Antonio Galves and Kádmo de S. Laxa
We introduce a new model for a highly polarized social network. This model is a system of interacting marked point processes with memory of variable length. Each point process indicates the successive times in which a social actor express a "favorable" or "contrary" opinion on a certain subject. The orientation and the rate at which an actor express an opinion is influenced by the social pressure exerted on them, modulated by a polarization coefficient. The social pressure on an actor is reset to 0, when they express an opinion, and simultaneously the social pressures on all the other actors change by one unit in the direction of the opinion that was just expressed. We prove that when the polarization coefficient diverges, this social network reaches instantaneous consensus. Here by consensus we mean the set of lists in which all the social pressures push in the same direction. This consensus has a metastable behavior. This means that the direction of the social pressures on the actors globally changes after a long and unpredictable random time.
"Simulation-based inference for neural network structure: a proposal" is the first seminar in the series Pathways to the 2023 IHP thematic project Random Processes in the Brain.
Área de conhecimento: Ciência da Computação
News
| Jul 21, 2021
NeuroMat's study on "Retrieving the structure of probabilistic sequences of auditory stimuli from EEG data" was featured on the last issue of science communication magazine Pesquisa FAPESP. The piece is called "The Statistician Brain" as NeuroMat's paper relates to a classical idea by Hermann von Helmholtz that the brain acts like a statistician as it models stimuli from the environment.
Predicting upcoming sensorimotor events means creating forward estimates of the body and the surrounding world. This ability is a fundamental aspect of skilled motor behavior and requires an accurate and constantly updated representation of the body and the environment. To test whether these prediction mechanisms could be affected by a peripheral injury, we employed an action observation and electroencephalogram (EEG) paradigm to assess the occurrence of prediction markers in anticipation of observed sensorimotor events in healthy and brachial plexus injury (BPI) participants. Nine healthy subjects and six BPI patients watched a series of video clips showing an actor’s hand and a colored ball in an egocentric perspective. The color of the ball indicated whether the hand would grasp it (hand movement), or the ball would roll toward the hand and touch it (ball movement), or no event would occur (no movement). In healthy participants, we expected to find distinct electroencephalographic activation patterns (EEG signatures) specific to the prediction of the occurrence of each of these situations. Cluster analysis from EEG signals recorded from electrodes placed over the sensorimotor cortex of control participants showed that predicting either an upcoming hand movement or the occurrence of a tactile event yielded specific neural signatures. In BPI participants, the EEG signals from the sensorimotor cortex contralateral to the dominant hand in the hand movement condition were different compared to the other conditions. Furthermore, there were no differences between ball movement and no movement conditions in the sensorimotor cortex contralateral to the dominant hand, suggesting that BPI blurred specifically the ability to predict upcoming tactile events for the dominant hand. These results highlight the role of the sensorimotor cortex in creating estimates of both actions and tactile interactions in the space around the body and suggest plastic effects on prediction coding following peripheral sensorimotor loss.
