AbstractBackgroundEye movement tasks, especially anti‐saccade tasks, have been used to assess cognitive function in patients with neuropsychiatric disorders. Although it has been shown that individuals with cognitive impairment perform worse on anti‐saccades tasks, there is a lack of systematic evaluation of the sensitivity of parameters of anti‐saccades to assess different subtypes of cognitive impairment.MethodsA total of 158 participants were enrolled in this study, consisting of 66 men and 92 women, with an average age of 50.2 ± 10 years. The comparison of pro‐saccade reaction time, anti‐saccade reaction time, and error rates in the saccade task between individuals with cognitive impairments and a normal group was conducted. Furthermore, we systematically analyzed the correlations between the performance in neurological function tests (Mini‐Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Stroop) and these anti‐saccade parameters. Especially, the correlation between these parameters and cognitive function in different domains of the MoCA task were also evaluated.ResultsThe pro‐saccade reaction time, anti‐saccade reaction time and error rate were negatively correlated with the MMSE and MoCA scores (P < 0.001), and positively correlated with the time used in Stroop tasks. Among them, the error rate had the strongest correlation with the performance of MMSE, MoCA and Stroop tasks (MoCA: P < 0.0001, r2 = −0.608; MMSE: P < 0.0001, r2 = −0.344; Stroop: P < 0.0001, r2 = 0.455). Among the seven cognitive domains examined by the MoCA task, error rates had relatively high correlations with visuospatial/executive (P < 0.0001, r2 = −0.4660) and delayed recall (P < 0.0001, r2 = −0.4228) compared to naming, language (P = 0.0004, r2 = −0.0788), attention (P = 0.0004, r2 = −0.0780), abstraction (P < 0.0001, r2 = −0.1515), orientation (P < 0.0001, r2 = −0.1075). Moreover, pro‐saccade reaction time, anti‐saccade reaction time and error rate of people with high MoCA scores were significantly higher than those of people with low MoCA scores, which can be used to identify people with mild cognitive impairment.ConclusionsOur study's results provide valuable clinical evidence supporting the effectiveness of anti‐saccades in assessing cognitive impairment, which is beneficial for screening and timely clinical intervention in individuals with specific cognitive impairment.