Theses Doctoral

The Behavioral and Neural Effects of Rejection Sensitivity on Selective Attention and Feedback-Based Learning

Crew, Christopher

Gaining acceptance and avoiding rejection is arguably one the most fundamental and challenging relational tasks that we face. Given the importance of close relationships, an especially serious threat is rejection, real or imagined, by significant others. Considerable research supports the idea that prolonged exposure to harsh rejection can have deleterious effects on one's physical and emotional wellbeing (Baumeister & Leary, 1995; see Dickerson & Kemeny 2004, for a full review). Research also suggests that early experiences with rejection can result in a bias to anxiously expect and readily perceive rejection in other's behavior - a disposition known to derail interpersonal relationships. This phenomenon is known as Rejection Sensitivity (RS; Feldman & Downey, 1994; Downey & Feldman, 1996).
There have been important advances in understanding psychological and physiological responses to interpersonal rejection (e.g., Downey & Feldman, 1996; Downey, Mougios, Ayduk, London, & Shoda, 2004; Dickerson & Kemney, 2004; Romero-Canyas & Downey, 2005; Powers, Pietromonaco, Gunlicks, Sayer, 2006; Richman & Leary, 2009). However, relatively less is known about patterns of attentional processes underlying reactions to rejection cues and events, as well as the extent to which RS impacts learning and memory. These unanswered questions are of critical importance as theory and research suggests that information-processing biases may provide an explanation for the maintenance of RS and disorders like social phobia and anxiety that share many of the characteristics of rejection sensitive individuals (See Bar Haim et al., 2007 for a meta-analytic review).
Study 1 uses a well-established attentional control paradigm (Attentional Network Task - ANT; Fan et al., 2002) to assess the relationship between RS and basic attentional mechanisms for alerting, orienting, and executive control. Results from study 1 suggest that RS is not associated with the functioning of attentional networks important for alerting, orienting, and executive control, raising the possibility that RS operates as a distinct system that interacts with attentional networks to influence attention deployment in the presence of social threat cues. This hypothesis is tested in study 2.
Study 2 uses a selective attention paradigm that measures eye movements during a visual probe task (e.g., MacLeod, Mathews, & Tata, 1986) in order to assess patterns of attention deployment to socially threatening stimuli in RS individuals. Study 2 also tests the attenuating effects of executive control on processing of social threat cues in RS individuals. The latter part of study 2 is designed to address important theoretical and empirical questions about the ability of attentional control to attenuate maladaptive information processing biases in RS individuals. Results suggest that RS is associated with initial vigilance and later avoidance for social threat cues but, as predicted, vigilance for social threat cues is attenuated by high executive control. That is, having good executive control (as measured by self-report and behavioral measures - the ANT) can help to reduce the extent to which social threat cues capture and hold the attention of RS individuals.
Study 3 was designed to answer the question of how the tendency of RS individuals to detect and react to social threat cues can affect more overt forms of learning and memory (i.e., declarative memories). In order to address this question, study 3 used an incidental-learning paradigm where participants answered general knowledge questions (What is the capital of Delaware?) followed by immediate performance accuracy (correct vs. incorrect) and the correct answer (Dover). Initially incorrect items were retested 24 to 48 hours later to determine if the correct answer had been successfully encoded.
Event-Related Potentials (ERPs) were used to measure neural responses to performance feedback (correct vs. incorrect at first test) and learning feedback (the correct answer) to assess whether (1) RS is associated with greater sensitivity to performance feedback in general or specifically for social performance feedback, (2) whether these reactions mediate successful learning (i.e., retrieval of corrective feedback), and (3) whether there are gender differences in how RS operates in an evaluative context, which would provide an explanation, based on neural mechanisms, to previously found differences in which RS females seem to be more vulnerable to reduced achievement in competitive academic settings (London et al., 2013).
Overall, behavioral results suggest that individuals were able to encode and retrieve corrective information after receiving social (face) performance feedback at the same rate as they were after receiving non-social (symbol) performance feedback, suggesting that contextualizing performance feedback within the social domain did not generally enhance or impair learning and memory. However, within females, higher RS scores were associated with poorer retrieval in the social performance feedback condition suggesting that RS moderates the effect of social performance feedback on retrieval in females but not in males. To better understand the mechanisms underlying these behavioral effects we examined the following ERP waveforms associated with processing of social and non-social performance feedback: the frontally-maximal feedback related negativity (FRN), the frontally-maximal orienting effect (P3a), and a centrally-maximal late positive potential (LPP). Respectively, these components have been shown to reflect more automatic processing of feedback valence, orienting responses to rare events, and sustained attention to motivationally relevant information. Finally, ERP waveforms associated with processing of the corrective feedback were also analyzed.
Consistent with previous research, the FRN was enhanced in response to performance feedback indicating that an incorrect response had been made while the P3a and LPP were enhanced in response to performance feedback indicating that a correct response, a rarer outcome in this challenging task, had been made. There were no gender differences in the overall amplitude of the FRN, P3a or LPP. However, within females, RS was associated with a smaller FRN amplitude in the social performance feedback condition. Analyses were also conducted on the relationship between these ERPs, encoding of the corrective feedback (i.e., seeing the correct answer on the screen), and subsequent memory (i.e., correctly answering the question at retest). Although the P3a and the LPP were not associated with encoding of the corrective feedback or subsequent memory, the FRN positively predicted greater processing of the corrective feedback and subsequent memory in the social feedback condition. However, within females, the FRN negatively predicted encoding of the corrective feedback and subsequent memory only in the social condition. Finally, a mediation analysis was used to further understand the process by which neural responses to the performance feedback might affect processing of the corrective feedback and subsequent memory overall and perhaps differently for RS females and males. Results suggest that social performance feedback reduces retrieval success in RS females by reducing the level of engagement with corrective feedback, ultimately resulting in poorer encoding into long-term memory.
This knowledge could help expand our understanding of how rejection cues may disrupt, by triggering maladaptive strategies, the attention deployment of individuals who are especially sensitive to social threat whether for personal reasons (e.g., a history of experience with harsh rejection from caregivers) or because of membership in a marginalized social group (e.g., women in law or STEM fields). In doing so, this research could identify important avenues for interventions that work to enhance interpersonal functioning in RS individuals by training them to use self regulatory strategies that reduce attentional biases and augment information processing (i.e., learning and memory).


  • thumnail for Crew_columbia_0054D_12060.pdf Crew_columbia_0054D_12060.pdf application/pdf 2.04 MB Download File

More About This Work

Academic Units
Thesis Advisors
Downey, Geraldine
Ph.D., Columbia University
Published Here
July 7, 2014