HOW ARE TRAIT EVALUATIONS PROCESSED IN THE BRAIN?

1. Sample Stimuli Trait Adjectives Personal Semantics Address College Major Mother’s Name Trustworthy Polite Cruel Tasks Evaluation Task: Evaluate whether each trait adjective was “Like Me” or “Not Like Me.” Memory Task: Yes/No recognition task on PS items (PS Me) and an equal number of comparable foils (PS Not Me). ERP Recordings Trait Complexity and Post-Retrieval Activity • ERP session was 1 week (mean = 8.2 days, SD = 4.9) after interview. • 82 scalp sites-referred to A1 (re-referenced to A1+A2) • Electrodes in rings spaced at 11.25% of vertex-nasion distance Traits were divided into low- and high-complexity items based on a median split of RT. Stimulus-Locked Averages • 2150 ms epoch • 150 ms baseline Behavioral Results Response-Locked Averages For these comparisons, trait items were collapsed over valence (e.g., Like Me, Not Like Me) Accuracy Reaction Time Accuracy R RT (ms) Distinct Peaks RT was significantly different across conditions, with PS Me having the fastest RT and Traits having the slowest RT (ps < 0.05). Although accuracy was high in all conditions, responses to PS Not Me items were significantly more accurate than Trait evaluations (p < 0.05). 450 – 650ms Role of Episodic Memory in Trait Evaluations Introduction To determine the role of episodic memory in trait evaluations, their ERPs were compared to those of PS Me and Not Me items in the 450-650 ms interval. Trait items elicited significantly smaller ERPs than PS Me items, but significantly larger ERPs than PS Not Me items (450-650 ms) (ps <.01), which suggests that retrieval of episodic memories plays a role in trait evaluations. Self-referential evaluations about traits determine how we think about ourselves and therefore affect how we approach and avoid things in our daily lives. Although our traits are part of the basis of our self-conceptions, the neural mechanisms that underlie how these traits are represented and processed in the brain remains poorly understood. Previously, Johnson et al. (in prep) assessed the role of episodic retrieval processes during attitude evaluations by measuring the electrical activity of the brain with event-related potentials (ERPs). That experiment looked for the presence of a parietal episodic memory (EM) effect, which is an ERP component shown to reflect the use of recollective processes, during attitude evaluations as compared to retrieval of autobiographical (AB) memories. These data revealed that the same episodic recollective processes were present during attitude evaluations and retrieval of AB memories. However, when trials were divided on the basis of reaction time (RT), the results showed differences in the relative timing of recollection and response only for attitude items. That is, evaluation of complex attitudes was characterized by a long retrieval-response delay suggestive of post-retrieval evaluation processes. By contrast, evaluations of simple, preference-like, items showed no delay between retrieval and response and thus appeared to not require post-retrieval evaluation. The data thus led to the conclusion that the timing of memory retrieval processes involved in attitude evaluations vary as a function of attitude complexity. The present experiment sought to determine whether a similar complexity-based difference existed for trait evaluations. That is, whether more complex trait items (e.g. Heartless) take longer to process than less complex trait items (e.g. Kind). Personal Semantic (PS) items (i.e., facts about one’s self) served as the comparison to test this because, while they are also retrieved from memory, they are thought to be, like traits, relatively context free. Hence, the retrieval processes used for PS items provides a better comparison for examining the role of memory in self-referential evaluations of traits. HOW ARE TRAIT EVALUATIONS PROCESSED IN THE BRAIN? As previously found, for the PS Me and Not Me items, responses occurred very soon after retrieval. However, there was a significant lag between retrieval and response for the Trait items, suggestive of the occurrence of post-retrieval processing. The fast items are presumed to be of low-complexity and to require little evaluation of retrieved memories. This is supported by the finding that the low-complexity items look similar to the PS items, in that the response occurs soon after the retrieval. By contrast, slow items are assumed to be high-complexity items and require more evaluation of retrieval products prior to one’s response. This is supported by the significantly greater lag between retrieval and response for the high-complexity trait items. Claudia Rodriguez, Heather Henkell, Deena Rendel, Adina Bitton, Christopher Schroeder & Ray Johnson, Jr. Department of Psychology, Queens College and The Graduate Center of CUNY For slower and/or more difficult retrievals, such as those about traits, the timing of retrieval-related processes is likely to be more tightly linked to the response than to stimulus onset. In addition, given the large differences in RT as a function of evaluation/retrieval time (Low complexity 686 ms; High complexity 966 ms), averages were created that were time-locked to the response. Methods • PARTICIPANTS: • 21 (16 females) right-handed, native English speakers were paid $10/hour • Mean Age = 23.7 years (S.D. = 4.0) • Informed Consent was obtained prior to experiment • PROCEDURE- INTERVIEW: • One week prior to the experiment, participants were interviewed and provided information on 2 types of stimuli: • Traits: Does this trait describe me? (1 = never, 6 = always) • PersonalSemantics: Basic personal factual information (e.g. occupation, birthday, address, relatives’ names, etc.) • Participants also rated each item on importance or centrality: • Traits: How important is the fact that this trait does/does not describe you as a person? (1 = not at all important, 6 = very important) • PersonalSemantics: How central is this fact to your personal identity? (1 = not at all central, 6 = very central) • Individualized stimulus files were created using items with the highest importance/centrality ratings thus obtaining 24-40 items for Traits (Like Me 5-6, Not Like Me 1-2) and 35 Personal Semantic items were chosen for each participant. These averages reveal clear amplitude differences between the Low- and High-complexity items. Low-complexity items elicit a single peak at the time of the response, which is consistent with previous findings of PS and autobiographical memories. By contrast, High-complexity items show a smaller peak at the time of the response and a secondary peak approximately 200 ms before the response, which offers additional support that these items require post-retrieval processing. PS Episodic Memory Effect Discussion PS Me – Not Me Potential Map • The typical parietal EM effect was found for Me PS items, indicating that personal semantic information is stored in episodic memory, rather than in semantic memory as some thought. • Comparing the brain activity elicited by trait evaluations with that elicited by items known to be and known not to be in episodic memory (i.e., PS Me and Not Me items, respectively) suggests that such judgments are based on representations stored in episodic memory. • An effect of complexity on trait evaluations was found, replicating previous findings. Low-complexity trait items were similar to PS Me items in that they appeared to come from memory and did not require post-retrieval evaluation prior to the response. • By contrast, a greater lag between retrieval and response for high-complexity trait items was found, indicating the need for greater evaluation of the retrieval products prior to the response. Response-locked ERPs revealed two distinct ERP components for these items, further supporting that after retrieval (i.e., the first peak) additional processing occurred prior to the categorization of the response (i.e., the second peak). A comparison of the ERPs elicited by PS Me and PS Not Me items revealed a significant parietal EM effect (450-650ms) (p<.001). The greater amplitude for PS Me items indicates that they were retrieved from episodic memory. Map of the PS Me-Not Me difference shows the typical parietal focus of the EM Effect http://brainlab.psych.qc.cuny.edu