Effects of age and sex on eye movement characteristics

Abstract Abnormal eye movements are often associated with psychiatric disorders. Eye movements are sensorimotor functions of the brain, and aging and sex would affect their characteristics. A precise understanding of normal eye movements is required to distinguish disease‐related abnormalities from natural differences associated with aging or sex. To date, there is no multicohort study examining age‐related dependency and sex effects of eye movements in healthy, normal individuals using large samples to ensure the robustness and reproducibility of the results. In this study, we aimed to provide findings showing the impact of age and sex on eye movement measures. The present study used eye movement measures of more than seven hundred healthy individuals from three large independent cohorts. We herein evaluated eye movement measures quantified by using a set of standard eye movement tests that have been utilized for the examination of patients with schizophrenia. We assessed the statistical significance of the effects of age and sex and its reproducibility across cohorts. We found that 4‐18 out of 35 eye movement measures were significantly correlated with age, depending on the cohort, and that 10 of those, which are related to the fixation and motor control of smooth pursuit and saccades, showed high reproducibility. On the other hand, the effects of sex, if any, were less reproducible. The present results suggest that we should take age into account when we evaluate abnormalities in eye movements.


| INTRODUC TI ON
The current diagnosis of mental illness is based on subjective symptoms, and hence, it often depends on the experience of physicians.
To overcome this problem, there is an effort to establish and disseminate guidelines to minimize these variations across Japan 1 . In addition to educational efforts, it is important to establish biological measures that can help support diagnosis/treatment along with such guidelines. Recently, various candidate biomarkers of psychiatric disorders have been proposed [2][3][4][5][6][7][8][9][10][11][12][13][14][15] , and eye movement is one of those with the greatest effect sizes for distinguishing schizophrenia cases from healthy controls 6 .
Eye movement abnormalities in patients with schizophrenia have generally been studied using data obtained for adult subjects without a distinction of sex. However, brain structures are dependent on the age and sex of subjects, and age and sex have often been considered nuisance variables in studying the effects of disease [16][17][18][19] . Because eye movements are sensorimotor functions of the brain, they might also be dependent on age and sex, as are brain structures. Thus, we sought to determine whether we should consider age and sex as confounding factors when we study eye movement abnormalities. Some previous findings have suggested that aging would affect eye movement characteristics, including lower saccade velocity 20 , lower visual tracking performance 21,22 , and poorer antisaccade performance 23 in elderly subjects. However, other studies have shown a nonsignificant effect of age on saccade velocity 24,25 . To date, there has been no multicohort study of the age-related dependency and sex effects on eye movements in adults using large samples to ensure the robustness and reproducibility of the results.
Overall, a precise understanding of normal eye movements is required to distinguish disease-related abnormalities from natural differences associated with aging or sex. In this study, we examined the influence of age and sex on eye movement measures in healthy, normal individuals. We evaluated effects on eye movement measures obtained from a set of eye movement tests involving fixation, smooth pursuit, and free-viewing tasks that were previously utilized for examining abnormalities in schizophrenia 4,6 . The present study uses three large independent cohorts to evaluate the effects of age and sex and determine their reproducibility among cohorts, and the findings will help in the design of future eye movement studies for not only patients with schizophrenia but also those with other psychiatric disorders and neurological diseases.

| Subjects
Three independent cohorts of healthy participants with no history of psychiatric illness and no neurological disorders were involved in this study (

| Eye movement examinations and data collection
The details of the eye movement examinations and data collection have been described elsewhere 6,26,27 . For cohort OSK1, eye movement examinations consisted of 7 tests (three kinds of fixation tasks with no, near, and far distractor stimuli, three kinds of smooth pursuit tasks involving different stimulus motions including horizontal, slow Lissajous, fast Lissajous, and free-viewing tasks with 56 images). For cohort OSK2 and the multisite cohort, a smaller battery of eye movement tests was used (fixation task with far distractor stimulus, smooth pursuit task with a fast Lissajous stimulus motion and free-viewing task with 56 images (OSK2) or 20 images (multisite)). Thirty-five eye movement measures obtained from the fixation task with a far distractor stimulus, smooth pursuit task with a fast Lissajous stimulus motion, and

| Effects of age
In cohorts OSK1, OSK2, and the multisite cohort, 4, 18, and 8 out of the 35 eye movement measures, respectively, were significantly correlated with age (see Table 2). We then classified eye move-

| Effects of sex
No significant effect of sex was found in OSK1 and OSK2, though a significant effect was observed in the multisite cohort (Table 3).
Thus, seven eye movement characteristics of smooth pursuit were classified into category 3 based on the same criteria used for the effect of age.

| D ISCUSS I ON
In this study, we utilized data from three large cohorts to determine Half of the measures in categories 1 and 2 were those from the smooth pursuit test. There was a marked negative correlation with the horizontal and vertical velocity gains (ie the ratio between the eye velocity and target velocity). This means that the higher the age was, the slower the velocity of smooth tracking relative to target motion. Similar relationships between smooth pursuit and age have been pointed out in previous studies 21,22 . The present result supports their finding using much larger samples obtained from multiple institutes. We also found that the properties of catch-up saccades were also affected by age, such that larger catch-up sac-  Table 2), albeit weak, suggesting that fixation is weakly influenced by age, and hence, that Morita's integrated eye movement score may also be influenced, at least in part, by age.

| LI M ITATI O N S
The findings from this study are limited to those of healthy, normal individuals, and the effects of age and sex in patients were remained unresolved. Further studies are required to clarify age-and sexrelated effects on patients with psychiatric/neurological disorders.
As the range of age was limited from 16 to 79 in this study, additional studies are necessary to determine the early development of eye movements during childhood and aging in older seniors.

| CON CLUS ION
We conclude that not all but some characteristics of fixations, saccades, and smooth pursuit performance depend on age in healthy, normal individuals. Thus, we should take age into account when we evaluate disease-related abnormalities of eye movements.

| APPROVAL OF THE RE S E ARCH PROTO COL BY AN IN S TITUTI ONAL RE VIE W BOA RD
The study was performed in accordance with the World Medical Association's Declaration of Helsinki and was approved by the research ethical committees of all the institutions that joined this study (Kyushu Univ, Osaka Univ, Univ Tokyo, Nagoya Univ, Nara Medical Univ) and of the National Center of Neurology and Psychiatry.

| INFORMED CON S ENT
All participants provided written consent to the study after a full explanation of the study procedures.

ACK N OWLED G M ENTS
We appreciate the cooperation of all individuals who participated in this study. This research was supported by AMED under grant numbers JP18dm0207006, JP20dm0207069, JP20dm0307002, JP20dm0307102 and JP20lm0203007 and by JSPS KAKENHI grant numbers 20K06920 and 20H03611. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated and/or analyzed during the current study are not publicly available because they contain information that could compromise research participant privacy/consent.