COMPUTATIONAL STUDY OF COMPENSATION

补偿的计算研究

• 批准号: 2053081
• 负责人: David Kirsh
• 金额: $ 14.57万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN DIEGO
• 依托单位国家: 美国
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-04-10 至 1998-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/2053081
• 关键词: aging; attention; behavior test; cognition; human old age (65+); human subject; mathematical ability; mathematical model; memory; neural information processing; perception; performance; psychological models; stimulus /response; videotape /videodisc; young adult human (21-34)

The proposed research is aimed at expanding our knowledge of how people compensate for their cognitive limitations, particularly as they age. Many actions are best understood as compensatory, they are undertaken because they reduce the memory and computational demands agents must cope with if they are to seed at their tasks. Our current research on college age students suggests that such compensations occur even in environments, such as computer arcade games, where quick response is required. We hypothesize that: (1) careful empirical and theoretical study will reveal that older people also perform compensatory actions in fast paced environments to reduce internal cognitive demands; (2) compensations spontaneously develop, although they may also be taught; (3) the differential value of compensations teaches us about the relation between computational strategies and cognitive architectures. To test our hypotheses we shall run experiments and construct computer models to simulate the consequences of lesioning key capacities. The primary experimental task is Tetris, a fast paced computer game that involves making quick decisions about rotating and translating simple shapes related to shapes described in the literature on mental rotation. We have constructed a special computational environment for automatically collecting Tetris data at millisecond levels and for creating arbitrary Tetris stimuli. Experiment I examines if the tendency to rotate video shapes in the world, rather than rotate images of them in the head -- a valuable compensatory strategy -- is a tendency that crosses age and motor performance. Experiment II examines if older subjects learn more slowly than young subjects and if their learning rate is sensitive to the initial pace of the game. Additional study of age-related performance will reveal if there are statistically significant differences in the style of play, such as, the average time to first touch the keys, average number of extra rotations, average drop height, percentage of zoids dropped, and so on. Experiment III tests if real time compensating strategies can be learned from explicit instruction. Experiment IV explores via simulations the question whether certain cognitive architectures are more robust to the cognitive deficits typical of aging. We study how the selective control of resources differentially effects performance in each of the models we shall build. In Experiment V we use experiment and theoretical analysis to discover if there are domain specific categories and concepts that human experts operate with in playing Tetris. In Experiment VI we confirm the general idea that interactive, compensatory strategies, outside the Tetris domain, can help agents improve performance, by observing if subjects can recognize letters from randomly oriented parts of those letters faster if they are allowed to rotate those parts. The relevance to health and the aging is that the more we understand how deficits can be compensated for the better we can design user friendly environments and devices for the aged; calibrate the workplace to keep pace with declining capacities, and teach the aged and cognitively limited to compensate for their deficits.

拟议的研究旨在扩大我们对人们的了解 弥补其认知局限性，尤其是随着年龄的增长。 最好将许多行动理解为补偿性，他们是进行的 因为它们减少了记忆力，并且代理必须应付计算需求 如果他们要播种他们的任务。 我们目前关于大学的研究 年龄的学生建议即使在环境中也会发生这种薪酬， 例如计算机街机游戏，需要快速响应。 我们 假设：（1）仔细的经验和理论研究将揭示 老年人还在快节奏中执行补偿性行动 减少内部认知需求的环境； （2）补偿 自发发展，尽管也可能被教导； （3） 补偿的差异价值教会我们有关 计算策略和认知体系结构。 测试我们的 假设我们将运行实验并构建计算机模型 模拟病变关键能力的后果。 主要 实验任务是Tetris，这是一款快速节奏的电脑游戏，涉及 快速决定旋转和翻译简单形状 与文献中描述的有关心理旋转的形状有关。 我们有 为自动构建了特殊的计算环境 以毫秒级收集俄罗斯方块数据，并用于创建任意 俄罗斯螺旋刺激。 实验我检查了是否旋转视频的趋势 世界上的形状，而不是旋转它们的头部 - 一个 有价值的补偿策略 - 是越过年龄和运动的趋势 表现。 实验II检查较老的受试者是否学习得更慢 比年轻学科，如果他们的学习率对初始学科敏感 游戏节奏。 与年龄相关的表现的其他研究将揭示 如果游戏风格在统计上存在显着差异， 例如，平均触摸钥匙的平均时间，平均额外数量 旋转，平均下降高度，zoid的百分比下降等等。 实验III测试是否可以学习实时补偿策略 从明确的说明中。 实验IV通过模拟探索 疑问某些认知架构是否对 典型的认​​知缺陷。 我们研究选择性控制如何 资源在每个模型中都会差异地影响性能 应建造。 在实验V中，我们使用实验和理论分析 发现是否有特定的类别和概念 人类专家在打俄罗斯方面工作。 在实验vi中，我们确认 交互式，补偿性策略的一般思想之外 俄罗斯方块域，可以通过观察 受试者可以识别这些主题的信件。 如果允许它们旋转这些零件，请更快。 相关性 对健康和衰老，我们越了解赤字如何 补偿我们可以设计用户友好环境的越好 老年人的设备；校准工作场所以保持步伐与下降 能力，并教年龄和认知限制以补偿 他们的赤字。