Just out: Modelling sensory attenuation as Bayesian Causal Inference across two datasets.
If you are interested in cool tactile/ proprioceptive experiments and Bayesian graphical networks, please check it out!
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0317924
If you are interested in cool tactile/ proprioceptive experiments and Bayesian graphical networks, please check it out!
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0317924
Comments
Exp. 1: A vibratory probe is presented to the finger right before stroking over a structured surface. When probe + surface vibration match in frequency, ppl are less likely to perceive it - i.e. sensory attenuation.
Main idea: if we caused it ourselves (internal cause), we process it less (compared to externally caused inputs).
Qualitatively, our model reproduces psychometric functions and behavior.
Quantitatively, it outperforms competitor models such as forced fusion, forced integration, & individual bias only.
What are the neural underpinnings of model-derived markers of causal inference (e.g. posteriors, surprisal)?