时空干扰不对称中量值信息的贝叶斯整合

发布时间：2018-06-23 18:43

本文选题：时空干扰不对称 + 量值理论　；参考：《西南大学》2017年硕士论文

【摘要】：时间和空间存在紧密联系。在心理学研究中,研究者主要关注时间加工和空间加工的相互影响。空间距离有空距离和实距离之分,空距离即相继呈现的离散刺激之间的距离,如点与点之间的距离;实距离即连续刺激形成的距离,如线段的长度。与此相对应,时间也有空时距和实时距两种形式,空时距与空距离相对应,即只有起止刺激,以间隔的方式呈现的时间;实时距与实距离相对应,即实距离从出现到消失之间持续的时间。对于空时距-空距离的研究表明,时间加工和空间加工之间存在相互干扰,典型的例子是kappa效应和Tau效应;在实时距-实距离的研究中,空间影响时间加工,时间却不影响空间加工,二者之间存在着不对称的干扰。现有研究对于这种不对称干扰现象的解释主要有两种理论取向:量值理论和隐喻理论。量值理论认为时间和空间同属一个广义量值系统,二者只是表征这个系统的维度,而时空干扰不对称现象的原因可能是实验的任务等无关因素,并不能真实反映二者之间的关系;与此相反,隐喻理论则认为时空干扰不对称现象则是由于日常生活中感知运动经验和语言的使用造成的,个体更多地使用空间词去描述时间词,语言中时间和空间的不对称扩展到了基本感知觉领域,使得个体对于空间距离的加工影响了其加工时间,而时间加工却不影响个体的加工空间距离。先前有关时空干扰不对称现象的研究多采用复制法,以秒以上的刺激呈现时间进行实验。由于复制法要求被试在行为上主动参与,当呈现时间较长时复制法容易受到记忆等因素的影响,因此在有些条件下其准确性低于简单的时距辨别任务。认知分段综合模型认为,时间认知具有分段性,不同的时间范围对应不同特点的计时机制。因此有必要扩大时间范围,以秒以下的时距作为刺激呈现时间,并以比较法进行实验,考察时距信息是否影响空间加工。其次,隐喻理论虽然能够解释时空干扰不对称现象,个体在复制时距时会受到无关的空间量值信息的干扰,但是空间量值信息是以怎样的形式影响个体最终确定的时间量值呢?量值信息的贝叶斯整合的观点启发我们可以利用贝叶斯估计对时空干扰不对称现象进行更深入的解释。我们用两个实验探讨上述的两个问题。实验一采用比较法和秒以下的刺激呈现时间考察时距信息对空间距离加工的影响。结果发现,比较时距的长短不能显著影响被试对于线段长度的知觉。这与先前研究的结果一致,即时距长短不会影响个体的空间距离知觉。实验二依然采用比较法进行实验,我们扩大了线段长度的范围至九种,基于隐喻理论构造了贝叶斯模型;并对理想观察者模型进行了改进,将个体的时距判断任务分为三个阶段:信号输入阶段,贝叶斯整合阶段,决策阶段。信号输入阶段中,物理刺激量值转换为心理量值时有两种可能的表征形式:线性表征和对数表征。因此我们构造了线性表征模型和对数表征模型,通过模型拟合来考察哪种模型能够更好地解释空间对时间加工的影响。实验二的结果分为两部分:首先采用和实验一同样的回归分析方法考察线段长度对时距知觉的影响,结果发现线段长度显著影响个体的时距知觉,符合隐喻理论的假设,此结果和实验一的结果一起再次证实了时空干扰不对称现象的存在;其次,数据拟合发现,对数表征模型能够更好地拟合实验数据。通过两个实验,我们得出以下结论:时空干扰不对称现象确实存在,且不受实验方法和刺激呈现时间的影响;以量值信息的贝叶斯整合对时空干扰不对称现象进行解释,对数表征模型更符合个体的行为表现;结合模式控制模型并对理想观察者模型进行改进,能够对时空干扰不对称现象进行更为深入的解释。
[Abstract]:In psychological research, the researchers focus on the interaction between time processing and space processing. Space distance has an empty distance and a real distance, and the distance is the distance between discrete stimuli, such as the distance between points and points; the real distance is the distance formed by a continuous stimulus, such as a line segment. Length. Corresponding to this, time also has two forms of space time distance and real time distance. The space time distance corresponds to the empty distance, that is, only starting and stopping stimulation, the time that is presented in a spaced way; the real-time distance corresponds to the real distance, that is the real distance from the appearance to the disappearance. There is mutual interference between inter processing, the typical examples are the kappa effect and the Tau effect. In the study of real distance real distance, space affects time processing and time does not affect space processing. There are asymmetric interference between the two. There are two kinds of theoretical orientations for the interpretation of this asymmetric interference phenomenon: quantity theory The theory of metaphorical theory. The theory of value value holds that time and space belong to a generalized value system. The two is only the dimension of the system, and the reason for the asymmetry of time and space interference may be unrelated factors such as the task of the experiment, and can not truly reflect the relationship between the two, and the metaphor theory, on the contrary, thinks that the interference of space and time is not correct. The phenomenon is caused by the perception of sports experience and the use of language in daily life. Individuals use more space words to describe time words. The asymmetry of time and space in language extends to the field of perception, which makes the processing time affected by the processing of space distance, while time processing does not affect the individual. Processing space distance. Previous studies on spatiotemporal disturbance asymmetry mainly use replication method, with more than second stimulus time to experiment. Because the replication method requires the participants to participate actively in behavior, the replication method is easily affected by memory and other factors when the time is longer, so the accuracy is lower in some conditions. A simple time distance discrimination task. The cognitive segmentation comprehensive model holds that time cognition is piecewise and different time ranges correspond to the timing mechanism of different characteristics. Therefore, it is necessary to expand the time range, take the time distance below second as the stimulus time, and to verify whether the time distance information affects the space processing. Although metaphorical theory can explain the asymmetry of spatiotemporal interference, the individual is disturbed by the unrelated spatial value information at the time of replication, but in what form does the spatial value information affect the individual's final time value? The view of Bayesian integration of the amount of information can be inspired by Bayesian estimation. We use two experiments to discuss the two problems mentioned above. In experiment one, the effect of time distance information on spatial distance processing is investigated by comparing the time interval of the comparison method and below the second. The results show that the length of the distance can not significantly affect the perception of the length of the line. The results of previous studies agree that the length of the distance does not affect the spatial distance perception of the individual. Experiment two still uses the comparison method to carry out the experiment. We expand the range of the length of the line segment to nine kinds, construct the Bayesian model based on the metaphor theory, and improve the ideal observer model, and divide the individual time distance judgment into three. In the phase of the signal input phase, the Bayes integration stage, the decision stage. In the signal input phase, there are two possible forms of characterization: the linear representation and the logarithm representation. Therefore, we construct the linear representation model and the logarithm representation model, and investigate which model can be better by the model fitting. The effect of space on time processing is explained. The results of experiment two are divided into two parts: first, the effect of line length on temporal distance perception is investigated by the same regression analysis method with experiment one. The result shows that the length of line segment significantly affects the temporal perception of the individual, which conforms to the hypothesis of metaphorical theory, and this result is again confirmed with the result of experiment one. Secondly, the data fitting shows that the logarithmic representation model can better fit the experimental data. Through the two experiments, we draw the following conclusion: the spatio-temporal interference asymmetry does exist, and is not influenced by the experimental method and the time of the stimulus presentation; the Bayesian integration of the amount of information on time and space The interference asymmetry is explained, and the logarithmic representation model is more consistent with the individual behavior performance, and the model control model combined with the ideal observer model can explain the spatiotemporal disturbance asymmetry more deeply.
【学位授予单位】：西南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842

【参考文献】