| Author: | Chan, Ka Siu |
| Title: | Neural mechanisms underlying information sampling in multiple choice decision making |
| Advisors: | Chau, Bolton (RS) |
| Degree: | M.Phil. |
| Year: | 2023 |
| Subject: | Decision making Cognitive neuroscience Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Rehabilitation Sciences |
| Pages: | 95 pages : color illustrations |
| Language: | English |
| Abstract: | In a multifaceted environment, it is common to select among options with uncertain consequences or outcomes. Often the uncertainty can be reduced by sampling information, such as gathering information from options and searching for new options. In other words, there are at least two ways in which we can sample information. Existing literature has discussed the option selection and information sampling processes independently. However, there is a lack of a unified framework to illustrate the dynamics between them, while these three decision-making processes should be considered at the same time. In Chapter 1, I illustrate a daily life example and state the importance of investigating the neural mechanisms of information sampling in multiple-choice decision-making. To address the major problem in this thesis, I reviewed three candidate brain regions: the ventromedial prefrontal cortex (vmPFC), the anterior cingulate cortex (ACC), and the intraparietal sulcus (IPS). The vmPFC is related to both valuation and value-comparison processes, which are essential in selecting better options. The former process allows us to assign an internal value to an option, while the latter process allows us to make the comparison between options according to their values. The ACC is related to searching for new options from the environment, while the IPS is related to uncertainties and information gain. I then designed a multiple-choice decision-making task with three possible decisions (i.e., to accept a current option; to clarify a current option; to search for new options) to investigate both computational and neural mechanisms of information sampling in Chapter 2 & 3 respectively. Behavioural results are discussed in Chapter 2. During the designed decision-making task, while the value of existing options was greater, but the outcome was largely uncertain at the same time, clarify decisions were more preferred. While the value of searching was greater, search decisions were more preferred. These results served as a fundamental framework to illustrate the sampling of information from existing options and the environment involved different mechanisms. In Chapter 3 I report a functional magnetic resonance imaging (fMRI) experiment to examine the neural signals of the vmPFC, IPS and ACC. The results show that during multiple-choice decision-making, the vmPFC guides the selection of the best option. Critically, the IPS and ACC were found to be crucial during information sampling. The IPS signals the demand for information and information gain, which guide clarify decisions to reduce uncertainty, while the ACC signals the incentive of searching, which guides search decisions. After the vmPFC, IPS and ACC were found to be involved in information sampling during multiple-choice decision-making, how these three regions formulated final decision was still not clear. To address this problem, in Chapter 4 I employ a convolutional neural network (CNN), which is a deep learning technique that is trained and dependent on human data, without any prior assumptions (model-free) to predict human decisions. The CNN involves feature extraction, integration and decision-making processes, which are particularly important in the decision-making task. By conducting a series of representational similarity analyses (RSA), which is multivariate analysis that allows us to compare the representational similarities between the multimodal representations in the CNN and the multi-voxel activation patterns of the human brain, which helps us better understand the computational processes in our brains. I demonstrate that the IPS, ACC and vmPFC are related to the early, intermediate, and late stage of decision formation respectively. To conclude, I demonstrate that the IPS is critical during the sampling of information from options, the ACC is important during the sampling of information from the environment to discover alternatives, while the vmPFC operates as a guide to stop sampling information and accept existing options. The unified framework to illustrate the dynamics between option selection processes and information sampling processes was demonstrated. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12395