Reconciling findings of emotion-induced memory enhancement and impairment of preceding items

Abstract

A large body of work has revealed that people remember emotionally arousing information better than neutral information. However, previous research has revealed contradictory effects of emotional events on memory for neutral events that precede or follow them: In some studies, emotionally arousing items have impaired memory for immediately preceding or following items, and in others arousing items enhanced memory for preceding items. By demonstrating both emotion-induced enhancement and impairment, Experiments 1 and 2 clarified the conditions under which these effects are likely to occur. The results suggest that emotion-induced enhancement is most likely to occur for neutral items that (a) precede (and so are poised to predict the onset of) emotionally arousing items, (b) have high attentional weights at encoding, and (c) are tested after a delay period of a week rather than within the same experimental session. In contrast, emotion-induced impairment is most likely to occur for neutral items near the onset of emotional arousal that are overshadowed by highly activated competing items during encoding.

Figure 1

Figure 1A. The experimental timeline used in Anderson et al. (2006) for each encoding trial. Modulator items were rated on arousal during each trial. Retrograde enhancement of memory was observed for faces and houses that preceded emotionally arousing scenes in a one-week delayed recognition test. Figure 1B. The experimental timeline used in Hurlemann et al. (2005) for each encoding trial (list). Retrograde amnesia was observed for items that preceded emotionally arousing oddballs in tests of immediate free recall after each trial.

Figure 2

Figure 2A. The experimental timeline used in Experiment 1 for each encoding trial (list) and the format of the recognition memory test. Within each encoding list, either three or four of the non-oddball items were real photographs of objects (natural stimuli) and the remaining items were computer-generated objects (artificial stimuli). Across all lists, half of the items were real photographs and the remaining half consisted of computer-generated objects. Participants were asked to indicate with a key press whether the non-oddball item presented was “natural” (a real photograph) or “artificial” (a computer-generated image). Each list also included one emotionally arousing oddball or one neutral oddball inserted into list position 3, 4, 5 or 6. During the run, participants were presented with 24 lists. Each list was followed immediately by a 25-second distracter task during which participants were presented with math problems and solutions and had to determine whether the solutions were correct or incorrect. Episodic memory was tested immediately after the 25-second math task and was signaled by the words “Please Recall.” At this time, participants had 30 seconds to write down as many items from the list as they could remember. The computer displayed the time remaining and a beep occurred at the end of the elapsed time to alert the participant that a new list would appear. Following the presentation of the 24 lists, half of the participants took an immediate recognition test for both the oddball and non-oddball items. The remaining half took the recognition test one week later. On each trial of the recognition test, participants were presented with a pair of items that shared the same verbal label (e.g., both were “apples”) but differed in other visual features (e.g., orientation, color, shape). The word “old” appeared below each item either in red (item on the left) or blue (item on the right) font. The word “new” appeared in yellow and was centered below the two “old” labels. Participants were instructed to indicate which of the two items matched the non-oddball item seen previously, or if both of the items were new. Figure 2B. The experimental timeline used in Experiment 2 for each encoding trial (list) and the format of the recognition memory test. Within each encoding list, two of the items were real photographs of objects (natural stimuli) and the two remaining items were computer-generated objects (artificial stimuli). Each list also included one emotionally arousing picture or one neutral oddball inserted into list position 2, 3 or 4. Because of the shorter lists than in Experiment 1, the math task was reduced from 25 to 20 seconds and participants were allotted 25 seconds (instead of 30) to write down as many items from the list as they could remember. Participants in the no-recall condition started with a new encoding list instead of recalling the list items after the distraction task. Following the presentation of the 24 input sequences, all participants returned one week later to take the recognition tests for oddball and non-oddball stimuli.

Figure 3

Probability of recall (hits - false alarms) for emotional and neutral oddballs and non-oddball items in list positions immediately preceding and following oddballs in Experiment 1. No significant difference emerged in memory for the neutral and emotional oddball pictures themselves. Recall for items in E+/-1 positions was depressed relative to memory for items in N+/-1 positions.

Figure 4

General recognition [hits (same + similar items collapsed) - false alarms] for non-oddball items preceding and following oddballs as a function of recall status at the immediate test (A) and the delayed test (B) in Experiment 1. Recognition performance was collapsed across the 4 positions that preceded oddballs to obtain an overall “before-oddball” measure. The same procedure was followed to yield an overall “after oddball” measure.

Figure 5

Probability of recall as a function of list position in Experiment 2.

Figure 6

Recognition accuracy [hits - false alarms] for non-oddball items that preceded and followed oddballs in the recall and no recall conditions in Experiment 2. Recognition performance was collapsed across the three positions that preceded and followed oddballs to obtain an overall “before-oddball” and “after-oddball” measure.

Figure 7

Recognition accuracy [hits - false alarms] for non-oddball items that preceded and followed oddballs as a function of memory type in Experiment 2. Recognition performance was collapsed across the three positions that preceded and followed oddballs to obtain an overall “before-oddball” and “after-oddball” measure.

Figure 8

Recognition accuracy [hits - false alarms] for non-oddball items that preceded and followed oddballs as a function of memory type in the recall condition (A) and the no recall condition (B) in Experiment 2. Recognition performance was collapsed across the three positions that preceded and followed oddballs to obtain an overall “before-oddball” and “after-oddball” measure.