OBJECTIVES Whether people with Alzheimer's Disease present with accelerated long term forgetting compared to healthy controls is still debated. Typically, accelerated long term forgetting implies testing the same participants repeatedly over several delays. This testing method raises the issue of confounding repetition effects with forgetting rates. We used a novel procedure to disentangle the two effects. METHODS Four short stories were presented during an initial in-person assessment of 40 patients with Alzheimer's Disease and 42 age-matched healthy controls. Our aim was for participants to reach a score of 70% correct (9 out of 13 questions) at encoding. If this criterion was not achieved after the first trial, the four stories were presented again (in a different order); participants took the 1 min filler task again and were then retested. We repeated this process until participants reached the 70% criterion or to a maximum of four trials. Cued recall memory tests were completed during follow-up telephone call(s) at different delay intervals. Study material was presented only at encoding, then probed with different question sets on all other delays. Each question set tested different sub-parts of the material. The experiment employed a mixed design. Participants were randomly allocated to either a condition without retrieval practice or a condition with retrieval practice. Participants in the condition without retrieval practice were only tested at two delays: post encoding filled delay and at one month. Participants in the condition with retrieval practice were tested at four delays: post encoding filled delay, one day, one week and one month. Our methodological design allowed us to separate the effects of retesting from the effects of delay. RESULTS Alzheimer's Disease patients showed a significant encoding deficit reflected in the higher number of trials required to reach criterion. Using Linear Mixed Models, we found no group by delay interactions between the post encoding filled delay retrieval and one month delays, with Alzheimer's Disease groups having a similar decline in performance to healthy controls, irrespective of testing condition. Significant condition by delay interactions were found for both groups (Alzheimer's Disease and healthy controls), with better performance at one month in the condition with retrieval practice. CONCLUSIONS Our data showed that Alzheimer's Disease is not characterised by accelerated long term forgetting, patients in our sample forgot at the same rate as healthy controls. Given the additional trials required by Alzheimer's patients to reach the 70% correct criterion, their memory impairment appears to be one of encoding. Moreover, Alzheimer's Disease patients benefited from repeated testing to the same extent as healthy controls. Due to our methodological design, we were also able to show that performance improved under repeated testing conditions, even with partial testing (sampling different features from each narrative on every test session/delay) in both healthy controls and Alzheimer's Disease.

中文翻译：

---

阿尔茨海默病的遗忘：速度快吗？是否受到重复检索的影响？

目标 与健康对照者相比，阿尔茨海默病患者是否会出现长期遗忘加速的情况仍有争议。通常，加速长期遗忘意味着在几次延迟中重复测试相同的参与者。这种测试方法提出了混淆重复效应和遗忘率的问题。我们使用了一种新颖的程序来消除这两种效应。方法 在对 40 名阿尔茨海默病患者和 42 名年龄匹配的健康对照者进行初步现场评估时，提出了四个短篇故事。我们的目标是让参与者在编码时获得 70% 的正确率（13 个问题中的 9 个问题）。如果第一次试验后没有达到这一标准，则再次呈现这四个故事（以不同的顺序）；参与者再次进行 1 分钟填充任务，然后重新进行测试。我们重复这个过程，直到参与者达到 70% 标准或最多四次试验。线索回忆记忆测试是在不同延迟间隔的后续电话呼叫期间完成的。研究材料仅在编码时呈现，然后针对所有其他延迟用不同的问题集进行探讨。每个问题集测试了材料的不同子部分。该实验采用了混合设计。参与者被随机分配到没有检索练习的条件或有检索练习的条件。没有检索练习的情况下的参与者仅在两次延迟时进行测试：编码后填充延迟和一个月。具有检索练习条件的参与者在四种延迟下进行了测试：编码后填充延迟、一天、一周和一个月。我们的方法设计使我们能够将重新测试的影响与延迟的影响分开。结果阿尔茨海默病患者表现出显着的编码缺陷，这反映在达到标准所需的试验数量较多。使用线性混合模型，我们发现编码后填充延迟检索和一个月延迟之间没有延迟相互作用，阿尔茨海默病组的表现与健康对照组有类似的下降，无论测试条件如何。两组（阿尔茨海默病组和健康对照组）都发现了延迟相互作用的显着状况，在一个月后进行检索练习时表现更好。结论 我们的数据表明，阿尔茨海默氏病的特点不是加速长期遗忘，我们样本中的患者遗忘率与健康对照者相同。鉴于阿尔茨海默病患者需要进行额外的试验才能达到 70% 的正确标准，他们的记忆障碍似乎是编码之一。此外，阿尔茨海默病患者从重复测试中受益的程度与健康对照组相同。由于我们的方法设计，我们还能够证明在重复测试条件下性能有所提高，