Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Methodological Insights into Declarative Memory Assessment: Investigating the Impact of Familiarity and Visual Complexity in Picture Recognition Tests
Poster Session A - Sandbox Series, Thursday, October 24, 10:00 - 11:30 am, Great Hall 3 and 4
This poster is part of the Sandbox Series.
Jiayi Lu1, Jueyao Lin1, Xiaocong Chen1, Lu Li1, Fumo Huang1, Caicai Zhang1; 1The Hong Kong Polytechnic University
INTRODUCTION: Declarative memory, crucial in language-related studies due to its link with vocabulary abilities (Hamrick et al., 2018), involves storing and recalling facts and events. Recognition tests, commonly used to evaluate this memory, present participants with a mix of previously studied (target) and new (distractor) items, asking them to identify the studied items after a delay. This method assesses declarative memory using signal detection theory and evaluates its retention by testing at different delays. Both verbal (e.g., words) and non-verbal (e.g., pictures) stimuli can be used, with pictures preferred to minimize verbal influences (Ullman, 2020). Controlling confounding factors is essential for valid test results. In picture-based recognition tests, familiarity and visual complexity are key factors, with familiarity negatively and complexity positively correlated with memory performance according to a previous study (Brodeur et al., 2017). The effects of these two variables on memory performance across different contexts (e.g., target vs. distractor pictures, short vs. long delays) and their potential interactions remain unclear. This research aims to clarify these questions, in order to enhance declarative memory assessment and inform future methodologies. METHODS: Eighty scene pictures were rated on familiarity and complexity by 20 native Cantonese-speaking adults (mean age: 22.3 ± 3.79 years; 10 female) using a 1 to 5 scale. These pictures were then used in recognition tests with a different group of 28 native Cantonese-speaking adults (mean age: 22.96 ± 3.54 years; 16 female). Participants studied 40 pictures and were tested with 20 studied pictures and 20 distractors after 10 minutes (immediate test) and with the remaining 20 studied pictures and 20 new distractors after 24 hours (overnight test). Each picture was tested in all four contexts (immediate-studied, immediate-distractor, overnight-studied, overnight-distractor) with seven participants per context. Recognition accuracy was analyzed using a logistic mixed-effects model, examining main effects and interactions of picture familiarity, complexity, target vs. distractor status, and test delay, with gender and age as control variables, and random effects for subjects and picture items. FINDINGS: An interaction between picture familiarity and complexity (p = 0.026) shows that accuracy improves with increased familiarity at high complexity (complexity = 3.66). This trend weakens (complexity = 2.91) or reverses (complexity = 2.15) as complexity decreases. This finding aligns with the study showing a negative correlation between familiarity and memory accuracy for low-complexity object pictures (Brodeur et al., 2017). The interaction is more pronounced in immediate tests but absent in overnight tests, as revealed by a marginal three-way interaction with test delay (p = 0.075). The model also predicts larger accuracy variations at extreme familiarity and complexity levels. Accuracy is higher in immediate tests (p < 0.001) due to memory decay, and distractor pictures show higher accuracy (p = 0.077), likely due to decision biases. DISCUSSION: These findings necessitate the careful selection of pictures with moderate familiarity and complexity, and the rigorous balancing of interactions between familiarity and complexity for both target and distractor pictures in short-delayed recognition tests. These strategies could enhance the accuracy of declarative memory assessments and support language-related studies.
Topic Areas: Methods, Language Development/Acquisition