老年人自传记忆更偏事实细节，新研究称未必是衰退

研究做了什么

这项研究关注的是自传体记忆，也就是人对自己生命经历的回忆系统。它通常被分成两类：情节记忆是带有具体时间、地点和感官细节的一次性事件，比如上个月在伦敦参加某场聚会；个人语义记忆则是关于自己生活的一般性事实，不绑定于某个具体场合。介于两者之间的是对重复事件的记忆，比如每周去朋友家做饭看电影，它概括了许多相似场合，但仍保留部分情境细节。

英国东安格利亚大学Louis Renoult团队让参与者在线完成记忆任务：屏幕上逐个出现21个中性、易于想象的提示词，参与者按提示回忆一个特定的一次性事件，或一个反复发生的类别性事件，然后口述最多两分钟。实验分两部分，基线块中参与者连续按同一规则回忆，切换块中规则每两到三次试次不可预测地改变。研究者据此比较规则刚改变与规则保持的试次，并统计回忆内容中各类细节的比例。

切换规则增加负担，但老年人没有更吃亏

最终样本为37名60至85岁老年人和37名18至33岁年轻人，两组教育水平相近，自报抑郁与焦虑症状无差异；其中各30人完成注意与工作记忆的标准测试，组间无统计显著差异。

整体准确率从基线块的82%降到切换块的73%，说明同时维持两套回忆规则确实增加了难度。但在切换块内部，规则刚改变的试次与规则保持的试次准确率没有差别，作者据此认为负担来自同时记住两种规则，而不是切换动作本身。更关键的是，老年组在切换条件下并未比年轻组下降更多。

在具体事件回忆上，两组准确率几乎相同，老年人81%，年轻人80%。差异出现在类别性重复事件上：老年人按指令给出合格回忆的比例为68%，年轻人为81%。当老年人未能给出合格的重复事件记忆时，最常转向的是一般性的个人事实。

叙述风格的年龄差异

老年人平均口述时间更长，约45秒对27秒，每条记忆产生的细节总数也更多，平均11.00个对7.41个，因此研究者比较的是各类细节所占比例而非绝对数量。

即便老年人成功回忆起一次性事件，其叙述中特定情节细节也只占66%，年轻人为79%；一般语义事实占老年人叙述的18%，年轻人占13%。在重复事件回忆中，老年人给出的目标重复事件细节比例为48%，年轻人为72%，而一般语义信息占到40%至42%，年轻人则为15%至24%。这一模式在基线块和切换块中都存在。

主观评分上，老年人虽然具体细节比例更低，却把特定记忆评得更加生动，平均3.81分对3.53分（5点量表），整体情绪强度也更高，2.92对2.46。回忆的时间分布同样不同：年轻人的记忆多来自过去五年，尤其是过去一年；老年人在39%的特定事件试次和48%的类别事件试次中回忆的是十年以上的往事，年轻人分别为15%和7%。

作者如何解释，以及哪些限制要紧

Renoult在采访中强调，老年人在大多数试次中仍能按指令提取正确类型的记忆，只是细节构成不同。他认为，老年人更多依赖事实性知识来讲述过去，可能是有意义、有适应性的叙述方式，而不一定是情节记忆效率下降后的补偿，也不等于记忆能力的全面衰退。他还提到，语义记忆在老年期保存较好，有些任务上老年人甚至优于年轻人。

解释这些结果时需要注意几点。研究的设计与分析方案未预注册；基线块总是先于切换块完成，以便参与者学习规则，因此疲劳或“容易回忆的往事先用掉”都可能贡献了切换阶段的表现下降。试次数量为控制在线测试疲劳而保持较少，导致无法分别检验规则切换对两类记忆的影响。

样本方面，这批老年人在工作记忆与注意测试上与年轻大学生相当，而执行控制通常随年龄下降，因此他们可能不代表更广泛的老年群体，这也可能解释他们为何在提示下顺利找到具体记忆。参与者以女性为主，且可自由选择人生任何时段的记忆。论文中有一处具体指令下切换效应的数字与别处重复，该效应大小因此不确定。

研究团队下一步计划考察交谈的社会情境如何影响老年人的记忆提取，包括听记忆的人是谁。Renoult提到，文献中存在一个相关偏向：老年人通常由年轻研究者访谈，他们可能因此加入更多一般性背景知识，以便向不同年龄的人解释。

A new study published in Psychology and Aging indicates that having to switch between recalling different types of personal memories makes it harder for both young and older adults to follow instructions accurately. Beyond the mental effort of juggling changing rules, the findings suggest that older adults tend to include more general personal facts and fewer specific details than younger adults, whether they are recalling a one-time event or a recurring routine.

Autobiographical memory is the mental system that allows people to recollect their own life history. It encompasses two main kinds of information: episodic memory and personal semantic memory. Episodic memories are rich recollections of unique events tied to a specific time and place, such as a particular birthday dinner last month. Personal semantic memories are broader facts about one’s life that are not tied to a single episode, such as general self-knowledge. Falling between the two are memories of repeated events, like a weekly habit of cooking with a friend, which summarize many similar occasions while still carrying some sensory and contextual detail.

As people age, the balance between these forms of recall tends to change. A benchmark 2002 study of younger and older adults found that when asked to describe past life events, younger adults mostly shared specific details about what happened, where they were, and how they felt, whereas older adults shared fewer time-and-place details and favored general, factual knowledge.

“A typical example would be that if a young adult is telling you about their last birthday party, they are more likely to include specific contextual and perceptual details (such as the specific time and place, what clothes people wore at that party, or specific details about what they said),” senior author Louis Renoult, a professor of cognitive neuroscience and the postgraduate research director in the School of Psychology at the University of East Anglia, told PsyPost. “In contrast, when remembering the same event, older adults would typically provide more factual background knowledge about similar birthday parties or other celebrations they would have attended, or more general knowledge they have about the people involved, for example.”

This pattern raised questions about whether the shift toward general memories comes from changes in how memories are stored or from age-related declines in executive control, which is the set of mental skills used to manage goals, stay focused, and filter out irrelevant thoughts. A 2013 study explored this by changing how instructions were given while participants listened to musical clips, finding that younger adults adjusted how specific their memories were to match the rules, while older adults did not adjust their recall as much when task demands grew more complex.

To test how mental flexibility affects memory recall, another 2013 study of university students introduced a test requiring participants to rapidly alternate between recalling unique, one-time events and repeated routines. The students were less accurate at providing specific memories when they had to switch back and forth compared to when they followed a single rule at a time.

“In the field, people generally assume that the greater reliance on factual knowledge (and semantic memory) in aging would constitute a compensation, due to a less efficient episodic memory,” explained Renoult, who also recently discussed memory and aging on a podcast episode. “However, tests commonly used to assess semantic and episodic memory are rarely well matched in difficulty, and most tests of autobiographical memory focus on episodic memory only. Due to these issues, my research group has developed new memory tests to study semantic and episodic memory in parallel, in conditions where they are well matched experimentally to better understand memory changes as we get older.”

To examine whether mental switching places an extra burden on older adults during memory recall and how it interacts with the broader shift toward factual details, Renoult and his colleagues carried out the new study. The work was led by Greta Melega while she was completing her doctorate at the University of East Anglia; she is now a postdoctoral researcher at Charité – Universitätsmedizin Berlin (“and also an elite European Ultimate Frisbee player (!),” Renoult noted).

The researchers recruited 45 younger adults and 45 older adults, all of whom were native English speakers without psychiatric or neurological conditions. After excluding participants who did not finish the study, had poor audio recordings, fell below a cognitive screening threshold, or did not pass an initial training check, the final sample included 37 older adults aged 60 to 85 and 37 younger adults aged 18 to 33.

The two age groups had similar levels of education and did not differ in self-reported symptoms of depression or anxiety, which the authors measured because mood can influence how specifically people recall the past. A subset of 30 younger and 30 older participants also completed standard tests of attention and working memory, showing no statistically significant differences between the age groups.

Participants completed an online memory task where they saw 21 neutral, easy-to-visualize cue words on a screen, one at a time. Depending on the prompt, they had up to one minute to think of either a specific memory (a unique event lasting less than a day, such as having “a good time at Jane’s party last month in London”) or a categoric memory (a repeated event that happened multiple times in a similar way, such as “When I go to Jane’s house, we usually cook together and watch a movie afterwards”).

Once a memory came to mind, they pressed a button and spoke aloud for up to two minutes to describe it in as much detail as possible. After each description, participants had five seconds per question to rate how vivid and emotional the memory was, whether they saw it through their own eyes or as an outside observer, and how long ago it took place.

The experiment was divided into two parts: a baseline block and a switching block. In the baseline block, participants followed one instruction at a time, recalling three specific memories in a row and three categoric memories in a row. In the 14-trial switching block, the instructions to recall specific or categoric memories changed unpredictably every two or three trials. This allowed the scientists to compare “switch trials,” where the rule had just changed from the previous prompt, with “remain trials,” where the rule stayed the same as the prompt right before it.

The research team evaluated the recorded narratives in two ways. First, they measured accuracy, defined as the percentage of memories that matched the prompt given on that trial. Second, for the memories that matched the instructions, the researchers broke the transcripts down to count the exact types of details produced: episodic details (specific sights, times, and feelings of a one-time event), repeated event details (context and sensory details of a recurring routine), semantic details (general facts and personal background knowledge), and other statements like repetitions. Because older adults spoke longer on average than younger adults (about 45 seconds versus 27 seconds) and produced more total details per memory (an average of 11.00 details compared to 7.41), the authors analyzed the proportion of each detail type rather than raw counts.

When looking at overall accuracy, both younger and older participants had a harder time following instructions when they had to juggle both rules in the switching block, dropping from 82 percent accuracy in the baseline block to 73 percent in the switching block. Within the switching block itself, accuracy did not differ between switch trials and remain trials, suggesting that the mental cost came from keeping two different rules in mind at once rather than the immediate act of switching from one trial to the next.

Older adults were not disproportionately harmed by the switching block compared to younger adults. Across the experiment, the two age groups were equally accurate when asked to recall specific, one-time events, with older adults succeeding on 81 percent of trials and younger adults on 80 percent. However, older adults were less accurate than younger adults when instructed to recall categoric, repeated events, matching the instructions 68 percent of the time compared to 81 percent for younger adults. When participants failed to provide a proper categoric memory, they most often defaulted to sharing general personal facts instead of an actual recurring event.

When the researchers examined how participants elaborated on their accurate memories, strong age differences appeared regardless of whether participants were in the baseline or switching block. Even when older adults successfully recalled a unique, one-time event, only 66 percent of the details they shared were specific episodic details, compared to 79 percent for younger adults. In turn, general semantic facts made up 18 percent of older adults’ descriptions of unique events, compared to 13 percent for younger adults.

A similar pattern emerged for categoric memories: older adults included a smaller proportion of on-target repeated event details than younger adults (48 percent versus 72 percent) and a much larger proportion of general semantic information across both blocks (40 to 42 percent for older adults, versus 15 to 24 percent for younger adults).

“As observed previously, older adults consistently produced more semantic (factual) details but less episodic (specific) elements than young adults,” Renoult said. “However, crucially, this was not only observed under specific (episodic) instructions, but also under categoric (semantic) instructions. This suggests that remembering the past using more factual knowledge is adaptive and meaningful to older adults, and not necessarily the result of a compensation, or of a general decline in their memory abilities.”

The study itself did not test what purpose, if any, this narrative style serves.

Increased task demands in the switching block also altered the makeup of memory details. Under categoric instructions, moving to the switching block reduced the proportion of target repeated event details from 67 percent to 54 percent across participants. The paper reports a similar shift under specific instructions, but its figures for that comparison duplicate those given elsewhere, so the size of that effect is uncertain.

The subjective ratings highlighted additional differences in how the two age groups experienced their memories. Even though older adults included proportionally fewer specific details in their stories, they rated their specific memories as more vivid than younger adults did (an average of 3.81 versus 3.53 on a 5-point scale) and rated their memories overall as more emotional (2.92 versus 2.46). Regarding the age of the memories, younger adults mostly drew from the past five years, especially the past year, whereas older adults recalled memories from more than 10 years ago on 39 percent of specific trials and 48 percent of categoric trials, compared to just 15 percent and 7 percent for younger adults.

“It is important to realize that, even though young and older adults remember their past differently (with different narrative style), in these studies participants of both age groups are able to follow the overall instructions and retrieve the relevant type of information in most trials (i.e., they produce the correct type of memory, even if with different amount of details),” Renoult noted. “Nonetheless, the results that we report in this study are consistent with previous findings from our group using different methods.”

Renoult also emphasized that differences in how older adults recount their past should not be mistaken for a broad loss of memory function. “Sometimes people have gloomy views of aging, but not all types of memory decline when we get older, as shown in studies such as ours: Semantic memory (general knowledge of the world) is well preserved in older adults (who sometimes outperform young adults in these tasks).”

As with all research, there are a few things to keep in mind when interpreting these results. The study’s design and analysis plan were not preregistered. Because the baseline block was always completed before the switching block so participants could learn the rules, fatigue or the possibility that the easiest memories were used up first could have contributed to the drop in performance during the switching phase. The number of trials was also kept relatively small to prevent tiredness during online testing, which meant the authors could not separately test whether switching rules hurt one memory type more than the other.

In addition, the older adults in this sample performed just as well as the younger undergraduate students on standard cognitive tests of working memory and attention. Because executive control often declines with age, this high-performing group of older adults might not reflect the broader aging population, which could explain why they had no trouble finding specific memories when prompted. The participants were also predominantly female, and they were allowed to choose memories from any point in their lives; future studies could restrict recall to specific time periods and use standardized follow-up prompts to see if older adults can retrieve more specific sensory details when encouraged to do so.

Looking ahead, the research team plans to explore how the social context of a conversation shapes the way people recount their lives. “We are interested to investigate social elements that may influence memory retrieval in aging, such as with whom one is sharing their memories,” Renoult said. “One related bias in the literature is that older adults are typically interviewed by a young adult (e.g., a PhD student, or a research assistant). It is possible that, in such cases, older adults include more general knowledge to provide additional background information for someone of a different age group.”

The study, “Age-Related Differences in Autobiographical Memory Recall: Impact of Retrieval Mode Switching,” was authored by Greta Melega, Alice Bush, Tabitha James, Riccardo Sacripante, and Louis Renoult.