Seeing is predicting: Testing a predictive coding account of visual perception involving saccadic eye movements

Projektinformationen

C-Pre

ID Finanzhilfevereinbarung: 846392

DOI

10.3030/846392

Projekt abgeschlossen

EK-Unterschriftsdatum 11 April 2019

Startdatum 1 Februar 2020

Enddatum 31 Januar 2022

Finanziert unter

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Gesamtkosten

€ 175 572,48

EU-Beitrag

€ 175 572,48

Koordiniert durch

STICHTING RADBOUD UNIVERSITEIT
Netherlands

Periodic Reporting for period 1 - C-Pre (Seeing is predicting: Testing a predictive coding account of visual perception involving saccadic eye movements)

Berichtszeitraum: 2020-02-01 bis 2022-01-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

We usually have the impression to see the world around us in clear detail. However, detailed vision is at any point in time limited to only a comparatively small area around the centre of the visual field, the part that corresponds to the fovea on the retina. In order to arrive at a clear impression of our complete surrounding, we scan the environment with quick and jerky, so-called saccadic, eye movements, of which we are usually not aware. These abrupt displacements of our gaze position partition the continuous stream of visual input into discrete periods during which gaze is comparatively stable, called fixations. These gaze dynamics imply that every time we perceive something foveally with high resolution we have previously seen the same thing with much worse resolution further in the visual periphery. There is always a so-called visual preview. Researchers have argued that peripheral visual preview information allows to predict upcoming foveal visual information and that this form of prediction could be understood under the currently influential frameworks of predictive processing, and more specifically, predictive coding. This has, however, not yet been tested empirically. The question addressed in the present project was, thus, whether predictive processing and predictive coding can actually explain trans-saccadic preview effects.
It is important for society to understand how we, as humans, perceive the world. The present project is a particular example for the more general idea that expectations affect perception. Understanding to what extent and how exactly expectations affect perception is of utmost importance for human interaction and mutual understanding. Regarding a more technical understanding of expectations, the present project tested the limits of a currently highly influential theory of brain function, i.e. predictive coding.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

The project was split into two objectives. First, the question was whether early stages of post-saccadic processing (preview effect) are affected by higher-level predictive mechanisms / expectations. A set of online and laboratory eye-tracking experiments was conducted to answer this question. It turned out that extremely predictive contexts indeed change post-saccadic processing, but the degree to which post-saccadic processing was affected depended how predictive the context was and did change over time. Second, the question was whether there are top-down prediction and bottom-up prediction error signals consistent with the idea of trans-saccadic predictions. This hypothesis was tested in a combined MEG and eye-tracking experiment that allowed to track brain dynamics time-locked to saccadic eye-movements. The most important lesson learnt from this study was that any top-down or bottom-up prediction-related signal has to be carefully distinguished from effects related to neural adaptation.
The results have been presented at international conferences and are to be published in peer-reviewed journals. The created knowledge will provide the basis for further, potentially clinically useful, applications related to typical and atypical perceptual processing in certain mental conditions, e.g. involving delusions or hallucinations, which have already been linked to alterations in predictive processing.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

From a conceptual perspective, this project progresses beyond the state of the art by delineating and operationally distinguishing the closely related and competing concepts of prediction and adaptation in the context of trans-saccadic perception. This careful conceptual distinction together with the methodologically unique combination of MEG recording, eye-tracking, and measures of information transfer allows to trace brain dynamics time-locked to saccadic eye-movements and thereby provides a way to determine the explanatory boundaries of one of the most influential theories of brain function, predictive coding.

Summary figure