uses artificial neural networks to learn key factors contributing to algal outbreaks, enabling accurate forecasts and better understanding of the factors involved.
Researchers found that blocking the interaction between two naturally occurring molecules restricts tumor growth and spread in mice. Inhibition of RAGE-amphoterin interaction decreased tumor cell growth, migration, and invasion without affecting angiogenesis.
Computational biologist Sorin Istrail shows that the solution of Ising's model cannot be extended into three dimensions for any lattice, making exact solutions unattainable. The model, developed by Ernst Ising in 1926, has been extensively studied in one and two dimensions but remains unsolvable in three dimensions.
A Cornell psychologist has challenged the idea of universal music experiences across cultures by finding a Western bias in the explanation. Research suggests that expectations play a crucial role in our emotional response to music, and that this may be specific to Western musical backgrounds. The study's findings have implications for ...
KSU civil engineer Steven K. Starrett developed the KTURF program to predict pesticide and nitrogen leaching, which is within 4% of testing case values. The tool helps golf course turf managers screen chemicals and reduce groundwater contamination.
Recent progress in Over-The-Horizon (OTH) radar technology is reviewed, including advancements in detection and tracking systems. Key findings include improved capabilities in clutter-dominated environments and the development of advanced algorithms to mitigate distortion mechanisms.
Scientists at Boston University report a new model of ventilator assisted lung function that incorporates 'noise' to improve gas exchange in patients with lung injury. The approach may also minimize additional trauma by enhancing the avalanche-like opening of airways and alveoli during inhalation.
VA-10,367 has been shown to enhance the recovery of function following spinal cord injury through effects on neuronal regeneration. In a peripheral nerve injury model, the compound accelerated the days to onset of foot movement and walking, leading to recovery of function.
Researchers at Johns Hopkins University use computer models to study serious cardiac disorders and test potential treatments. By mimicking the heart's physiological functions in a highly detailed model, they aim to find medicines that can prolong the lives of millions suffering from congestive heart failure.
The study reveals that chemical activity within brain cells exhibits behavior characteristic of self-organized criticality, which may help identify disease processes. Researchers used modeling techniques to simulate brain networks and found a power law distribution of wave sizes and durations, adhering to the signature of self-organize...
Scientists created a 3D computer model of a glutamate receptor, which can be used to develop new drugs for Alzheimer's disease and other disorders. The model accurately represents the receptor's structure and behavior.
A computer model developed by researchers at WashU Medicine suggests that one-time carotid artery screening is beneficial for asymptomatic men with high-risk factors, preventing strokes at a relatively low cost. However, annual screening would be prohibitively expensive and may even increase stroke risk in low-risk individuals.