The Bayesian Integration Of Magnitude In The Spatio-temporal Asymmetric Interference

There is a close relationship between time and space.In the psychological study,the the researchers focus on the interaction between time processing and spatial processing.The spatial distance divides blank distance and full distance.The former means the length of the continued stimulus(such as the length of a line)and the latter means the distance between successive discrete stimulus(such as the distance between dot and dot).Correspondingly,the duration divides blank duration and full duration.The former is the presentation time of the blank distance and the latter means the duration of the full distance.The study about blank distance and blank duration show that space processing influences time processing and time processing influences space processing,and the typical examples are kappa effect and tau effect.Unlike this Symmetric interference,the study about full distance and full duration show that space processing influences time processing,but not vice versa.ATOM(a theory of magnitude)and MT(metaphor theory)are often used to explain this Asymmetric interference phenomenon.ATOM argues that space,quantity and time are linked by a common metric for action,and this Asymmetric interference results could be task-related artifacts.By contrast,MT argues that representations of time and number depend asymmetrically on representations of space.The claim that some domains tend to be asymmetrically dependent on others in our mental representations,which is at the core of MT,was originally supported by patterns in metaphorical language.First,the method of reproduction was often used to study the Asymmetric interference between space and time in previous studies,and the duration in the experiments was longer than one second.It should be noted that the method of reproduction in previous studies is easy to be affected by memory and other factors because this method requires the participants to participate actively in the experiments.According to the range-synthetic model,different time range correspond to different timing mechanism.As a result,it is essential to expend the time range and use different paradigm,such as discrimination task.Second,MT argue that people incorporated irrelevant spatial information into their temporal judgments even in purely non-linguistic tasks,which could explain the Asymmetric interference to a certain extent,but what MT can not tell us is that how the space magnitude influence the time estimation.A Bayesian perspective on magnitude estimation can used to explain the spatio-temporal Asymmetric interference more deeply.Two experiment were conducted to answer these two questions above.In experiment 1,discrimination task was used to investigate whether the phenomenon that space processing affected by time magnitude exists no matter which paradigm is conducted.The results showed that the duration of the line could not significantly affect the perception of the length.This is consistent with the results of previous studies,in which space processing influences time processing but not vice versa.In experiment 2,we used nine kinds of length and constructing Bayesian model based on the MT.A Bayesian model was modified from an ideal observer model for discrimination task,which include three stages: Sensory input,Bayesian integration,Response.In the first stage,a sample interval is measured and the corresponding cause-sensation function has two possible forms: linear representation and logarithmic representation.We constructed two corresponding models and fitted the models to the experiment data.Results show that space processing does influence time processing,which is consistent with the hypothesis of the MT,and the logarithmic model is superior to the linear model in data fitting.Through the two experiments,we deserve the results as follows: The Asymmetry interference phenomenon between spatial and temporal does exist,and it is not affected by the experimental method and the stimulus presentation time.Combined with the Mode control model,a Bayesian perspective on magnitude estimation can explain the Asymmetry interference more deeply,and the cause-sensation function is logarithmic representation.