认知与发展心理学研究室知识库

    视觉工作记忆是指对短暂出现过的视觉信息进行存储加工的记忆系统，与诸如推理、问题解决、流体智力等高级认知能力有密切的关系，对人们日常生活中处理大量信息和同时完成多种任务有重要的作用；它既是一种重要的认知障碍指标，又是人才选拔中衡量个体认知能力的重要指标。虽然视觉工作记忆的影响如此之大，但其所能存储和加工的信息量却极为有限，这种有趣的现象引来研究者对其存储加工机制多年来的探讨。已有存储机制的模型争论集中在工作记忆精确度与广度的关系，近年来主要以成人为被试探索二者关系的研究得到了矛盾的结果。文献分析发现这可能是已有模型的比较忽略了被试自身认知资源差异对精确度和广度之间关系可能产生的影响，而且不同模型所适用的人群也可能有所不同；而认知资源的差异主要来源于个体自身认知资源的发展差异和个体间的智力差异，并且抑制控制能力在其中可能起到重要的促进作用。因此本研究从个体发展差异和个体间智力差异的角度出发，考察不同认知资源个体的工作记忆广度和精确度的特点及关系，并进一步考察抑制控制能力在其中可能产生的影响，以此来验证不同存储机制模型的假设。
研究一分别从横断比较和纵向追踪的角度逐步探讨6-15 岁儿童视觉工作记忆精确度和广度的发展关系；研究二则以智力超常儿童为被试，考察不同智力水平个体的工作记忆精确度和广度特点；研究三在研究一、二的基础上，采用新的实验范式考察不同年龄、智力水平个体的抑制控制能力对其工作记忆精确度和广度可能产生的影响。根据不同条件下工作记忆精确度和广度间的关系来检验两类模型的假设。
    研究一从个体发展的角度考察工作记忆精确度和广度的发展关系。实验1采用Corsi 组块任务、经典延迟反应任务和镜面延迟反应任务对6-15 岁儿童的工作记忆任务表现进行横断比较，结果发现视觉工作记忆广度在儿童12-13 岁时就已达到成人水平，视觉工作记忆精确度则在15 岁后才会接近成人水平；即使控制智力的影响，这种发展趋势依然显著存在。这说明视觉工作记忆精确度的发展不受视觉工作记忆广度发展的制约，并且儿童后期工作记忆能力的变化主要体现在工作记忆精确度的提升。实验2 对实验1 的被试进行了为期1 年的追踪研究，比较工作记忆精确度及广度的纵向发展趋势。实验2 与实验1 结果相似，发现视觉工作记忆精确度在工作记忆广度停止发展后仍呈线性加速发展趋势，而且工作记忆精确度的发展轨迹不受年龄差异和性别的影响。研究一的结果表明，视觉工作记忆精确度的发展并不受记忆插槽数量变化的影响；对被试的反应概率分析发现该年龄段工作记忆精确度的发展可能主要受抑制控制能力的影响。
    研究二从个体间智力差异的角度考察不同智力水平儿童的视觉工作记忆精确度及广度特点。实验3 的结果发现，智力超常儿童的视觉工作记忆广度与智力良好儿童无显著差异，但显著高于同龄的平均智力儿童，且与成人的表现无显著差异；超常儿童的视觉工作记忆精确度显著高于同龄儿童，而且在高认知负荷条件下超常儿童的视觉工作记忆精确度与成人的表现无显著差异。这说明智力超常儿童的工作记忆精确度显著高于同龄儿童，但智力超常和智力良好儿童的工作记忆广度无显著差异。研究二的结果表明，高智力个体的视觉工作记忆精确度要显著高于同龄的普通智力个体，分析发现这种表现差异源于对非目标刺激的反应抑制，但工作记忆广度对高智力个体间工作记忆能力的差异区分效度不高。这再次证明工作记忆精确度可以独立检测个体的工作记忆能力，其与工作记忆广度并非从属关系。
    研究三在研究一和研究二的基础上，进一步考察抑制控制能力对工作记忆精确度和广度的影响。实验4 通过出现在编码加工阶段或固化阶段的无关信息，来测量不同年龄、智力被试的工作记忆精确度和广度受抑制控制能力的影响。结果发现干扰信息对儿童或成人的工作记忆广度无显著影响；工作记忆精确度受无关信息的干扰，尤其是出现在固化阶段的无关记忆表征对儿童的影响程度最大；但儿童与成人抑制加工无关刺激的能力没有差异。研究三的结果表明，儿童抑制无关记忆表征能力的发展要晚于抑制无关刺激加工的能力，而前者是导致不同年龄和智力水平儿童工作记忆精确度产生差异的主要原因。
    本研究的结果表明视觉工作记忆精确度与工作记忆广度既相关又独立，工作记忆精确度的发展变化并不能由记忆插槽数量的变化解释，数量化模型无法合理解释现有结果，实验结果支持质量化模型。实验结果还表明视觉工作记忆精确度是比工作记忆广度更为敏感的认知能力指标，可以检测出个体间或个体发展过程中认知能力的细微变化，这挑战了仅以工作记忆广度为衡量指标的传统视觉工作记忆能力观点。本研究结果从个体认知资源差异的角度进一步加深了对视觉工作记忆的存储机制了解，为日后工作记忆训练提供理论依据和实践指导。

Visual working memory (VWM) refers to the limited-energy memory system which encodes and manipulates instant information. Complex cognitive abilities, such as reasoning, problem-solving and fluid intelligence are related closely to VWM. Slot model assumes that memory representation is stored individually in memory slots, and the number of slots could be considered as visual working memory capacity. In that view, working memory span should be expected to constrain working memory precision’s development. In contrast,qualitative model assumes that VWM is a limited-resource memory system, and its limitation cannot be presented in slots number. It also assumes that working memory precision, which would decrease by the increasing number of stimulus,should correlate with working memory span in a degree, not completely. Yet recent research deploying young adults as participants yield incongruent results.The analysis of previous results suggests that cognitive differences because of individual development and individual differences in intelligence may confuse the results.
In order to investigate relationship between working memory precision and span, here we used original and mirror delayed response task to measure children’s working memory precision and span in different age and intellectual levels. And irrelevant stimulus were used during encoding or insolidating phase of revised delayed response task to investigate inhibition’s influence on VWM’s development and individual differences.
Study1 comprised Experiment1 and Experiment2, which investigated how children’s VWM precision and span developed cross-sectionally and longitudinally. Exp1 found that VWM span reached adult level around age 12-13,and precision would not stop increase at least around 14-15 years’ old. The precision under high-load condition would develop into adolescence, and this pattern still exist even controlling for intelligence. It demonstrated that VWM precision developed independently from span and it took more time to become mature. Exp2 further tracked the participants from Exp1 for 1year and found the similar results longitudinally. VWM precision still developed even after span stopped which also demonstrated that precision was a more sensitive indicator for cognitive development in late-childhood. Study1 showed that VWM precision did grow influenced by the inhibition rather than span.
Study2 compared VWM precision and span among different intellectual
level children. Exp3 taking the same tasks from Study1 found that intellectually gifted children’s span did not significantly differ from peers with good intelligence, yet precision was much higher. In high-load conditions, gifted children’s precision was even similar to adults’ level. This suggested that VWM span was more easily failed to detect the difference of cognitive individualdifferences for high-intelligence group of children. VWM precision, however, is a validated cognitive indicator without the limitation of span.
On the basis of Study1 and 2, Study3 further examined how inhibition would influence children’s VWM precision and span. Exp4 changed the time phase where distractor appeared, appearing at the encoding or insolidating phase of the new revised task to see how VWM precision and span of individuals with different age and intellectual levels would change. Results showed that distractors did not affect children’s or adults’ span, but it damged precision, especially in insolidating phase. Additionally, children’s ability to prevent from encoding distracting stimulus was similar with adults, yet the ability to retrieve with distracting representation was much worse. This demonstrated that development of inhibition in insolidating phase was probably the main reason for VWM’s development and individual differences during late childhood.
In general, this study finds that VWM precision is not dependent on span which makes slot model’s hypothesis not reasonable to explain the results. Hence,the results here support the qualitative models. Moreover, precision turns out to be a more sensitive and validated cognitive indicator than span for late-childhood and high-intelligence children. Traditional view of VWM capacity only taking span as indicator should be reconsidered. Additionally, this study helps better understanding the possible storage mechanism of VWM and results could provide valuable suggestions for future VWM evaluation and training.