眼动技术在个体认知能力差异研究中的应用

本文从可能性和可行性两方面介绍和论述了眼动技术在个体认知能力差异研究中的应用。在可能性方面，个体眼动特征具有的良好的重测信度和内部一致性信度；在可行性方面，已有研究利用注视和眼跳相关指标、眼动轨迹和瞳孔直径这三类指标对该问题进行探索，结果均发现这些眼动特征与认知能力之间存在着密切关系。今后研究应继续深入探索两者关系的内在认知和神经机制，并尝试结合人工智能算法开发高信效度的眼动认知测验。

The application of eye-movement technique in researching individual differences in cognitive abilities

Eye-movement technique has been widely used in the study of cognitive psychology. The application of eye-movement technique in researching individual differences in cognitive ability has also attracted researchers' attention. Previous studies have successfully identified individuals with psychological disorders (such as autism spectrum disorder and schizophrenia) by using eye-movement analysis. Whether eye-movement characteristics can accurately reflect the cognitive ability of normal people is also worth exploring. In this review, we first introduced the stability of individual’s eye-movement characteristics across time and trial. These studies indicate that eye-movement characteristics have good test-retest reliability and internal consistency reliability, which lays the foundation for using eye-movement characteristics to measure cognitive abilities. Subsequently, we systematically summarized the relationship between eye-movement characteristics and cognitive abilities from three aspects: fixation and saccade related indicators, scan patterns, and pupil size. These studies provided substantial evidence that eye-movement characteristics could reflect individual’s cognitive abilities such as intelligence and working memory. To be specific, individuals with high cognitive abilities show more efficient information processing in eye-movement tasks such as visual searching and reading, reflecting in longer saccade amplitude and lower proportion of long fixation duration. In intelligence test, individuals with high cognitive abilities tend to select the more efficient strategies according to test’s rules. Take Raven's advanced progressive matrices test as an example, high intelligence individuals spend longer time on encoding the problem and demonstrate less toggling rate between problem area and response alternatives area. In addition, studies focusing on scan patterns developed powerful quantitative analysis tools in recent years, which brings new vitality to qualitative scan patterns. These studies found that indicators based on quantitative scan patterns perform better than conventional fixation or saccade indicators in predicting individual’s cognitive abilities. Moreover, pupil size, a more physiological indicator indirectly reflecting the activity of locus coeruleus norepinephrine (LC-NE) system, also has close relations with cognitive abilities. However, the correlation between cognitive abilities and tonic (baseline) pupil size remained controversial for several reasons, such as the diverging definition of baseline and different measurements for cognitive abilities. The correlation between phasic (task-related) pupil size and cognitive abilities was close to agreement according to the control hypothesis. For unfamiliar tasks that require more exploration, individuals with high cognitive abilities tend to fully utilize their cognitive resources to complete these tasks, causing their phasic pupil size to dilate more that low cognitive abilities individuals. In contrast, for familiar tasks that require exploit knowledge, they tend to utilize fewer resources to efficiently complete these tasks, causing their phasic pupil size to dilate smaller than individuals with low cognitive abilities. Studies have revealed the close relationship between various eye-movement characteristics and cognitive abilities, indicating eye-movement technique is a promising tool for measuring individual cognitive differences with good reliability and validity. Future studies need to further explore the causal relationship, internal cognitive mechanism, and underlying neurobiological basis between eye-movement characteristics and cognitive abilities. Also, the influences of possible moderating variables (e.g., age, task type, task difficulty and physical properties of stimulus materials) should be fully considered in relative studies. Furthermore, it is worth exploring to develop reliable and valid cognitive ability tests combined with artificial intelligence algorithm based on eye-movement characteristics.