Author ORCID Identifier

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Neuroscience Institute

First Advisor

Jordan Hamm, PhD

Second Advisor

Gennady Cymbalyuk, PhD

Third Advisor

Anne Murphy, PhD

Fourth Advisor

Marise Parent, PhD

Fifth Advisor

Aras Petrulis, PhD


The predictive coding theory of sensory processing posits that the brain generates and consistently updates models about expectations of future sensory input based on previously encountered stimuli. This theory argues that the brain is organized hierarchically, allowing each level to generate predictions about likely sensory information (to be received from lower levels) and to compare incoming sensory information against those predictions. By suppressing responses to sensory data that fits its own predictions, hierarchical brain networks effectively minimize the amount of incoming sensory information that is processed at any given time. This theory is best supported by imaging, electrophysiology, and electroencephalography studies of visual and auditory information in isolation. However, the sensory world is rarely experienced naturally unimodally, and certain corollaries of predictive coding have yet to be tested as they apply to multisensory integration and information processing. The overarching hypothesis of this dissertation is that the mammalian brain responses support a theory of predictive coding mechanisms at multiple levels of sensory processing such that both unisensory and audiovisual multisensory information are processed within the context in which the information is presented. We used two-photon calcium imaging and local field potential recordings in adult mice to identify individual neuronal and regional responses, respectively, in brain regions associated with unisensory (primary visual cortex, V1) and multisensory (parietal associative area, PTLp) information processing. Using a novel multisensory oddball paradigm, we uncovered populations of pyramidal and somatostatin neurons in PTLp that detect contextually deviant multisensory stimuli. Further, we propose that, because schizophrenia (SZ) is a fundamentally a disorder of information integration, predictive processing of multisensory information is disrupted by genetic and molecular risk factors for the disease. We tested this by regional responses to contextual deviance in animals expressing a SZ-relevant missense mutation (Kalirin P2255T). We identify altered context processing in this model, in the form of dampened regional responses to multisensory deviance. Together, these results provide evidence to support the theory of predictive coding and hierarchical sensory information processing in the brain. Finally, we provide further evidence to support the impact of the Kalrn P2255T mutation on SZ-relevant cognitive functioning.


File Upload Confirmation


Available for download on Friday, December 06, 2024