Abstract: PURPOSE: Although extensive research in neurophysiology, neuropsychology, and neuroanatomy suggests the involvement of widespread brain regions in ocular tracking, there is still a lack of anatomical evidence for the development of ocular tracking abilities. Our study aims to fill this gap by investigating the relationship between performance in ocular tracking an unpredictable target and gray matter volume in preadolescent children and young adults.
METHODS: We used an 8-minute ocular-tracking task in which participants tracked the step-ramp motion of a cartoon character (0.64°H × 0.64°V) with its speed (16°/s-24°/s) and direction (2°-358°) randomly varied from trial to trial. A total of 81 children aged 8-9 years (47 females and 34 males) and 77 adults aged 18-30 years (43 females and 34 males) completed the ocular-tracking task. Among them, 52 children (34 females and 18 males) and 72 adults (42 females and 30 males) had valid structural MRI data.
For the ocular-tracking task, we computed 12 oculometric measures to assess different aspects of ocular-tracking performance. We also combined the 12 oculometric measures to compute the ocular-tracking performance index that indicates the overall tracking ability. For the structural MRI data, we first obtained cortical grey matter volume using the Desikan Killiany atlas, with 34 cortical regions per hemisphere. We then transformed the regional cortical volumes into centile scores using the lifespan chart of the human brain derived from the largest MRI samples by far (Bethlehem et al., 2022). The centile score evaluates to what extent an individual deviate from the normative distribution of reference samples with the same sex and similar age. We assessed the developmental state of the 34 brain regions in our child and adult cohorts based on the age at peak regional volume from the lifespan chart. This enabled us to roughly evaluate whether those regions exhibited volume enlargement or reduction in both child and adult cohorts.
For the data analysis, we first examined whether children and adults differed on the ocular tracking performance metrics (i.e., performance index and the 12 oculometric measures). We then performed Spearman’s rank correlation analysis between the centile scores of the 34 brain regions and the performance metrics that exhibited intergroup differences. Regarding the individual oculometric measures, we explored three key questions: (1) Which brain regions are specifically involved in adults during unpredictable ocular tracking? To address this, we examined whether any brain regions showed significant correlations between volume centile scores and two or more oculometric measures exclusively in adults, not in children. (2) Which brain regions are specifically involved in children? To this end, we explored whether any brain regions demonstrated significant correlations between volume centile scores and two or more oculometric measures exclusively in children, not in adults. (3) Are there any brain regions that play a significant role in both adults and children? To address this, we investigated whether the volume percentile scores of any brain regions exhibited a significant correlation with at least one identical oculometric measure in both adults and children.
RESULTS: Children demonstrate inferior performance in the ocular tracking task compared to adults, as indicated by both the performance index and individual oculometric measures. Correlation analysis revealed distinct brain regions associated with the performance index in adults and children. In adults, the centile scores of the caudal middle frontal region (including DLPFC) and the pars opercularis in the frontal cortex specifically exhibited correlations with the performance index. In contrast, in children, the superior parietal region (including IPS and V3a) in the parietal cortex specifically showed correlations with the performance index. According to the age at peak regional volume from the lifespan chart, the brain regions associated with performance index in adults develop later compared to those in children. Nevertheless, the respective brain regions in adults and children all have reached a considerable level of development, i.e., in developmental stages characterized by cortical volume reduction. This aligned with our correlation results that for both children and adults, the smaller the volume percentile scores of the related brain region, the better the overall ocular tracking abilities (i.e., the larger performance index).
Correlation analysis on individual oculometric measures revealed a similar pattern. Specifically, the brain regions associated with oculometric measures in adults develop later compared to those in children. In adults, these brain regions were primarily located in the frontal cortex, including the caudal middle frontal, medial orbitofrontal, and frontal pole. In children, these brain areas were located in the pericalcarine of the occipital cortex and the isthmus cingulate. In both adults and children and, a smaller volume percentile score of the related brain region were corelated with better ocular tracking performance, such as shorter pursuit latency and larger proportion of smooth pursuit. These findings align with the fact that the respective brain regions in both adults and children are in stages of cortical volume reduction.
In addition to the distinct correlation patterns found in children and adults, the correlation analysis on oculometric measures also revealed certain brain regions that exhibit overlapping correlations in both age groups. These shared regions include precentral (including FEF) and superior temporal areas. Interestingly, the precentral were in different developmental stages in adults (i.e., reduction) and children (i.e., enlargement). Accordingly, in adults, a smaller volume percentile score of the precentral is associated with larger open-loop acceleration, whereas in children, a larger volume percentile score of the precentral is associated with larger open-loop acceleration. The developmental state in children contrasts with findings mentioned above that the respective brain regions in both adults and children entered the stage of cortical volume reduction.
CONCLUSIONS: This study represents the first investigation into the underlying neural basis of ocular tracking development. We found that the ocular-tracking brain network in adults includes regions that are absent in the network of children, and these regions generally develop later. Conversely, the ocular-tracking brain network in children contains regions that are absent in adults, and these regions usually develop earlier. This may be due to the fact that certain brain regions in children have not yet reached a considerable level of development. Consequently, the connections between these brain regions and ocular tracking performance observed in adults have not yet been established. Instead, children rely on brain regions that have already developed and are functional for them but absent in adults. Additionally, a critical brain region, i.e., the FEF responsible for generating oculomotor commands and having major outputs to subcortical regions for controlling ocular tracking behavior, exhibits a significant correlation with children's performance despite their incomplete development in this region. In general, the relationship between brain structure and ocular-tracking performance can be described as follows: if the cortical volume of these brain regions has not yet reached their peak, then the larger volume centile score, the better the ocular-tracking performance, and vice verse.

Key words: eye movements, structural MRI, cortical volume, brain development