All articles tagged with #computational model
Originally Published 1 year ago — by Neuroscience News
A Yale study found that the brain is more likely to remember images that are difficult to interpret, using a computational model to show that scenes hard to reconstruct are more memorable. This insight could help develop more efficient AI memory systems.
Originally Published 2 years ago — by Medical Xpress
A new study by researchers at the Paris Brain Institute explores how individual preferences influence the speed and creativity of idea generation. Using a behavioral study and computational modeling, the researchers found that the subjective evaluation of ideas plays a crucial role in creativity. The study also revealed that individuals inclined towards original ideas tend to suggest more inventive concepts. The findings challenge the notion that creative thinking is a mysterious process and suggest the possibility of precisely describing the mechanisms of creativity at a neurocomputational level. The researchers aim to further investigate different creativity profiles and explore how environments can foster or inhibit creativity.
Originally Published 2 years ago — by ScienceAlert
A new study by researchers from the University of Exeter and the University of Sussex suggests that certain optical illusions can be explained by limitations in our visual neurons rather than higher-level processing in the brain. The study found that our visual neurons have a finite bandwidth, which affects our perception of patterns on different scales. The researchers developed a computational model that shows how this limitation in processing and metabolic energy forces neurons to compress visual data, resulting in visual illusions. The findings challenge long-held assumptions about how visual illusions work and could help explain why we perceive contrasts in modern televisions with built-in HDR.
Originally Published 2 years ago — by Medscape
Intense exercise may increase the risk of stroke in patients with moderate to severe stenosis, according to a new study. The research used a computational model to simulate blood flow in the internal carotid artery and found that intense exercise increased arterial wall shear stress at the stenosis zone, potentially leading to plaque rupture and ischemic stroke. The study suggests that exercise regimens should be carefully prescribed for individuals with stenosis or a history of strokes, and regular monitoring of arterial health is recommended for those engaging in intense workouts.
Originally Published 2 years ago — by SciTechDaily
MIT neuroscientists have developed a computational model that predicts human emotions in social scenarios, outperforming other emotion prediction models. The model considers individuals’ desires, expectations, and the influence of observers, deducing motivations, comparing outcomes with expectations, and predicting emotions based on these factors. The researchers used scenarios from a British game show called “Golden Balls” to train the model, which performed much better than any previous model of emotion prediction. The researchers aim to adapt the model for broader applications.
Originally Published 2 years ago — by Neuroscience News
MIT neuroscientists have developed a computational model that can predict human emotions, including joy, gratitude, confusion, regret, and embarrassment, based on the prisoner’s dilemma game theory. The model uses factors such as a person’s desires, expectations, and whether their actions are being observed to predict emotional responses. The researchers used scenarios from a British game show called “Golden Balls” to train the model, which outperformed previous models in predicting emotions, signaling significant progress in emotional artificial intelligence.