Gain Modulation and Perceptual Switches

For many years now James Danckert and I have been interested in the notion of a mental model, by which we mean, loosely, your model of the world and how it works. In some cases this may be an explicitly statistical model (like in our Rock, Paper, Scissors work), but it doesn't have to be. Somehow we also make inferences about things that are not statistical, e.g. whether to go to graduate school or what country, of the ones we have never visited before, would we most enjoy on our next vacation.

Assuming such mental models, exist two obvious questions are how to do they come to be and what are the circumstances that lead to their revision, which we have called "updating." To get away from the purely statistical character of our earlier work we partnered with Elisabeth Stöttinger on the design of a set of visual stimuli which morphed from one image to another and passed through an ambiguous point.

link to paper describing stimuli

A population of people will show fairly consistent points of transition from image 1 to image 2, and after having established this consistency we went looking for physiological correlates with both fMRI and pupillometry. Through a sabbatical connection James Danckert made with Mac Shine of the University of Sydney we provided our imaging, pupil, and behavioral data so that Shine and his lab could apply some of their network methods. The result is a manuscript under review, but now available as a publically accessible preprint. The principal conclusion is that there seem to be temporally linked changes in networks of neuromodulatory transmitters, like noradrenalin, that are linked to the physiological changes (like pupil dilation) and transitions in the object a participant in our experiments reports. The paper concludes that these changes are also accompanied by changes in the energy landscape and dimensionality of brain state dynamics.

If you read the paper, please let us know what you think.