The Reliability Of Dot-Probe Word Tasks In Clinical Research Of Depression

Background:The classical dot-probe task is an experimental paradigm,and the attentional bias index based on reaction time is widely used in the clinical research of emotional disorders.Although foreign studies have found that in the clinical study of anxiety disorders,the attentional bias indicators of dot-probe emotional pictures task are lacking in reliability,whereas whether or not dot-probe task attention bias indicators,in particular the indicators of emotional word dot-probe task have good reliability in clinical research in depression has not been confirmed by research.The purpose of this study was to investigate the reliability of attention bias indicators of Chinese emotional word dot-probe task in the clinical study of depression.Method:Individuals with depression and depressive symptoms were selected as subjects in this study.Study 1 selected adults(ie,college students)with depression(n=20)to complete the neutral-negative word pair dot-probe task test,10 of whom completed 13 times(probably every other day)retest.Study 2 selected college students with depressive symptoms(n=67)to complete 3 types of word pairs(ie,negative-neutral,positive-neutral,and neutral concrete-neutral abstract word pairs),and 17 of them completed the neutral-negative word pair dot-probe task test,completed 13 times(probably every other day)retest.The two studies separately calculated the split-half reliability,parallel-forms reliability,and test-retest reliability of the traditional attention bias index.The study also explored various aspects factor such as the duration of stimulus presentation(ie,1000 ms and 2000 ms),different pairs of emotional words(eg,neutral-negative,neutral-positive pairs of words),abnormal value processing methods,and attentional fields that influences the the reliability of attention bias index.Therefore,it can provide suggestions for improving the reliability of the attention bias test.Results:1)The reliability of attention bias in the presentation time of 2000ms neutral-negative word pairs:The split-half reliability of depression(study 1)and depressive symptoms(study 2)individual was r=-.41 and r =-.03,respectively;The parallel-forms reliability was r =-.09 and r =-.09,respectively;and the test-retest reliability of the average of all measures ICC was.05(P =.40)and-.55(P =.74),respectively.2)Reliability of attention bias in positive-neutral words pairs with 1000ms and 2000ms presentation time:The split-half reliability of depressive symptoms(study 2)individuals was r =-36 and r =.06,respectively;The parallel-forms reliability r =-.16 and r =.19,respectively.3)the split-half reliability of attention bias by different outlier processing methods:the half-reliability of the traditional outlier and rescale outliers retention processes for depressive individuals(Study 1)is r =-.41 and r ?-1.01,respectively.After three-week retesting,the split-half reliability of traditional processing was r =-.76;the split-half reliability of rescale outliers was r =-1.51.the half-reliability of the traditional outlier and retention processes for depressive symptoms individuals(Study 2)is r =-.03 and r =-.36,respectively.After three-week retesting,the split-half reliability of traditional processing was r =.64;the split-half reliability of rescale outliers was r =.59.4)Upper and lower visual field split and half reliability:depressive symptoms individuals with lower vision field of attention bias split-half reliability range was-.49-.50;upper vision field of attention to the split-half range-.83-.35.Conclusion:The reliability of the indicators of the attention bias in dot-probe words task has no consistency and stability in depression and depressive symptoms individuals,reliability coefficient is in unacceptable range;different outlier processing methods,different attention fields,different design features have little effect on reliability of the dot-probe words task.How to improve the reliability of dot-probe task is an important challenge of attention bias in clinical studies.