Abstract:

During the past decades, an increasing amount of studies have focused on the emotional effect of old adults’ memory. However, not a consensus has been reached yet on the effect of emotional valence on false memory in older adulthood. To explain this inconsistency among previous studies, the present study suggested a lack of consideration on both participants' memory quality and response bias. In addition, few studies have ever taken the temporal factor into account; the emotional effect of memory may alter with time. To fill this gap, the present study examined how emotional valence and time interval would influence older adults’ false memory (FM) for emotional pictures within the framework of Signal Detection Theory (SDT). SDT allows for a simultaneous inspection on both discriminability (d′; an index of memory quality) and judgment criteria (β; an index of response bias). Twenty-one elderly participants (aging 67.17 ± 5.03) completed a recognition memory task consisting of one learning phase and two follow-up recognition tests. The learning materials consisted of 180 pictures (60 positive, 60 negative and 60 neutral) selected from the International Affective Pictures System (IAPS). After the learning phase, participants were asked to finish two follow-up recognition tests half an hour later (short time-interval) and three weeks later (long time-interval), respectively. Two different sets of 90 pictures (30 positive, 30 negative and 30 neutral) selected from the IAPS were respectively used as interfering materials in the two recognition tests. It was found that (1) both discriminability (d′) and judgment criteria (β) were negatively correlated with FM (using false alarm rate as its index) in the short time-interval follow-up recognition test, suggesting a joint contribution of these two factors to FM. However, in the long time-interval recognition test, only β could predict older adults’ FM, suggesting that β, rather than d′, took a critical role in FM as memory blurred with time. (2) Regardless of the length of time- interval, no evidence of distinct d' was revealed between negative and positive pictures. To clarify, d′ might have nothing to do with the divergence of emotional effects on older adults’ FM. (3) Negative emotion caused a lower β and a higher FM in the recognition test with a short time-interval. However, when it came to the long time-interval condition, positive emotion took a similar effect. These results altogether suggest that the emotional valence influence older adults’ FM by impacting response bias (i.e. β) but not memory quality (i.e. d′). The effects of positive and negative valence on FM may reverse as encoding-retrieval time-interval prolongs. Specifically, when the time-interval is short, negative emotion leads to a lower β and a higher level of FM. However, as the time-interval prolongs, it may be positive emotion that causes the same effect. In the domain of FM, older adults’ disposition for positive information, named as “the positivity effect”, may be reflected as an increasing tendency to inaccurately reconstruct positive information into their memory with time.

Key words: older adults, emotional valence, false memory, time interval, discriminability, judgment criteria