A Study Of Cognitive Component In Numerical Inductive Reasoning

Inductive reasoning is an inference from a particular instance of a general rule orrelationship process, and is one of the important higher cognitive functions of thehuman brain. Currently, the knowledge of its cognitive component is still little.Therefore, in the domain of numerical inductive reasoning, the study usednumber-series completion problems and carried out the following studies:Part1: An eye-movement study of cognitive component in numericalinductive reasoningBased on the reaction time and accuracy, previous behavioral study has proposedthat there are3sub-components involved in numerical inductive reasoning, i.e.,recognition of memorized series, calculation, and checking. Eye-movement technology has the advantages of reflecting the real time visual processing,thus, the present study aims at providing eye-movement evidences of the3sub-components in number-series completion problems. A2(series type:counting, non-counting)×3(rule effectiveness: true, invalid, anomalous) study wasperformed to15participants by Tobii T120eye tracker. The results showed that forthe counting series and true series, the fixation duration, fixation count, and thenumber of regression were significantly less than those of the non-counting series andthe invalid series, and the tasks of the role of3sub-components lack of similarityof eye movements. These findings are consistent with the hypothesis andfirstly provide the eye movement evidence for the existence of the3sub-componentsin numerical inductive reasoning.Part2: Computational cognitive modeling on the cognitive component innumerical inductive reasoningPrevious behavioral study has confirmed the existence of learning. Based on theeye-movement study, we further study the learning effect and the form of it about the3sub-components in numerical inductive reasoning. A2(series type:counting, non-counting)×3(rule effectiveness: true, invalid, anomalous) study wasperformed to9participants by E-prime. The results showed that for the s recognition,the form of learning complied with the constants model, for the calculationcomponent, the form of learning complied with the constants model, and for thechecking component, the form of learning complied the incremental model.Part3: An eye-movement data automatic analysis software Since Tobii T120eye tracker s eye movement data collected is complicated, andit is inconvenient to be analyzed. Therefore, in order to facilitate the analysis of TobiiT120eye tracker’s eye-movement data, and getting the corresponding results ofeye-movement，an eye-movement data automatic analysis software is developed. Inthe software, you can just by a simple operation to complete the analysis of complexdata. The software not only can display a single subject’s eye movement results, butalso can export to the eye-movement results in all subjects.In summary, the present study provides eye movement evidences of the3sub-components in number-series completion problems. Then, by a computationalcognitive modeling study, the evidence of learning effect is given. Finally, anautomatic analysis software of eye-movement data was developed.