Information theoretic perspectives on en- and decoding in audition and vision

Daube, Christoph (2022) Information theoretic perspectives on en- and decoding in audition and vision. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2022DaubePhD.pdf] PDF
Download (34MB)

Abstract

In cognitive neuroscience, encoding and decoding models mathematically relate stimuli in the outside world to neuronal or behavioural responses. While both stimuli and responses can be multidimensional variables, these models are on their own limited to bivariate descriptions of correspondences. In order to assess the cognitive or neuroscientific significance of such correspondences, a key challenge is to set them in relation to other variables. This thesis uses information theory to contextualise encoding and decoding models in example cases of audition and vision. In the first example, encoding models based on a certain operationalisation of the stimulus are relativised by models based on other operationalisations of the same stimulus material that are conceptually simpler and shown to predict the same neuronal response variance. This highlights the ambiguity inherent in an individual model. In the second example, a methodological contribution is made to the problem of relating the bivariate dependency of stimuli and responses to the history of response components with high degrees of predictability. This perspective demonstrates that only a subset of all stimulus-correlated response variance can be expected to be genuinely caused by the stimulus, while another subset is the consequence of the response’s own dynamics. In the third and final example, complex models are used to predict behavioural responses. Their predictions are grounded in experimentally controlled stimulus variance, such that interpretations of what the models predicted responses with are facilitated. Together, these three perspectives underscore the need to go beyond bivariate descriptions of correspondences in order to understand the process of perception.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Subjects: B Philosophy. Psychology. Religion > BF Psychology
Colleges/Schools: College of Science and Engineering > School of Psychology
Supervisor's Name: Gross, Professor Joachim and Ince, Dr. Robin
Date of Award: 2022
Depositing User: Theses Team
Unique ID: glathesis:2022-82893
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 20 May 2022 09:16
Last Modified: 24 May 2022 11:15
Thesis DOI: 10.5525/gla.thesis.82893
URI: https://theses.gla.ac.uk/id/eprint/82893

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year