Use this URL to cite or link to this record in EThOS:
Title: Testing the predictive coding account of temporal integration in the human visual system : a computational and behavioural study
Author: Aitken, Fraser
ISNI:       0000 0004 8502 3913
Awarding Body: University of St Andrews
Current Institution: University of St Andrews
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
A major goal of vision science is to understand how the visual system maintains behaviourally relevant perceptions given the level of uncertainty in the signals it receives. One proposed solution is that the visual system applies predictive coding to its inputs based on the integration of prior knowledge and current stimulus features. However, support for some vital aspects of predictive coding in the temporal domain is lacking and simpler accounts of temporal integration also exist. The aim of this thesis was to test two key attributes of predictive coding in time a) does the visual system apply adaptive weighting to prediction errors and b) can the visual system apply probabilistic information learnt from stimulus sequences when making predictions. In chapters 3 & 4, we tested predictive coding's ideas of how prediction errors are weighted under the theoretical guidance of a temporal integration model linked to predictive processing, called the Kalman filter. Here, both experiments supported predictive coding. We showed that, consistent with the Kalman filter, visual estimates and the way estimation errors were corrected, adapted to stimulus behaviour and viewing conditions. In chapter 5, we assessed the ability of the visual system to integrate conditional relationships present in sequences of stimuli when making predictions. To do this, we inserted a stimulus sequence that changed and omitted trials based on Markov transition probabilities that made some transitions more or less probable and assessed reaction times and omission trial responses. Reaction time data was consistent with predictive coding, in that more predictable changes elicited faster responses. Omission trials data, was though, less clear. When faced with no stimulus, participants did not apply the conditional probabilities in their decisions optimally, instead applying non optimal decision strategies, inconsistent with predictive coding. In summary, this thesis supports the predictive coding of temporal integration but questions its application in all situations.
Supervisor: Ales, Justin Michael Sponsor: Wellcome Trust
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral