Date of Award

2019

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Neuroscience

College

College of Graduate Studies

First Advisor

Thomas Naselaris

Second Advisor

Leonardo Bonilha

Third Advisor

Nathan Rowland

Fourth Advisor

Prakash Kara

Fifth Advisor

Thomas Jhou

Abstract

Cerebral processing of visual stimuli is characterized by a complex circuitry involved in processing visual inputs while simultaneously contextualizing, categorizing or modulating these inputs by higher-order cortical centers. The sensory input and the cognitive control over this input can potentially be differentiated spatially, i.e. different brain regions, or spectrally, i.e. different frequencies of neuronal activity. In this work, we used object-category based visual tasks to investigate the spectral and spatial patterns of encoding of visual inputs and cognitive control at the level of visual attention and mental imagery using local field potentials in human subjects across widely distributed recording sites and a broad frequency spectrum(1-100Hz). Using PCA, we demonstrate that during a task involving both visual attention to an object category with varied observed category a broadband, and two narrowband low-frequency explained the main variance in the data. When comparing response to attended versus seen categories, we did not observe a spatial difference in location of encoding sites, but using decoding models, the broadband signal decodes vision better than attention in visual cortex and vision and attention equally in the temporal lobe. However, narrowband delta-theta decodes the best for both vision and attention, and alpha-beta differentially decodes for attention better than vision. Using an alternate task that involves image memorization, imagery, and passive viewing, we demonstrate that the main power spectral modulation among those mental states are represented by a broadband, gamma band, and low-frequency band. Encoding sites were similarly spatially widely distributed. Decoding models showed that gamma band predicts attentive viewing and mental imagery with best accuracy. Both broadband and low frequency band accurately decode for passive and attentive viewing. Our findings demonstrate that encoding of vision, attention and mental imagery is not dependent on a single spectral domain and optimal decoding of visual processes should consider the co-contribution of narrowband and broadband spectral patterns to account for the different co-occurring top-down and bottom-up processes. Therefore, although gamma-significantly studied in vision-is involved in visual working-memory tasks, both broadband and low-frequency narrowband patterns of neuronal activity co-participate in visual processing in the context of object-based visual attention and mental imagery.

Rights

All rights reserved. Copyright is held by the author.

Share

COinS