Exploring differences in the processes and mechanisms of building face familiarity

Abstract

The human visual system appears is finely tuned for processing facial information, leading to enhanced perceptual sensitivity and discrimination abilities for faces compared to non-face stimuli (Rossion, 2008). The current thesis focuses on exploring and understanding face learning, in particular the building of face familiarity and how unfamiliar faces become familiar. It investigates the cognitive mechanisms underlying human face recognition, focusing on how individuals learn unfamiliar faces through various stimuli types and tests while considering the influence of personality traits and socio-emotional functioning. Through a series of eight experiments, the research demonstrates that exposure to multiple static images of unfamiliar faces, each exhibiting high variability, leads to superior outcomes in building face representation compared to alternative stimuli. In Study 1, best learning performance is observed with multiple static images, indicating that diverse exposure enhances memory formation. This highlights the importance of image or face variability in fostering more robust facial representations. Study 2 confirms this advantage by demonstrating that even when static images are presented as a single recognition episode, recognition accuracy is greater relative to motion clips. Study 3 further emphasizes that recognition rates for static multiple images remain higher than those for motion clips, illustrating that variability is a more crucial factor than motion in building effective face representations. Subsequent studies focus on the impact of contextual factors on recognition accuracy. Both Study 4 and 5 reveal no significant influence of background context on performance, indicating that facial features remain the primary cue for recognition, independent of the surrounding environment. In Study 4 and 5, recognition accuracy decreases when different background contexts are introduced, suggesting that context may introduce cognitive noise. And further confirms the consistency of static multiple images, achieving the highest hit rates. The remaining studies delve into the advantages of presenting multiple images using a visual search methodology. Study 6 highlights enhanced identification efficiency in complex visual arrays, where participants were provided with conceptual information as identity identifiers during the learning phase. This approach facilitates more accurate identification of target faces among distractors, emphasizing the role of cognitive integration in recognition tasks. Study 7 and 8 demonstrate that recognition performance remains robust even with increased task complexity, showcasing the adaptability of face recognition strategies. Empathy and social anxiety's influence on face recognition performance varied across the studies. While some studies observed that higher empathy correlated with improved recognition accuracy, others found that social anxiety negatively impacted performance in others, though this effect was not consistent throughout. A key general finding across the studies was the negative correlation between empathy and social anxiety. This suggests that these emotional traits are inversely related and may influence cognitive processing in distinct ways during face recognition tasks. In conclusion, this thesis underscores the critical role of image variability in forming robust facial representations and achieving accurate recognition. The findings indicate that effective face learning is significantly enhanced through high variability in static multiple images. Overall, the work offers a comprehensive perspective on the complexities of face recognition.

Date of Award	17 Dec 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Karen Lander (Main Supervisor) & Daniela Montaldi (Co Supervisor)