A Systematic Investigation of Feature-Based Visual Learning: Mechanisms and Application

Wang, Zhiyan

Full Metadata

Overview

Year:: 2021
Contributor:: Wang, Zhiyan (creator); Watanabe, Takeo (Advisor); Frank, Michael (Reader); Sasaki, Yuka (Reader); Brown University. Department of Cognitive, Linguistic, and Psychological Sciences (sponsor)
Genre:: theses
Subject:: Perceptual learning; Cognitive neuroscience
Extent:: 12, 96 p.

Files

Description

Abstract:: Visual perceptual learning is defined as visual performance improvement after visual perceptual experiences. It is controversial how changes in association with feature representation change, namely feature-based plasticity, occurred. In Aim 1, the dissertation investigated where do changes occur in association with feature-based plasticity. In Experiments 1 & 2, we used decoded fMRI neurofeedback to induce a pattern in V1/V2 similar to that of a Sekuler motion stimulus. The Sekuler motion consisted of both local motion directions and a global motion direction which was the summation of local motion signals. Previous psychophysical studies demonstrated that local motion directions corresponded to feature-based plasticity. Aim 1 demonstrated that after neurofeedback training in early visual areas (V1/V2), improvement was found among the local motion ranges, suggesting changes in early visual areas are associated with feature-based plasticity. In Aim 2, we performed three experiments and trained four groups of subjects: the Reward Before group, the Reward After group, the Arousal Before group, and the Arousal After group. Each group differed in the temporal order of how reward or arousal cues were paired with the trained orientation feature. Aim 2 demonstrated that reward and arousal have differential effects on feature-based plasticity. Moreover, fitting the normalization model of vision to the psychophysical data indicated an excitatory effect of arousal and an inhibitory effect of reward. In Aim 3, the dissertation demonstrated that feature-based learning of complex visual images such as faces involves feature representation changes. In Aim3, a group of BDD and healthy control subjects were trained on low spatial frequency faces to enhance holistic processing. The experiment showed dissociable blood-oxygen-level-dependent (BOLD) changes in body dysmorphic disorder patients and healthy control subjects consistent with their pre-training representation of holistic face processing. To conclude, our results demonstrated how feature-based visual learning occurred and can be potentially applied to clinical populations.
Notes:: Thesis (Ph. D.)--Brown University, 2021

Content

Citation

Wang, Zhiyan, "A Systematic Investigation of Feature-Based Visual Learning: Mechanisms and Application" (2021). Cognitive, Linguistic, and Psychological Sciences Theses and Dissertations. Brown Digital Repository. Brown University Library. https://repository.library.brown.edu/studio/item/bdr:c6n29vz8/

Relations

Collection:

Cognitive, Linguistic, and Psychological Sciences Theses and Dissertations

Theses and Dissertations for the Cognitive, Linguistic, and Psychological Sciences department.
...