Moderating effects of self-perceived knowledge in a relevance assessment task: An EEG study

Abstract

Relevance assessment, a crucial Human-computer Information Retrieval (HCIR) aspect, denotes how well retrieved information meets the user's information need (IN). Recently, user-centred research benefited from the employment of brain imaging, which contributed to our understanding of relevance assessment and associated cognitive processes. However, the effect of contextual aspects, such as the searcher's self-perceived knowledge (SPK) on relevance assessment and its underlying neurocognitive processes, has not been studied. This work investigates the impact of users' SPK about a topic (i.e. ‘knowledgeable’ vs. ‘not knowledgeable’) on relevance assessments (i.e. ‘relevant’ vs. ‘non-relevant’). To do so, using electroencephalography (EEG), we measured the neural activity of twenty-five participants while they provided relevance assessments during the Question and Answering (Q/A) Task. In the analysis, we considered the effects of SPK and specifically how it modulates the brain activity underpinning relevance judgements. Data-driven analysis revealed significant differences in cortical electrical activity modulated by searchers' SPK in the context of relevance assessment, suggesting that SPK affects cognitive processes associated with attention, semantic integration and categorisation, memory, and decision formation that underpin relevance assessment formation. Our findings are an important step toward a better understanding of the role users' SPK plays during relevance assessment.

Keywords

• Human-computer information retrieval (HCIR)
• Relevance assessment
• Binary relevance
• EEG
• ERPs
• Self-perceived knowledge