Long-Term Memory : Latency Predicts Recognition
Long-Term Memory : Latency Predicts Recognition
复制标题
DOI：
--
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
Ray Johnson;Adolf Pfefferbaum;K. BERT S;Campbell Picton;Baribeau - Braun
中科院分区：
文献类型：
--
作者：
Ray Johnson;Adolf Pfefferbaum;K. BERT S;Campbell Picton;Baribeau - Braun
研究测试范式用于研究记忆获取过程以及重复对长期识别记忆的影响。在该过程中，向受试者提供要记忆的单词列表（“目标”）（研究系列）。随后在单词列表上进行识别测试，该单词列表由随机混合有相同数量的新干扰单词的目标单词组成（测试系列）。 Botb反应时间和tbe事件相关脑电位的tbe P300分量被用作处理时间的测量。在该研究系列中，尽管词类别概率为 1.0，但仍引出了较大的 P300。当研究系列中的单词根据识别性能进行划分时，随后识别的单词会导致无法识别的单词的延迟降低 P300。研究系列中的 P300 单词幅度在保持恒定潜伏期的同时增加了重复次数。在 tbe 测试系列中，P300 延迟和反应时间减少了 botb 目标词和干扰词的重复次数。与测试重复相比，所有单词的 P300 幅度显着增加，其中包含干扰词的 P300 会引起更大的 P300。在 tbe 测试系列中，单词 tbat 的识别更加一致，引起了更大和更早的 P300 tban 单词 tbat 的识别不太一致。来自 botb 研究和测试系列的 P300 幅度和潜伏期结果被解释为反映了随着记忆痕迹强度的增加，目标词的辨别力也随之增加。描述：事件相关电位、P300、识别记忆。研究发现，事件相关脑电位 (ERP) 的 P300 成分可以索引多种范式中的认知处理。 P300 有两个参数提供了信息处理的独特指标：幅度和延迟。 P300 幅度与刺激引起的信息处理量（即处理负载）直接变化，并与刺激的主观概率成反比变化（Duncan-Johnson & Donchin, 1977; Johnson & Donchin, 1980, 1982: Picton, Campbell, Baribeau-Braun, & Proulx, 1978; Tueting & Sutton, 1976; Tueting, Sutton 和 Zubin，我们要感谢 Connie C. Duncan-Johnson 仔细阅读了原稿的早期版本，并提出了许多有益的建议和批评。我们还要感谢 Christopher D. Wickens 和 James F. Juola 对本文早期草稿的评论。有关这些数据的初步报告已在 1980 年心理生理学研究学会第二十届年会上提出。这项研究得到了退伍军人管理局医学研究服务处的支持。转载至：Ray Johnson, Jr., National Institutes of Health, NINCDS/Medical Neurology Branch, Building 10, Room 4N246, Bethesda, MD 20205. 1970），这些变量的影响反过来又受到影响受试者在刺激后模棱两可的因素的调节（Johnson, 1979, 1984）。越来越多的证据表明，P300潜伏期可以用作刺激评估时间的指标（Duncan-Johnson，1981；Duncan-Johnson＆Donchin，1982；Johnson＆Donchin，1985；Kutas，McCarthy，＆Donchin，1977；Pfefferbaum，Ford，Johnson，Wencgrat，＆Kopell，1983），其他研究发现P300 延迟与响应选择和执行过程无关（Duncan-Johnson & Kopell，1981；Ford, Mohs, Pfefferbaum, & Kopell，1980；McCarthy & Donchin，1981）。因此，已经证明，刺激因素的操纵会影响反应时间（RT）和 P300 潜伏期，而反应因素的操纵会影响 RT，并且对 P300 的影响很小（如果有的话）（Ford 等，1980；McCarthy & Donchin，1981）。受试者对刺激性质的后验不确定性（即模棱两可）的影响在确定所产生的 P300 的幅度和延迟方面起着重要作用。例如，Johnson 和 Donchin (1978) 描述了一个实验，其中操纵了两种反馈音的可区分性。结果表明，两种刺激引起的 P3OO 振幅与其辨别能力直接相关。正如预期的那样，P300 潜伏期随着刺激辨别力的增加而减少（Johnson & Donchin，1985）。他们还发现，正反馈引发的 P3OO 比负反馈引发的 P3OO 更早且更大。许多识别记忆理论假设，放置在记忆中的项目以某种方式用连续变量标记，该变量提供了受试者对测试项目熟悉程度的衡量标准（Atkinson & Juola，1973；克雷克和图尔文，1975；拉特克利夫，1978）。将对某个项目的熟悉程度等同于将该项目与不相关项目区分开来的容易程度是合理的。此外，其中一些记忆理论认为，项目熟悉度值的增加伴随着识别该项目的速度的增加。鉴于 P300 和刺激辨别力之间的关系，该 ERP 组件特别适合研究记忆过程。因此，人们可以预测，对熟悉度维度的任何操作都会导致 P300 的振幅和延迟发生变化。一些 ERP 实验致力于确定 P300 测量值和记忆过程之间的关系。在 Stemberg 范式 (1969) 中，P300 延迟随内存搜索时间变化，与 RT 相同：P300 延迟内集（目标）项目的延迟比起始（干扰项）项目的延迟更短（Adam & Collins，1978；Ford、Roth、Mohs、Hopkins 和 Kopell，1979；Ford 等人，1980；Gomer、Spicuzza 和 O'Donnell， 1976）。此外，Ford 等人 (1980) 报告说，虽然刺激辨别性的变化（即正常刺激与退化刺激）不会影响 RT 和 P300 潜伏期函数的斜率，但退化刺激的两种测量的截距都会增加（参见 Sternberg，1969），在此范例中也观察到了 P300 幅度差异。例如，戈篾等人。 (1976) 报告说，目标比干扰物引发更大的 P3OO。可以合理地假设，随后被识别的项目或事件的处理方式与那些未被识别的项目或事件不同。由于信息处理行为测量的性质，检验这一假设很困难。也就是说，虽然反应时间对于量化记忆提取过程的时间很有用，但当受试者必须学习一系列项目时，它不太适合阐明记忆获取过程。然而，因为P300 amVol。 22、第 5 号强度和潜伏期指数表示刺激处理的程度和时间，它独特地能够提供对这些心理操作的洞察。本实验的目的是评估单词习得过程中 P300 的活性，作为长期识别记忆任务中后续表现的函数。由于识别是可辨别性的函数，我们预测在测试过程中识别得更好的单词会在记忆获取过程中引发更大、更早的 P3OO。为了评估记忆获取和检索阶段的刺激处理，我们使用了研究测试程序（Murdock & Dufty，1972；Ratcliff，1978）。在这种范例中，向受试者提供要记忆的单词列表（研究系列），然后在由目标单词与相同数量的新干扰单词（测试系列）随机混合组成的列表上进行识别测试。通过使用相同的目标单词列表重复学习和测试系列，还评估了重复对学习和记忆的影响。
Tbe study-test paradigm was used to investigate memory acquisition processes and tbe effects of repetition on long-term recognition memory. In tbis procedure, subjects are presented witb a list of words ("targets") to be memorized (Study series). Tbey are later tested for recognition on a word list comprised of tbe target words mixed randomly witb an equal number of new, distractor words (Test series). Botb reaction time and tbe P300 component of tbe event-related brain potential were used as measures of processing time. During tbe Study series, large P300s were elicited despite a word category probability of 1.0. Wben tbe words from tbe Study series were divided on tbe basis of recognition performance, words tbat were subsequently recognized elicited P300s witb sborter latencies tban unrecognized words. P300 amplitude to words in tbe Study series increased witb repetition wbile maintaining a constant latency. During tbe Test series, P300 latency and reaction time decreased witb repetition for botb target and distractor words. P300 amplitude to all words increased substantially over Test repetitions witb tai^et words eliciting larger P300s tban distractor words. Words tbat were recognized more consistently during tbe Test series elicited larger and earlier P300s tban words tbat were recognized less consistently. Tbe P300 amplitude and latency results from botb tbe Study and Test series are interpreted as reflecting tbe increased discriminability of tbe target words as tbe memory trace increases in strengtb. DESCRIPTORS: Event-related potentials, P300, Recognition memory. The P300 component of the event-related brain potential (ERP) has been found to index cognitive processing in a variety of paradigms. There are two parameters of the P300 that provide unique indices of information processing—amplitude and latency. P300 amplitude varies directly with the amount of information processing invoked by a stimulus (i.e., processing load) and varies inversely with the subjective probability of a stimulus (Duncan-Johnson & Donchin, 1977; Johnson & Donchin, 1980, 1982: Picton, Campbell, Baribeau-Braun, & Proulx, 1978; Tueting & Sutton, 1976; Tueting, Sutton, & Zubin, We wish to thank Connie C. Duncan-Johnson for her careful reading of earlier versions of the manuscript and making numerous helpful suggestions and criticisms. We also thank Christopher D. Wickens and James F. Juola for their comments on an earlier draft of this paper. A preliminary report on these data was presented at the Twentieth Annual Meeting of the Society for Psychophysiological Research, 1980. This research was supported by the Medical Research Service of the Veterans Administration. Address requests for reprints to: Ray Johnson, Jr., National Institutes of Health, NINCDS/Medical Neurology Branch, Building 10, Room 4N246, Bethesda, MD 20205. 1970), The effects of these variables are in turn modulated by factors that affect the subject's equivocation after a stimulus (Johnson, 1979, 1984). A growing body of evidence has demonstrated that P300 latency can be used as an index of stimulus-evaluation time (Duncan-Johnson, 1981; Duncan-Johnson & Donchin, 1982; Johnson & Donchin, 1985; Kutas, McCarthy, & Donchin, 1977; Pfefferbaum, Ford, Johnson, Wencgrat, & Kopell, 1983), Other studies have found that P300 latency is independent of response selection and execution processes (Duncan-Johnson & Kopell, 1981; Ford, Mohs, Pfefferbaum, & Kopell, 1980; McCarthy & Donchin, 1981). Thus, it has been demonstrated that whereas manipulations of stimulus factors affect both reaction time (RT) and P300 latency, manipulations of response factors affect RT and have little, if any, impact on P300 (Ford et al., 1980; McCarthy & Donchin, 1981). The effect of a subject's a posteriori uncertainty (i.e,, equivocation) about the nature of a stimulus plays a major role in determining both the amplitude and latency of the resulting P300, For example, Johnson and Donchin (1978) described an experiJohnson, PfefFerbaum, and Kopell ment in which the discriminability of two feedback tones was manipulated. The results demonstrated that the amplitudes of the P3OOs elicited by both stimuli were directly related to their discriminability. As would be expected, P300 latency decreased as stimulus discriminability increased (Johnson & Donchin, 1985), They also found that positive feedback elicited P3OOs that were both earlier and larger than those elicited by negative feedback, A number of theories of recognition memory assume that items placed in memory are tagged in some way with a continuous variable that provides a measure of the subject's familiarity with the test item (Atkinson & Juola, 1973; Craik & Tulving, 1975; Mandler, 1980; Ratcliff, 1978). It is reasonable to equate familiarity with an item with the ease with which the item can be discriminated from irrelevant items. In addition, some of these memorytheories hold that increases in the familiarity value for an item are accompanied by increases in the speed with which that item can be identified. Given the relationship between P300 and stimulus discriminability, this ERP component is particularly well suited to studying memory processes. Thus, one would predict that any manipulation of the familiarity dimension would result in changes in both the amphtude and latency of the P300, Several ERP experiments have been devoted to determining the relationship between measures of P300 and memory processes. In the Stemberg paradigm (1969), P300 latency varies with memory search time in the same way as RT: P300 latency in-set (target) items have shorter latencies than outset (distractor) items (Adam & Collins, 1978; Ford, Roth, Mohs, Hopkins, & Kopell, 1979; Ford et al., 1980; Gomer, Spicuzza, & O'Donnell, 1976). Moreover, Ford et al, (1980) reported that while changes in stimulus discriminability (i,e,, normal vs, degraded stimuli) do not affect the slopes of the RT and P300 latency functions, the intercepts for both measures increased for the degraded stimuli (cf Sternberg, 1969), P300 amplitude differences have also been observed in this paradigm. For example, Gomer et al. (1976) reported that targets elicited larger P3OOs than distractors. It is reasonable to assume that items or events that are subsequently recognized are somehow processed differently from those that are not. Testing this hypothesis has been difficult due to the nature of behavioral measures of information processing. That is, whereas reaction time is useful for quantifying the timing of memory retrieval processes, it is not well suited to illuminating the processes involved in memory acquisition when subjects must study a list of items. However, because P300 amVol. 22, No. 5 plitude and latency index the extent and timing of stimulus processing, it is uniquely capable of providing insights into these mental operations. The purpose of the present experiment was to assess P300 activity during word acquisition as a function of subsequent performance in a long-term recognition memory task. Since recognition is a function of discriminability, we predicted that words that were better recognized during tests would elicit larger and earlier P3OOs during memory acquisition. To evaluate stimulus processing during the acquisition as well as the retrieval phases of memory, we used the study-test procedure (Murdock & Dufty, 1972; Ratcliff, 1978). In this paradigm, subjects are presented with a list of words to be memorized (Study series) and subsequently tested for recognition on a list comprised of the target words randomly mixed with an equal number of new, distractor words (Test series). By repeating the Study and Test series with the same list of target words, the effects of repetition on learning and memory were also assessed.