Thesis
Exploring the neurocognitive spectrum of users' information needs in the information retrieval domain
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2022
- Thesis identifier
- T16412
- Person Identifier (Local)
- 201885463
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- A body of NeuraSearch literature, comprised of interdisciplinary user-based investigations between Information Retrieval (IR) and Neuroscience, is growing. This research's novel perspective is driven by the increasing accessibility and applicability of neuroimaging techniques to objectively capture and understand the neurocognitive manifestations subserving the user's information search behaviours. One such is the complex concept of Information Need (IN). Given its fundamental role as a mental perception of the user's information anomaly that the user needs to resolve and the user's trigger to engage in search, it is timely to deepen the understanding of the origin of INs. The NeuraSearch perspective comes into place to explore the cognitive mechanisms behind the realisation of INs, which would be using the traditional techniques of user-based IR research not viable. This thesis, in particular, explores the user's cognitive context utilising the paradigms of cognitive mechanisms to indicate different states of knowledge, analyses types of knowledge anomalies and discusses their implications on user search behaviour and expectations. Our enquiry is inspired by the theoretical IR concept of various user levels of knowledge supporting different variants of anomalies (i.e., insufficiencies) in the knowledge. Moreover, the functional framework of knowledge, memory and cognition synergies the user's informativeness about their current capabilities and, thus, shows potential in addressing the variants of INs in the IR context. In summary, this thesis presents a series of investigations sourcing from a lab-based study modelled as an interactive textual Question-Answering scenario. The textual stimuli were controlled to evoke the information processing subserving the user's cognitive memory search. We acquired two categories of data from twenty-four (24) participants: 1) the behavioural data that describe subjects' interactions with the system and their responses and 2) the simultaneously acquired cortical activity of the same subjects using the Electroencephalography (EEG) technique associated with the outcomes of behavioural responses. We aimed at two factors of neurocognitive memory mechanisms which offer different perspectives on the knowledge capabilities of the user. First, we approached the metamemory, which refers to the user's introspective epistemic feelings with the prospective (future) quality of knowing. Second, we approached the associated concept of memory retrieval to emphasise the user's ability to retrieve factual knowledge involving a more in-depth memory search. In addition, we accounted for a modality of memories in interaction with the variability of the user's subjective confidence, potentially impacting a further decomposition of memories and indicating an affiliation to INs. We also constructed a data-driven analytical framework and conducted a quantitative analysis of Event Related Potentials (ERP) components, namely N1, P2, N400, and P6. These were found to be activated over the timeline of information processing to obtain significant spatio-temporal differences between the levels of these factors and their interactivity. In the light of the current understanding of the user's cognitive context in IR, this research aims to increase the informativeness about the variability of the perceived states of knowledge regarding INs by providing the evidence about their associated significant neurophysiological signatures and detectability in the brain. In addition, we used the links of ERP with cognitive operations to construct the qualitative models subserving the significant differences between particular levels of these factors. Overall, the contribution of the research presented in the current thesis is twofold. Firstly, it provides a multidisciplinary view on the metamnemonic and mnemonic drivers of INs, not addressed to date, by developing a comprehensive framework that can advance the understanding of the behaviours linked to INs and the associated neural mechanisms. Secondly, it contributes towards the evaluation of objective user signals and reflects on their applicability in the domain of IR, for example proactive IR systems.
- Advisor / supervisor
- Rodriguez, Mario Parra
- Moshfeghi, Yashar
- Resource Type
- DOI
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T16412 | 2023-02-28 | Public | Download |