Sage Journals: Discover world-class research

Abstract

Visual search difficulties are common in children with cerebral visual impairment (CVI), due to higher-order visual selective attention (VSA) deficits. However, little is known about children with CVI below 6 years. This international multi-centre study explored VSA through search performance and efficiency in preschool children aged 3–5 years with CVI (n = 24), or a CVI-risk (n = 20) compared with neurotypical children (n = 47). Search performance on the paper–pencil NEPSY Visual Attention task was measured by accuracy, commission errors, and completion time. Search efficiency was assessed by reconstructing the cancellation path to obtain inter-target distances, intersections, and cluster visits. Children with CVI demonstrated significantly lower accuracy, longer completion times, greater inter-target distances, and more revisits to clusters of targets compared with both CVI-risk and neurotypical children. We conclude that by using a modified approach of a paper–pencil search task, first signs of global and local VSA deficits can be detected, offering clinical insights.

Keywords

Cerebral visual impairment preschool children search efficiency visual search visual selective attention

Introduction

Imagine not being able to find your parents in a crowd, the cheese on the breakfast table, or your favourite toy in a toybox. These everyday challenges are often experienced by children with cerebral visual impairment (CVI), a brain-based visual disorder caused by damage to or abnormal development of the brain (Sakki et al., 2017). Although CVI can develop after postnatal brain damage, such as from brain tumours or traumatic brain injury (TBI), in most cases, the risk for CVI is identified around the time of birth. The most prevalent causes of CVI are hypoxic-ischemic encephalopathy (HIE; Good et al., 2001) and periventricular leukomalacia (PVL; Fazzi et al., 2004), placing children born preterm and children with cerebral palsy (CP) at an increased risk for developing CVI (Chau et al., 2013; Ego et al., 2015). CVI can have many manifestations, ranging from lower-order visual sensory deficits to higher-order visual processing deficits. The most profound higher-order deficits are those concerning visual selective attention (VSA), that is, difficulties in focussing on relevant visual elements in a scene while filtering out visual distractors. As a result, children with CVI often find it difficult or even impossible to locate people or objects, especially in visually crowded environments (Dutton et al., 2017; Lueck & Dutton, 2015; Macintyre-Béon et al., 2010) which may lead to frustration, anxiety, or disruptive behaviour (Dutton & Jacobson, 2001). Early identification of VSA or visual search deficits is beneficial for diagnostic accuracy and increases opportunities of providing early intervention and improving developmental outcomes.

Although infants already possess the basic perceptual abilities needed for visual search (Gerhardstein & Rovee-Collier, 2002), even young neurotypical children require help from caregivers to find objects, such as their favourite toy. Visual search tasks provide a structured way to study the development of these abilities by asking participants to locate specific items or targets among other distracting items. Search performance is typically assessed through the number of found targets (accuracy) and completion time (Treisman & Gelade, 1980). Neurotypical search performance continues to develop well into adolescence, with the most significant improvements occurring around 12 years of age and stabilizing just before adulthood (Woods et al., 2013). Compared with neurotypical peers, children with CVI tend to take longer to find targets and they miss or overlook them more often. In addition, their performance decreases significantly with increasing set size, less spacing between items and unstructured placing of items (Bennett et al., 2018; Geldof et al., 2015; Hokken, Stein, Kooiker, & Pel, 2024; Hokken, Stein, Pereira, et al., 2024; Manley et al., 2023; Zhang et al., 2022). Although some clinical search tasks, such as the Visual Attention subtest of the NEPSY (Korkman, 1998) or the Cancellation subtest of the Wechsler Preschool and Primary Scale of Intelligence (WPPSI; Wechsler, 2014), are designed for younger children, CVI-research has solely focused on school-aged children, adolescents, and young adults. As a result, little is known about the search performance of children with CVI below the age of 6, which limits our understanding of their early search deficits.

To increase insight into search behaviour, it is important to consider search efficiency as well, rather than focusing on search performance alone. For instance, when a child searches for a specific set of five toys in a toy box, systematically scanning the box and avoiding revisiting areas increases the likelihood of finding all toys more quickly. After finding one toy, an efficient search would mean finding the nearest undetected toy, rather than one further away. In contrast, signs of inefficient search have been observed in school-aged children, adolescents, and young adults with CVI. Both eye-tracking-based and paper–pencil tasks revealed that children with CVI searched over larger portions of the display (Hokken, Stein, Kooiker, & Pel, 2024; Hokken, Stein, Pereira, et al., 2024; Manley et al., 2023; Zhang et al., 2022) and tended to focus on one stimulus at a time (McDowell, 2020). In clinical practice, their search strategy is often described as ‘chaotic’, or unstructured, as they frequently fail to detect neighbouring targets and often revisit areas of the paper or display.

To date, search efficiency has not yet been studied in children with CVI. Here, we want to explore this through including the order of the found targets, the so-called cancellation path, into analysis. A first exploration to quantify search efficiency is by calculating the number of intersections within this path, the average distance between consecutively found targets (Woods et al., 2013), or the number of visits needed to cancel out a cluster of targets (Woods et al., 2013). The aim of this study was to investigate the search performance and search efficiency of children with a CVI-diagnosis, a CVI-risk, and neurotypical children using the NEPSY Visual Attention paper–pencil search task. We hypothesized that, in terms of search performance, children with CVI would find fewer targets and require more time to complete the task. For search efficiency, we expected their cancellation paths to show more signs of inefficient search, that is, more intersections, longer inter-target distances, and more revisits to clusters of targets compared with neurotypical children.

Method

Participants

We recruited children between 3 and 5 years of age from October 2022 to July 2024. Children with a medical risk of CVI, such as preterm birth, CP, PVL, traumatic brain injury (TBI), or other brain damage or abnormal brain development were recruited from (1) Royal Dutch Visio (a specialized centre for visually impaired and blind individuals), (2) Basalt Revalidatie (a Dutch expert centre for rehabilitation), and (3) the Department of Neonatology, Erasmus MC – Sophia’s Children Hospital. Ophthalmic and neurological data were extracted from their medical files. In addition, it was checked whether they had an official CVI diagnosis or CVI working diagnosis, the latter indicating a high likelihood of CVI, but acknowledging that further assessments may be required to confirm the diagnosis at an older age. Both diagnoses were established after a multidisciplinary assessment that led to visual rehabilitation. These children form the CVI group. The neurotypical group consisted of children from (1) the MARCS Institute for Brain Behaviour and Development of the Western Sydney University, (2) colleagues of Royal Dutch Visio, or (3) neurotypical siblings of the children in the clinical group. Children were excluded if they had a visual acuity below 0.1 decimal, a visual field smaller than 30 , or a diagnosed intellectual disability (intelligence quotient [IQ] < 70). Written informed consent was obtained from the parents of all included children. The study was approved by the Medical Ethical Committee of the Erasmus Medical Centre, Rotterdam (MEC-2020-0680) and the Western Sydney University Human ethics committee (approval #H14328). The principles outlined in the Declaration of Helsinki (2013) regarding research involving human subjects were followed.

Procedure and data analysis

The total experiment had a duration of 35 min and consisted of five computer tasks and two paper–pencil tasks. The NEPSY Visual Attention task was assessed at the end of the experiment. After the experiment, the child’s parents or caregivers were asked to complete the Vineland Questionnaire (VQ) online.

Vineland Questionnaire

The VQ (Sparrow et al., 2021) is a questionnaire that comprises three main domains: communication, daily living skills, and socialization. Together, these domains assess the adaptive behaviour skills of children. The outcome measure of the VQ is the Adaptive Behaviour Index (ABI), which is moderately associated (r = .46) with the IQ (Tassé & Kim, 2023).

NEPSY Visual Attention

The NEPSY Visual Attention subtest (Korkman, 1998) is a paper–pencil cancellation task and consists of two tasks: the Bunny and Cat search task (see Figure 1). In both tasks, children were presented with a paper that consisted of 96 pictures of animals or everyday objects such as apples, flowers, houses, and trains (1.2 × 1.2 cm in dimension). In the Bunny-Search task, 20 out of the 96 pictures were bunnies. All pictures were placed in structured rows and columns. In the Cat-search task, 20 out of the 96 pictures were cats. Here, the pictures were placed in an unstructured manner on the paper. Children were unaware of the number of targets (Bunnies or Cats) and were instructed to cancel out as many targets as possible with a maximum testing time of 3 min, using a marker. Opposite to the child, the assessor had the same paper on which she or he assigned each target a letter (A to T) to keep track of the target cancellation sequence. In addition, the assessor noted which hand the child used to complete the task and the viewing distance between the child and the paper. Both trials ended after 180 s or when the child indicated completion, either verbally or by putting down the marker.

Figure 1.

Visual representation and task set up of the NEPSY Visual Attention: bunny-search (left) and cat-search (right).

Data analysis

Search performance

Overall search performance was reflected by scores of accuracy, that is, the number of correctly cancelled out targets; commission errors, that is, the number of incorrect cancelled out targets; and the completion time.

Search efficiency

Search efficiency was mapped by reconstructing the cancellation path for each participant. We first determined the x and y coordinates (in pixels) for each target. Using Python 3 (Spyder), we calculated the mean inter-target distance, that is, the average distance (in pixels) between sequentially cancelled out targets. In addition, the number of intersections within each cancellation path was computed.

In both search tasks, two clusters of three neighbouring targets were present (Bunny-search: H-J-L, P-R-S; Cat-search: B-C-D, I-K-M, see Figure 1). For each cluster, we assessed how many targets were found by each child. If all three targets in a cluster were found, we analysed how many separate visits the child needed to cancel out all three targets. For instance, one visit means that the child cancelled out all three targets at once, while two visits means that the child cancelled out one or two targets, then cancelled out a target outside the cluster, before returning to complete the cluster.

Statistical analysis

We compared three groups of children: (1) CVI-group: children with a CVI (working) diagnosis, (2) CVI-risk group: children with a medical risk, and (3) neurotypical group. Non-parametric Kruskal–Wallis tests were used to compare the groups in terms of both search performance and search efficiency. Post hoc group comparisons were conducted using Mann–Whitney U tests with Bonferroni correction (α = .017) to account for multiple testing. Effect sizes were calculated with η² and were interpreted as small (η² < .01), moderate (η² > .06), or large η² > .14). For the CVI-group and CVI-risk group, Spearman correlation coefficients were computed to assess the linear relationship between visual search outcomes and adaptive behaviour and visual factors. All statistical analyses were carried out using SPSS Statistics package version 29.0.0.

Results

Group characteristics

In total, 52 children with a medical risk of CVI and 47 neurotypical (NT) children aged between 3 and 5 were included in the study. However, eight children with a medical risk of CVI were not able to complete the NEPSY Visual Attention task, due to fatigue (n = 1), lack of time (n = 2), an inability to understand the instructions (n = 3), or an inability to cancel out targets due to motor delays (n = 2). Of the 44 children with a medical risk of CVI, 24 children were placed in the CVI group as they had a (working) diagnosis of CVI, stated by a multidisciplinary team.

Table 1 presents the demographic, neurologic, and ophthalmic characteristics of the three groups. Children in the neurotypical group were significantly younger compared with the CVI group (Z = –2.074, p = .04) and the CVI-risk group (Z = –2.153, p = .031). There was no significant difference regarding sex distribution across groups (χ² = 4.24, p = .12).

Table 1.

Demographic, neurologic, and ophthalmic characteristics of the participating groups.

	CVI	CVI-risk	NT
Demographic findings
N	24	20	47
Age, M (SD)	4.9 (0.1)	5.0 (0.1)	4.5 (.1)
Girls, n (%)	7 (29%)	12 (60%)	20 (43%)
Hand-dominance
Right-handed	15 (63%)	18 (90%)	45 (96%)
Left-handed	7 (29%)	1 (5%)	2 (4%)
No Preference	2 (8%)	1 (5%)	0 (0%)
Neurologic findings, N (%)
Gestation age
Extremely preterm (<28 weeks)	4 (15%)	2 (10%)
Very preterm (28–32 weeks)	1 (4%)	7 (39%)
Moderate to late preterm (32–37 weeks)	8 (31%)	4 (22%)
Term (>37 weeks)	11 (46%)	7 (39%)
Cerebral Palsy	9 (38%)	2 (10%)
Periventricular leukomalacia	4 (15%)	0 (0%)
Brain tumour	2 (8%)	0 (0%)
Epilepsy	2 (8%)	0 (0%)
Ophthalmic findings, N (%)
Visual acuity
Low acuity (0.25–0.30)	2 (8%)	1 (5%)
Suboptimal acuity (0.30–0.60)	5 (21%)	1 (5%)
Normal acuity (0.60–1.50)	17 (71%)	18 (90%)
Rethinopathy of Prematurity	1 (4%)	1 (5%)
Hemianopia	2 (8%)	0 (0%)
Nystagmus	2 (8%)	2 (10%)

Search performance and search efficiency

Figure 2 shows two typical examples of the Bunny cancellation path of a neurotypical child (left), a child at CVI-risk (middle), and a child with CVI (right). The neurotypical child had an accuracy of 19 targets and made zero commission errors. The cancellation path consisted of a mean inter-target distance of 291 pixels, and one intersection. In both clusters, the neurotypical child found all three targets in one visit. The child at CVI-risk had an accuracy of 19 targets and made zero commission errors. The cancellation path consisted of a mean inter-target distance of 214 pixels, and one intersection. In cluster HJL, they found all three targets in one visit. In cluster PRS, they needed two visits to complete the cluster. The child with CVI had an accuracy of 16 targets and made zero commission errors. Here, the cancellation path consisted of a mean inter-target distance of 582 pixels and nine intersections. In cluster PRS, the participant found all three targets, but needed two visits, in cluster HJL they only found one target.

Figure 2.

Cancellation paths of a neurotypical participant (left: 4.3 years of age), a child at CVI-risk (middle: 4.11 years of age) and a participant with CVI (right: 4.7 years of age) on the bunny-search.

Table 2 represents the group means and group comparisons on search performance and efficiency for both tasks. Overall, significant differences were found in accuracy (Bunny-search: H = 13.17, p = .001; Cat-search: H = 9.98, p = .007), completion time (Bunny-search: H = 18.44, p < .001; Cat-search: H = 16.27, p < .001), and inter-target distance (Bunny-search: H = 9.72, p = .008; Cat-search: H = 9.41, p = .009). More specifically, compared with the CVI-risk and neurotypical group, the CVI group found fewer targets, took longer to complete the tasks, and had longer inter-target distances in their cancellation paths on both tasks (all ps < .016). In addition, the CVI group exhibited a shorter observed viewing distance compared with the neurotypical group (p = .001), but there was no significant difference in viewing distance between the CVI group and the CVI-risk group (all ps > .060). No group differences were found in the number of commission errors (Bunny-search: H = 4.37, p = .11; Cat-search: H = 0.34, p = .84) and number of intersections (Bunny-search: H = 1.58, p = .45, Cat-search: H = 0.09, p = .96).

Table 2.

Means, standard deviations and group comparisons of search performance and efficiency.

		CVI	CVI-risk	NT	CVI- CVI-risk	CVI-NT	CVI-risk–NT
Parameter	Task	M (SD)	M (SD)	M (SD)	p (η²)	p (η²)	p (η²)
Accuracy	Bunny	13 (1.3)	17 (.5)	17 (.6)	.013* (.20)	<.001* (.18)	.747
Accuracy	Cat	16 (1.1)	18 (.9)	18 (.5)	.016* (.13)	.003* (.12)	.881
Completion Time	Bunny	159 (7.5)	117 (11.6)	99 (6.2)	.016* (.14)	<.001* (.34)	.331
Completion Time	Cat	153 (7.9)	100 (9.1)	98 (7.2)	< .001* (.32)	<.001* (.25)	.817
Commission Errors	Bunny	2 (0.8)	0.26 (0.2)	0.66 (0.4)	n.s.	n.s.	n.s.
Commission Errors	Cat	0.63 (0.5)	0.44 (0.2)	0.71 (0.3)	n.s.	n.s.	n.s.
Intertarget distance	Bunny	526 (31.9)	428 (18.9)	443 (14.5)	.017* (.13)	.003* (.13)	.880
Intertarget distance	Cat	252 (14.9)	210 (9.5)	212 (6.2)	.010* (.16)	.005* (.12)	.782
Intersections	Bunny	5 (1.14)	4 (1.07)	6 (0.9)	n.s.	n.s.	n.s.
Intersections	Cat	8 (2.26)	5 (1.0)	5 (0.6)	n.s.	n.s.	n.s.
Viewing distance	Bunny	20 (2.1)	27 (2.1)	30 (0.6)	.060	<.001* (.22)	.222
Viewing distance	Cat	20 (2.4)	27 (2.1)	30 (0.7)	.079	.001* (.16)	.386

Figure 3 illustrates the distribution of children who successfully located all targets within a cluster, as well as the number of visits required per cluster. Overall, a smaller proportion of children with CVI (49%) were able to find all three targets in a cluster, compared with the CVI-risk (74%) and neurotypical children (80%). When they did find all targets, fewer children with CVI (45%) managed to locate them in a single visit compared with the other groups (CVI-risk: 72%, NT:74%). No notable differences were observed between the CVI-risk group and the neurotypical group.

Figure 3.

Distribution of target completion in clusters and number of visits across groups.

Effects of child characteristics on search performance and efficiency

In the CVI group, visual acuity significantly correlated with accuracy (r = .468, p = .002) and the number of intersections (r = .400, p = .009) but showed no significant correlation with completion time, inter-target distance, or viewing distance (all ps > .05). The VQ Adaptive Behaviour Index also significantly correlated with accuracy (r = .347, p = .026) but did not show significant correlations with other measures of search performance or search efficiency (all ps > .05). Boys with CVI made significantly more commission errors compared with girls with CVI (Z = –2.599, p = .009, η² = .63). In the CVI-risk group, both visual acuity and the VQ Adaptive Behaviour Index significantly correlated only with completion time (visual acuity: r = –922, p < .001; adaptive behaviour: r = –.459, p = .012), with no significant correlations found for other search performance or efficiency measures. No differences between boys and girls were found (all ps > .05). In the neurotypical group, boys had significantly longer completion times compared with girls (Z = –2.806, p = .005, η² = .73).

No significant differences were found between children with and without CP (in both the CVI and CVI-risk groups) for any of the search performance or search efficiency parameters (all ps > .05). No distinct search patterns were observed regarding the location of found targets in the four participants with a visual field deficit (e.g., hemianopia or Rethinopathy of Prematurity).

Discussion

The aim of this study was to investigate the search performance and search efficiency of young children with CVI, by comparing them to children with a medical risk of CVI and neurotypical peers on a search task with modified analysis. As hypothesized, children with CVI demonstrated significantly lower search accuracy, longer completion times, and less efficient search strategies, reflected by longer inter-target distances and more revisits to clusters. These results align with previous findings in children with CVI older than 6 years of age, which also show visual search deficits and challenges in efficiently scanning visual scenes (Bennett et al., 2018; Hokken, Stein, Kooiker, & Pel, 2024; Hokken, Stein, Pereira, et al., 2024; Manley et al., 2022; McDowell, 2020; McDowel & Butler, 2023; Zhang et al., 2022).

Visual selective attention deficits in children with CVI

The discrepancies in search performance and efficiency across groups reflect the visual selective attention (VSA) deficits often characteristic of children with CVI. Typically, VSA functions like a spotlight, focusing attention on specific parts of the visual field (Förster & Dannenberg, 2010; Liechty et al., 2003; Posner et al., 1980). In children with CVI, however, this attentional spotlight can either be too narrow, limiting their ability to oversee a full visual scene, or too broad, causing an overload of visual stimuli. These are known as global and local VSA deficits, respectively, and both impair visual search (van der Stigchel et al., 2009; Zuidhoek, 2015). Our study aligns with clinical and daily observations of older children with CVI (Atkinson, 2017; Hokken et al., 2025; Moore & Zirnsak, 2017; Philip & Dutton, 2014) and provides support that both global and local VSA deficits may be present in young children with CVI.

A global VSA deficit, characterized by a narrow attentional spotlight, limits the ability to process multiple visual stimuli simultaneously. In our study, this was observed in children with CVI who demonstrated a longer distance between two consecutive found targets and more revisits to clusters of targets (see the cancellation path of the child with CVI in Figure 2). These findings suggest that children with CVI focused on isolated individual targets, unable to perceive nearby stimuli simultaneously, consistent with a global VSA deficit. Similar behaviour has been observed in older children with CVI, who fixated on one stimulus at a time in a card-sorting search task (McDowell, 2020). Furthermore, eye-tracking studies have shown that children with CVI make significantly more eye movements within a larger area of the search display, further indicating that their attentional spotlight is too narrow to effectively encompass larger areas of the visual scene (Hokken, Stein, Pereira, et al., 2024; Manley et al., 2023; Zhang et al., 2022).

A local VSA deficit, where the attentional spotlight is too broad, causes children with CVI to process too many visual stimuli at once. This can make it challenging to select single targets between distractors. The shorter viewing distances that were found in children with CVI in this study might reflect an attempt to compensate for this overload. Importantly, this behaviour was not explained by visual acuity, suggesting that it serves as a strategy to consciously filter out irrelevant elements by moving closer to the target (Zuidhoek, 2015). Previous eye-tracking research, where viewing distance is typically kept consistent across all participants, has shown that children with CVI often miss targets even when fixating them (Hokken, Stein, Pereira, et al., 2024). A possible explanation is that this reflects an attentional spotlight that is too broad: without the ability to adjust their viewing distance, children with CVI process irrelevant visual details alongside the target.

In addition to local and global VSA deficits, other higher order visual function (HOVF) deficits in children with CVI warrant consideration. For example, although the absence of commission errors suggests that visual identification deficits are not present, children with CVI may still struggle to recognize a target bunny or cat, which could result in lower accuracy or longer completion time (Boot et al., 2010; Dutton & Jacobson, 2001; Manley et al., 2023). Furthermore, visual working memory deficits could explain their increased number of revisits to previously searched areas, as they may have difficulty retaining information about previously explored locations (Mannan et al., 2005). Finally, visual processing speed deficits could contribute to the longer completion times observed in children with CVI. In sum, although our findings point towards VSA deficits, it is important to acknowledge that multiple HOVFs may be affected. Therefore, a comprehensive multidisciplinary assessment of all HOVFs remains crucial in clinical practice for a thorough evaluation of CVI.

The role of age, cognitive and motor development, and visual factors on visual search outcomes

Age is an important factor to consider because search performance and efficiency typically improve as children grow older (Woods et al., 2013). In our study, neurotypical children were significantly younger than both the children with CVI and children at CVI-risk. This suggests that the differences in search performance and efficiency between children with CVI and neurotypical children may be even greater than observed. In addition, it may explain why the children with CVI-risk and neurotypical children showed no significant differences, potentially underestimating the visual search challenges faced by the children at CVI-risk.

Cognitive developmental delays are common in children with CVI and those at CVI-risk (Chokron & Dutton, 2023; Huo et al., 1999). Such delays could impact their visual search abilities or comprehension of the task. In this study, adaptive behaviour scores, which reflect cognitive developmental level or IQ, correlated with accuracy in children with CVI and with completion time in the CVI-risk group. These findings suggest cognitive delays may hinder search performance. However, the lack of association with search efficiency measures implies that visual processing challenges may have a higher impact on efficiency than general cognitive development per se.

Although we anticipated that children with CP might take longer to complete the task due to motor deficits or delays, no significant differences were found between children with and without CP. However, given the small number of children with CP in our study, and the exclusion of children who were unable to perform the paper–pencil task due to motor impairments (n = 2), no major conclusions can be drawn on the complex interplay of visual and motor impairments in preschool children with CVI.

Alongside higher-order visual challenges, some children with CVI experience lower-order visual deficits. For example, reduced visual acuity or visual field defects can hinder target detection in complex scenes. In this study, lower visual acuity was linked to lower accuracy and more intersections in the cancellation path, indicating poorer performance and a more disorganized search strategy. However, visual acuity was not associated with completion time, inter-target, or viewing distance. Notably, only a few children in the study had a low visual acuity of 0.25 decimal, which suggests caution in interpreting these findings. Similarly, the small number of children with visual field defects did not show distinct search patterns. These findings indicate that search inefficiencies in CVI are not solely due to lower-order visual deficits. Indeed, prior studies have also reported visual search deficits in children with CVI, even when acuity is intact (Chandna et al., 2021; Hokken et al., 2024; Manley et al., 2023; Zhang et al., 2022), highlighting the impact of higher-order visual deficits, such as VSA deficits.

Study strengths and limitations

This study is the first to explore visual search performance and search efficiency in children below 6 years of age and introduces a novel approach to assess search efficiency through cancellation path analysis.

One limitation is that several children (8%) were unable to participate in the study due to motor or language impairments, suggesting that the paper–pencil search task is not inclusive for all children with or at risk of CVI.

Moreover, the clinical group was divided into children with and without a confirmed CVI (working) diagnosis, to acknowledge that not all children with a medical risk will develop CVI. However, many children in the CVI-risk group did not receive a multidisciplinary CVI assessment, which may have introduced an inclusion bias. Still, these children were plausibly not referred to visual rehabilitation due to an absence of observable CVI symptoms in daily life or in previous general assessments. In line with this, children in the CVI-risk group demonstrated better search performance and efficiency than those with a confirmed CVI diagnosis, suggesting that observable deficits are more pronounced in children with a confirmed diagnosis or those experiencing notable visual challenges in daily life, who are more likely to be referred for diagnostic assessments. Finally, although the cancellation path provided valuable insights, it does not fully capture a child’s visual scanpath. Children with CVI may have fixated on nearby or clustered targets without recognizing them, and we lacked data on the time between target cancellations. This limits our ability to assess the timing of their search behaviour, which could have provided information on whether children are pausing, struggling, or hesitating during the search process, which in turn may be indicative of abnormal search.

Recommendations and practical implications

Although the medical risk for CVI is often identified around birth, most CVI assessments are not conducted until children are at least 6. Our study demonstrates that visual search tasks can be effectively used to assess children from 3 years of age. Although more research on CVI diagnostics in young children is needed, we recommend that these tasks can be employed to initiate the CVI assessments at a younger age. We also recommend that clinicians evaluate not only search performance but also search efficiency by reconstructing a child’s cancellation path. This approach offers insights into why children with CVI struggle to find targets or require more time, which can support the development of more tailored interventions. In addition, visual representations of the cancellation path could help parents and caregivers better understand their child’s daily visual challenges.

Besides analysing paper–pencil cancellation paths, a tablet could also be helpful to automatically track and reconstruct cancellation paths. Moreover, presenting search tasks on a monitor with an incorporated device to track eye movements further eliminates the need for verbal or motor responses. Previous work showed feasibility of this to nonverbally assess visual attention orienting and early-stage visual processing in young children with CVI risks (Kooiker et al., 2019; van Gils et al., 2020). Using eye-tracking technology could therefore also make higher-order visual assessments more inclusive for children with motor or language impairments, or for children below 3 years of age. Eye movement analysis could also provide a much more detailed and continuous scanpath both on and off targets, which enables a deeper investigation of individual search strategies in children with CVI.

Conclusion

This study showed that young children with CVI demonstrated poorer search performance and search efficiency compared with children at CVI-risk and neurotypical children. These findings suggest that underlying global and local VSA deficits emerge early in childhood and hinder effective visual search from a young age. Moreover, our results show that visual search tasks can be successfully used to assess children as young as 3 years, providing a valuable tool for early identification. By considering both search performance and search efficiency, clinicians can gain deeper insights into the unique deficits, organization, and strategies of search in young children with CVI. This could enhance our understanding of how children with CVI interact with visual environments.

Footnotes

Acknowledgements

The authors would like to thank the children and their families for their participation in this study. The current paper is part of the research project ‘Visual selective attention and processing deficits in children’, which is a collaboration of Royal Dutch Visio, Bartiméus, and Erasmus MC and is funded by the Visio Foundation, a Dutch non-profit organization providing financial support to (research) projects that improve the quality of life of individuals with a visual impairment (). They also thank Basalt Revalidatie, the Department of Neonatology of the Erasmus MC and the MARCS Institute for Brain, Behaviour and Development at the Western Sydney University for their collaboration, and Dee Ahn Sako, Karlijn Huinink, Mirjam Hulsman, Marthe Frank, Hesham Elbaz, Leonie van Keulen, and Lisa Taams for their help in data collection.

Author contributions

M.J.H.: Writing – original draft, Writing – review & editing, Visualization, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, and Conceptualization. S.V.: Writing – original draft, Writing – review & editing, Visualization, Formal analysis. C.G.: Writing – review & editing and Resources. P.E.: Writing – review & editing, Supervision, Resources, and Funding acquisition. M.J.G.K.: Writing – review & editing, Supervision, and Funding acquisition. J.J.M.P.: Writing – review & editing, Supervision, Resources, Methodology, Conceptualization.

Data availability statement

The authors do not have permission to share patient data.

Declaration of conflicting interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by the Visio Foundation (OI039055) and International Visiting Scholarship of the MARCS Institute for Brain, Behaviour and Development at the Western Sydney University. Data collection at the MARCS Institute for Brain, Behaviour and Development (i.e., Deeanh Sako’s salary and participant payment) was funded by an Australian Research Council grant (FT160100514) awarded to P.E.

Ethical approval and informed consent statements

Written informed consent was obtained from the parents of all included children. The study was approved by the Medical Ethical Committee of the Erasmus Medical Centre, Rotterdam (MEC-2020-0680) and the principles outlined in the Declaration of Helsinki (2013) regarding research involving human subjects were followed.

ORCID iDs

Marinke J Hokken

Silke Verboom

Christiaan JA Geldof

Paola Escudero

Marlou JG Kooiker

Johan JM Pel

References

Atkinson

(2017). The Davida Teller Award Lecture, 2016: Visual brain development: A review of ‘dorsal stream vulnerability’—Motion, mathematics, amblyopia, actions, and attention. Journal of Vision, 17(3), 26–26.

Bennett

C. R.

Bailin

E. S.

Gottlieb

T. K.

Bauer

C. M.

Bex

P. J.

Merabet

L. B.

(2018). Assessing visual search performance in ocular compared to cerebral visual impairment using a virtual reality simulation of human dynamic movement. In Proceedings of the Technology, Mind, and Society (pp. 1–6). Association for Computing Machinery. https://doi.org/10.1145/3183654.3183674

Boot

F. H.

Pel

J. J. M.

Van der Steen

Evenhuis

H. M.

(2010). Cerebral visual impairment: Which perceptive visual dysfunctions can be expected in children with brain damage? A systematic review. Research in Developmental Disabilities, 31(6), 1149–1159.

Chandna

Ghahghaei

Foster

Kumar

(2021). Higher visual function deficits in children with Cerebral Visual Impairment and good visual acuity. Frontiers in Human Neuroscience, 15, Article 711873.

Chau

Taylor

M. J.

Miller

S. P.

(2013). Visual function in preterm infants: Visualizing the brain to improve prognosis. Documenta Ophthalmologica, 127(1), 41–55. https://doi.org/10.1007/s10633-013-9397-7

Chokron

Dutton

G. N.

(2023). From vision to cognition: Potential contributions of Cerebral Visual Impairment to neurodevelopmental disorders. Journal of Neural Transmission, 130(3), 409–424.

Dutton

Chokron

Little

McDowell

(2017). Posterior parietal visual dysfunction: An exploratory review. Vision Development and Rehabilitation, 3(1), 10–22.

Dutton

G. N.

Jacobson

(2001). Cerebral Visual Impairment in children. Seminars in Neonatology, 6(6), 477–485. https://doi.org/10.1053/siny.2001.0078

Ego

Lidzba

Brovedani

Belmonti

Gonzalez-Monge

Boudia

Ritz

Cans

(2015). Visual-perceptual impairment in children with cerebral palsy: A systematic review. Developmental Medicine & Child Neurology, 57(Suppl. 2), 46–51. https://doi.org/10.1111/dmcn.12687

10.

Fazzi

Bova

S. M.

Uggetti

Signorini

Bianchi

P. E.

Maraucci

Zoppello

Lanzi

(2004). Visual–perceptual impairment in children with periventricular leukomalacia. Brain and Development, 26(8), 506–512. https://doi.org/10.1016/j.braindev.2004.02.002

11.

Förster

Dannenberg

(2010). GLOMOsys: A systems account of global versus local processing. Psychological Inquiry, 21(3), 175–197.

12.

Geldof

van Wassenaer-Leemhuis

Dik

Kok

J. H.

Oosterlaan

(2015). A functional approach to Cerebral Visual Impairments in very preterm/very-low-birth-weight children. Pediatric Research, 78, 190–197. https://doi.org/10.1038/pr.2015.83

13.

Gerhardstein

Rovee-Collier

(2002). The development of visual search in infants and very young children. Journal of Experimental Child Psychology, 81(2), 194–215. https://doi.org/10.1006/jecp.2001.2649

14.

Good

W. V.

Jan

J. E.

Burden

S. K.

Skoczenski

A. M.

Candy

T. R.

(2001). Recent advances in cortical visual impairment. Developmental Medicine & Child Neurology, 43(01), 56–60. https://doi.org/10.1017/s0012162201000093

15.

Hokken

M. J.

Stein

Kooiker

M. J.

Pel

J. J.

(2024). A novel gaze-based visual search task for children with CVI: A twin study. British Journal of Visual Impairment. Advance online publication. https://doi.org/10.1177/02646196241247973

16.

Hokken

M. J.

Stein

Pereira

R. R.

Rours

I. G.

Frens

M. A.

van der Steen

Pel

J. J.

Kooiker

M. J.

(2024). Eyes on CVI: Eye movements unveil distinct visual search patterns in Cerebral Visual Impairment compared to ADHD, dyslexia, and neurotypical children. Research in Developmental Disabilities, 151, 104767. https://doi.org/10.1016/j.ridd.2024.104767

17.

Hokken

M. J.

van der Zee

Y. J.

van der Geest

J. N.

Kooiker

M. J.

(2025). Parent-reported problems in children with Cerebral Visual Impairment: Improving the discriminative ability from ADHD and dyslexia using screening inventories. Neuropsychological Rehabilitation, 35, 54–74.

18.

Huo

Burden

S. K.

Hoyt

C. S.

Good

W. V.

(1999). Chronic cortical visual impairment in children: Aetiology, prognosis, and associated neurological deficits. British Journal of Ophthalmology, 83(6), 670–675.

19.

Kooiker

M. J.

Swarte

R. M. C.

Smit

L. S.

Reiss

I. K. M.

(2019). Perinatal risk factors for visuospatial attention and processing dysfunctions at 1 year of age in children born between 26 and 32 weeks. Early Human Development, 130, 71–79.

20.

Korkman

(1998). NEPSY. A developmental neuro-psychological assessment. Test materials and manual. Wiley.

21.

Liechty

Pieters

Wedel

(2003). Global and local covert visual attention: Evidence from a Bayesian Hidden Markov model. Psychometrika, 68(4), 519–541.

22.

Lueck

A. H.

Dutton

(2015). Vision and the brain: Understanding Cerebral Visual Impairment in children (Vol. 2). American Foundation for the Blind.

23.

Macintyre-Béon

Ibrahim

Hay

Cockburn

Calvert

J. N.

Dutton

Bowman

(2010). Dorsal stream dysfunction in children. A review and an approach to diagnosis and management. Current Pediatric Reviews, 6(3), 166–182.

24.

Manley

C. E.

Bennett

C. R.

Merabet

L. B.

(2022). Assessing higher-order visual processing in Cerebral Visual Impairment using naturalistic virtual-reality-based visual search tasks. Children, 9(8), 1114.

25.

Manley

C. E.

Walter

Micheletti

Tietjen

Cantillon

Fazzi

E. M.

Bex

P. J.

Merabet

L. B.

(2023). Object identification in Cerebral Visual Impairment characterized by gaze behavior and image saliency analysis. Brain and Development, 45(8), 432–444.

26.

Mannan

S. K.

Mort

D. J.

Hodgson

T. L.

Driver

Kennard

Husain

(2005). Revisiting previously searched locations in visual neglect: Role of right parietal and frontal lesions in misjudging old locations as new. Journal of Cognitive Neuroscience, 17(2), 340–354.

27.

McDowell

(2020). A pilot study of the Austin Playing Card Assessment: A tool to detect and find the degree of visual perceptual difficulties related to clutter. British Journal of Visual Impairment, 38(2), 118–136. https://doi.org/10.1177/0264619619896008

28.

McDowell

Butler

(2023). Validation of the Austin Assessment: A screening tool for Cerebral Visual Impairment related visual issues. PLOS ONE, 18(11), Article e0293904.

29.

Moore

Zirnsak

(2017). Neural mechanisms of selective visual attention. Annual Review of Psychology, 68, 47–72.

30.

Philip

S. S.

Dutton

G. N.

(2014). Identifying and characterising Cerebral Visual Impairment in children: A review. Clinical and Experimental Optometry, 97(3), 196–208. https://doi.org/10.1111/cxo.12155

31.

Posner

M. I.

Snyder

C. R. R.

Davidson

B. J.

(1980). Attention and the detection of signals. Journal of Experimental Psychology: General, 109(2), 160–174.

32.

Sakki

Dale

Sargent

Perez-Roche

Bowman

(2017). Is there consensus in defining childhood Cerebral Visual Impairment? A systematic review of terminology and definitions. British Journal of Ophthalmology, 102(4), 424–432. https://doi.org/10.1136/bjophthalmol-2017-310694

33.

Sparrow

S. S.

Cicchetti

D. V.

Saulnier

C. A.

(2021). Vineland-3-NL: Vineland adaptive behavior scales–third edition–Nederlandse bewerking: handleiding. Pearson.

34.

Tassé

M. J.

Kim

(2023). Examining the relationship between adaptive behavior and intelligence. Behavioral Sciences, 13(3), 252. https://doi.org/10.3390/bs13030252

35.

Treisman

A. M.

Gelade

(1980). A feature-integration theory of attention. Cognitive Psychology, 12(1), 97–136.

36.

van der Stigchel

Belopolsky

A. V.

Peters

J. C.

Wijnen

J. G.

Meeter

Theeuwes

. (2009). The limits of top-down control of visual attention. Acta Psychologica, 132(3), 201–212.

37.

van Gils

M. M.

Dudink

Reiss

I. K.

Swarte

R. M.

van der Steen

Pel

J. J.

Kooiker

M. J

. (2020). Brain damage and visuospatial impairments: Exploring early structure-function associations in children born very preterm. Pediatric Neurology, 109, 63–71.

38.

Wechsler

(2014). Wechsler Preschool and Primary Scale of Intelligence–4th ed. Manual Del 1. Svensk version (Manual part 1. Swedish version). NCS Pearson.

39.

Woods

A. J.

Göksun

Chatterjee

Zelonis

Mehta

Smith

S. E.

(2013). The development of organized visual search. Acta Psychologica, 143(2), 191–199. https://doi.org/10.1016/j.actpsy.2013.03.008

40.

Zhang

Manley

C. E.

Micheletti

Tesic

Bennett

C. R.

Fazzi

E. M.

Merabet

L. B.

(2022). Assessing visuospatial processing in Cerebral Visual Impairment using a novel and naturalistic static visual search task. Research in Developmental Disabilities, 131, 104364. https://doi.org/10.1016/j.ridd.2022.104364

41.

Zuidhoek

(2015). The role of attention and executive brain functions in seeing and behavior in children with CVI. In Hall Lueck

Dutton

G. N.

(Eds.), Vision and the brain (pp. 124–144). AFB Press.

Exploring visual search performance in preschool children with Cerebral Visual Impairment: A modified approach

Abstract

Keywords

Introduction

Method

Participants

Procedure and data analysis

Vineland Questionnaire

NEPSY Visual Attention

Data analysis

Search performance

Search efficiency

Statistical analysis

Results

Group characteristics

Search performance and search efficiency

Effects of child characteristics on search performance and efficiency

Discussion

Visual selective attention deficits in children with CVI

The role of age, cognitive and motor development, and visual factors on visual search outcomes

Study strengths and limitations

Recommendations and practical implications

Conclusion

Footnotes

Acknowledgements

Author contributions

Data availability statement

Declaration of conflicting interest

Funding

Ethical approval and informed consent statements

ORCID iDs

References