Influence of spatial perception abilities on reading in school-age children

Abstract Spatial perception abilities enable individuals to explore a visual field, to detect spatial position and to infer relationships between visual stimuli. Written words and text are conceptualized spatially along a horizontal mental line, but little is known about the way children develop these representations. The exact relationship between visuo-spatial perception and academic achievement has never been directly assessed. Therefore, our aim was to study the developmental trajectory of space perception abilities by assessing perceptual, attentional and memory components, the relationship between these abilities and reading achievement in school-age children. Forty-nine children aged between 6.5 and 11 years old were divided into four age groups and were assessed with visual bisection, visual search and visual memory location tasks. The results showed that the groups of older children, from the age of nine, improved significantly on the bisection and visual search tasks with respect to all visual fields, while the groups of younger children showed more errors in the left visual field (LVF). Performances on these tasks were correlated with reading level and age. Older children with a low reading score showed a LVF bias, similar to the youngest children. These results demonstrate how abnormal space perception might distort space representation and in turn affect reading and learning processes.


ABOUT THE AUTHOR
For some years now, Sc. Arnaud Saj has studied the perception and the representation of space in patients with right brain lesions, with or without spatial neglect, in studies with and without rehabilitation. He has a high level of expertise in both clinical neuropsychology and research. This work has led to several clinical publications and confirmed that spatial neglect patients suffer from an impaired internal representation of space with an ipsilesional deviation of their egocentric reference. Recently, in the MRI protocol in patients and healthy controls, he explored the neural correlates of attentional deficits and changes after prism therapy.

PUBLIC INTEREST STATEMENT
This is the first study examining the developmental trends of visuo-spatial perception abilities and their relation to reading performances. This study demonstrates the correlation between the age-related improvement of space perception and the development of the left visual field (LVF), allowing the coverage of a larger peripheral visual area. This is reflected by higher accuracy in spatial exploration and visual attention, which are important for spatial organization when processing visual cues. In terms of application, this study shows the importance of specific spatial perception abilities for learning to read. More precisely, it shows the existence of a relation between reading levels and spatial perception processing abilities for the LVF. On a neuronal level, results show a "bias" towards the LVF in young children and in older children with poor reading performances, which could be related to the immaturity of the contralateral fronto-parietal network of the right hemisphere.

Introduction
Left-to-right asymmetry is common across several cognitive domains (e.g. language, reading and time perception, for a review, see Boroditsky, 2011). It can be observed during performances on visual tasks involving spatial perception and attention abilities. This asymmetry is educationally reinforced in cultures that read and write from left to right. For example, the left-to-right axis forms the basis for systems representing time and numbers. Spatial perception abilities enable the processing of spatial information by exploring stimuli in the visual field, detecting their spatial position and the relationships between them. These visuo-spatial processing abilities are important for the quality of spatial analysis and spatial organization that, in turn, are fundamental to the development of several specific visuo-spatial abilities, such as spatial attention, orientation, memory and spatial imagery (Barisnikov, 2009;Booth & Siegler, 2006;de Hevia, Izard, Coubart, Spelke, & Streri, 2014;Piazza, Fumarola, Chinello, & Melcher, 2011;Smith & Chatterjee, 2008). These abilities constitute spatial cognition, which is essential for the development of spatial representations, allowing the individual to perform complex spatial tasks (e.g. pattern construction and drawing) and to acquire academic skills (e.g. reading, geometry and numerical skills).
When reading, a word is identified via the central visual acuity system; however, the process of reading also involves the recognition of cues in peripheral areas close to the central visual field (CVF). Therefore, the time of fixation on a word (or part of a word) includes an information extraction time necessary for reading, as well as a planning time for the saccade to the next word (Rayner & Pollatsek, 1981). Therefore, the simple task of reading is not only based on visual acuity but also on the perception of the space adjacent to the word. Basic numerical processing abilities also rely on these visuo-spatial processing abilities. Developmental studies have shown that ~10% of school children present visuo-spatial difficulties (e.g. perception, attention and memory) and learning difficulties that cannot be attributed to language development (Atkinson et al., 2003).
The visuo-spatial deficit hypothesis is a potential explanation for reading difficulties. More generally, the acquisition of reading is often reported to be impaired in many developmental disorders characterized by visuo-spatial deficits. A significant alteration in reading abilities (such as developmental dyslexia, see Mumeaux & Ducrot, 2014) has been observed in 5-6% of school-age children (von Aster & Shalev, 2007). Recent studies have demonstrated a causal link between visuo-spatial attention and learning how to read (Franceschini et al., 2013). Furthermore, it has been suggested that the orientation of visual attention is central to emerging reading abilities among children Hari & Renvall, 2001;Roach & Hogben, 2007;Vidyasagar & Pammer, 2010). However, few studies have explored spatial perception abilities and reading in this population, and the role of basic visuo-spatial processing abilities (attention, orientation and memory) in reading acquisition throughout development needs to be examined.
The main goal of the present study was to examine the relationship between space perception processing abilities and reading skills in school-age children. Based on the literature, we expected performances on spatial perception tasks (visual bisection and visual search) to be associated with reading level but not with visuo-spatial memory. We also expected developmental changes regarding asymmetry in the visuo-spatial perception field (left/right) to be associated with reading performances.

Participants
A total of 49 children, recruited from the same public elementary monolingual school in Geneva following mainstream education, took part in the study; none of the children repeated their academic year. All children were included in the analyses (24 girls; mean age, M = 8.69 years; standard deviation, SD = 1.42 years; age range: 6.5-11.20) and were divided into four groups. Each group comprised of 12 children part of the same class group with the same education level. The first group was composed of children from the third grade of primary education: 3P (M = 6.96; SD = 0.34; eight boys and four girls); the second group corresponded to the fourth grade: 4P (M = 8.25; SD = 0.66; six boys and six girls); the third group corresponded to the fifth grade: 5P (M = 8.91; SD = 0.39; six boys and six girls); and the fourth group corresponded to the seventh grade: 7P (M = 10.62; SD = 0.58; five boys and seven girls). The expected reading level was based on school grades ( Figure 1). Reading abilities were evaluated with an exam based on 21 points (cut-off at 14 points). The Switzerland reading score (1-6 points) was communicated to the school head master who obtained them from the class teacher in each assessed class (range: 1.25-5.75). There are five categories in the Swiss academic scoring system: 1-3.95 (not acquired); 4-4.45 (sufficient); 4.5-4.95 (good); 5-5.45 (very good); and 5.5-6 (excellent).
The Cantonal Authorities for Primary Education and the School Administration Authorities acquired the informed consent of the child participants through the adults (parents and teachers). The children participated as volunteers and could leave the study at any time. The study was approved by the Ethics Committee of the Department of Psychology (University of Geneva). This study was performed in accordance with the principles expressed in the Declaration of Helsinki.

Apparatus and procedure
The material consisted of triplet visuo-spatial processing task (adapted from Saj, Cojan, Vocat, Luauté, & Vuilleumier, 2013), which was administered individually, in a quiet room, by a trained psychologist. The participants faced the computer screen at a distance of approximately 50 cm and were instructed to respond as quickly and accurately as possible to all tasks.

Triplet visuo-spatial perception task
This visual paradigm is specifically designed to probe three distinct spatial processes, including perceptual, attentional and memory components, and to keep sensory inputs constant across different task conditions. The main characteristic of the task is the use of the same stimuli to assess these processes with three spatial tasks: a bisection task, a visual search and a spatial memory task.

Material
The stimuli were presented on a 19-cm colour monitor connected to a laptop computer running E-prime software. Visual stimuli consisted of simple shapes (i.e. triangles, squares or diamonds, all white on a black background, each at a visual angle of ~3° and presenting the same total surface area). At each trial, these shapes were presented in a row (triplet) of three items (for a duration of 1s), preceded by a fixation point at the centre of the screen (1 s) and followed by a varying inter-trial interval (3-6 s). In addition, to avoid predictability or systematic strategies, the position of the shapes on the screen was varied so that they could be presented in the left, right or CVFs (with either the rightmost, leftmost or central stimulus in the triplet being presented in the centre of the screen, respectively). These factors were independently manipulated for each trial in order to provide different tasks with the same set of stimuli ( Figure 2).

Task and procedure
Three tasks using the same visual stimuli were presented to the participants. Each task required a similar binary response (indicated by a key-press): (1) The bisection task: assessed spatial perception processing; participants were asked to indicate whether the central item was located on the midpoint between the two other items or not (possible responses: yes or no).
(2) The visual search task: assessed spatial attention processing; participants were asked to indicate whether the single-odd item in the triplet was a square or a diamond (possible responses: square or diamond).
(3) The visual memory task: assessed memory for spatial location (position); participants were asked to indicate whether or not the three items appeared in the same location as the previous trial (possible responses: same or different). Successive trials were controlled so that the position of items on trial n would correspond to positions on trial n − 1 on 50% of all trials (one-back task), irrespective of the actual shape (square, diamond or triangle) of the items Each task was presented in blocks of 30 trials (10 × 3 triplet positions on the screen) separated by short rest blocks (fixation point only), and repeated in a different order in two different sessions. This resulted in a total of 30 trials (10-min duration) with stimuli presented in the left visual field (LVF), the right visual field (RVF) or centrally (CVF) for each of the three tasks. All responses were recorded by a keypad for subsequent analyses of the participants' performances. The performances (corrects responses) for the three visual fields for each task were calculated separately.

Statistical analyses
Analyses of variance were performed with Statistica software (StatSoft, Maisons-Alfort, France) by considering the mean values of the correct response (CR). We analysed the influence of the factors group (3P, 4P, 5P and 7P), task (bisection, search and memory) and visual field (left, centre and right). In order to see at which age we could observe developmental changes in performances, post hoc comparisons were performed with the Newman-Keuls correction. The threshold for statistical significance was set at p < 0.05.
Regression and correlation analyses were conducted on individual performance scores for LVF, RVF and CVF for each task in function of the reading level (Swiss scoring system out of six) in order to respond to the hypothesis of asymmetry in visuo-spatial perception through development.

Figure 2. Example of trial used for the bisection, visual search and spatial memory tasks.
Note: The three tasks were performed in different blocks, each task requiring a similar binary yes/no response.

Results
The analysis of the CRs revealed a significant main effect of the factors group (p < 0.05), task (p < 0.05) and visual field (p < 0.05), and a significant interaction between group and visual field (p < 0.05). The group-task interaction was not significant (p = 0.461) but the group-task-visual field interaction was (p = 0.045).
The mean comparison between groups showed that the youngest group made more errors than the oldest group (% mean ± SD-3P: 67.3 ± 2.5; 4P: 73.1 ± 2.2; 5P: 77.8 ± 2.2; 7P: 81.9 ± 2.2). The 3P group showed a significant difference to the 5P and 7P groups (p < 0.05) but not with the 4P group (p = 0.08). The 4P group was significantly different to the 7P group (p < 0.05) but not to the 5P group (p = 0.14). Finally, the 5P group did not show any difference in comparison to the 7P group (p = 0.21).
The post hoc comparisons for each task showed that the performances of the groups of younger children (3P and 4P) were significantly different compared to the performances of the groups of older children for the bisection and visual search tasks in respect to visual field. For the bisection task (Figure 3(A)) and the visual search task (Figure 3(B)), the group of young children (3P) made more errors than the other groups, particularly in the LVF. The groups did not differ in the memory location task (Figure 3(C)).
We then explored correlations between the difference in performances in LVF, RVF and CVF in each task as well as the scores for each reading level. Significant relationships were found for the LVF and reading level since low reading levels correlated with LVF errors in the bisection (r = 0.50, p = 0.045) and search tasks (r = 0.56, p = 0.030). The group of youngest children made more errors in the LVF (Figure 4).

Discussion
To our knowledge, this is the first study that has shown the developmental trends of space perception abilities. It is also the first study to report a link between specific space perception biases and reading performances in a child population.
The triplet visuo-spatial perception tasks were used in order to study space perception abilities during development and to assess their possible link with reading capacity. In these tasks, the same stimuli were used to investigate three distinct visuo-spatial processing abilities (perceptual, attentional and memory). In the search task, the group of youngest children showed a significant bias towards the LVF, while no differences in performances were observed between the groups of older children with regard to different visual fields. On the bisection task, the performances of young children showed a bias towards the LVF and CVF. In contrast, on the visual memory task, no differences between groups and visual fields were observed. Furthermore, the groups' performances on the bisection and visual search tasks correlated with the children's academic grade and their individual reading scores. Increasing reading experience with age seems to improve space perception abilities. This dynamic relationship of mutual influence seems to result in significant improvements with age in space perception and reading abilities. Notes: The green line and the diamonds correspond to the 3P group; the blue line and the squares to the 4P group; the black line and the triangles to the 5P group; and the red line and the crosses to the 7P group.
In the first school years, it appears that children present "pseudo-spatial neglect" of the LVF. This bias seems to distort space perception and to disturb correct coding of visuo-spatial information, which affects the process of learning to read and causes increased reading errors. Maturation processes could explain the observed asymmetry for space perception and its developmental changes. In contrast, low reading performances exhibited by some older children could result from a specific deficit in spatial perception abilities. These children showed a bias towards the LVF, similar to the youngest group of children. Furthermore, their low scores for spatial perception and spatial attention were associated with low reading scores.
Correlations between poor attention and a bias towards line bisection have been associated with right hemisphere ineffectiveness (Manly, Cornish, Grant, Dobler, & Hollis, 2005). Developmental studies have also shown that the neuronal system (e.g. fronto-parietal regions) mediating spatial perception processing abilities develops early (Cohen & Cashon, 2001), and that neurological abnormalities in these regions underlie behavioural deficits observed in developmental dyslexia (Waldie & Hausmann, 2010) and dyscalculia (Rotzer et al., 2008).
Several studies have highlighted the role of attention in reading, demonstrating that visual attention skills are one of the most important predictors for reading skills (Bosse, Tainturier, & Valdois, 2007;Franceschini, Gori, Ruffino, Pedrolli, & Facoetti, 2012;Plaza & Cohen, 2007;Valdois, Bosse, & Tainturier, 2004;Valdois, Lassus-Sangosse, & Lobier, 2012). For example, a line bisection test (i.e. marking the centre of a horizontal line with a pencil) demonstrates potential deficits in the orientation of spatial attention in some dyslexic children. Typically, developing readers and adults reading from left to right tend to place their mark of bisection to the left of the true centre of the line (Jewell & McCourt, 2000). Nonetheless, there is no evidence for such a bisection bias in dyslexic children (Sireteanu, Goertz, Bachert, & Wandert, 2005).
With regard to developmental dyslexia, the majority of authors consider language deficits at the level of processing and representing phonological information to be important causal factors, although the precise nature of these deficits is not yet well defined (Castles & Coltheart, 2004;Ramus & Szenkovits, 2008). Simultaneously, there is a growing literature on visuo-spatial deficits being associated with developmental dyslexia (Cohen, Dehaene, Vinckier, Jobert, & Montavont, 2008), including oculomotor deficits (Getman, 1985;Jainta & Kapoula, 2011), visual attention deficits (Bosse et al., 2007;Facoetti, Paganoni, & Lorusso, 2000) and abnormalities in orientation perception (Meerwaldt & Vandongen, 1988;Romani, Ward, & Olson, 1999). Several recent studies have shown that individuals with poorer reading capacities present a deficit in visual search and coding tasks (Franceschini et al., 2012(Franceschini et al., , 2013Gori & Facoetti, 2015 for review). These authors demonstrated that visuo-spatial attention impaired before reading acquisition could be an explanatory cause of developmental dyslexia. Furthermore, this could be related to the abnormal development of visual magnocellular nerve cells involved in visual guidance control of attention and eye fixation (Stein, 2014). Therefore, difficulties in these abilities could be the consequence of a specific impairment of the magnocellular dorsal stream (Stein, 2014;Vidyasagar & Pammer, 2010 for review), which seems to be involved in learning to read Stein, 2014;Vidyasagar & Pammer, 2010 for review). Functional imaging studies in children with developmental dyslexia have shown deficiencies in the bilateral temporo-parietal junction areas (Eden & Zeffiro, 1998), with auditory and speech-sound processing for the left part (Gabrieli, 2009). On the other hand, the right temporo-parietal junction is the spatial attention region, responsible for spatial neglect (Corbetta & Shulman, 2011;Mort et al., 2003).
Specific spatial perception biases potentially associated with neglect dyslexia have been largely reported in adult patients with unilateral spatial neglect following (right) brain lesion. Reading errors exhibited by these patients seem to be related to the part of the visual field affected by neglect, usually the left part (for a review, see  of the text and/or words (Friedmann & Nachman-Katz, 2004;Lee et al., 2009). However, it has been shown that neglect developmental dyslexia may be confounded with learning disability (Friedmann & Nachman-Katz, 2004).
It is interesting to note that abnormalities in the parietal regions (for a review, see Vuilleumier, 2013) are more often associated with specific learning disabilities such as dyslexia or dyscalculia (Vallar, Burani, & Arduino, 2010). Therefore, in dyslexic patients (children or adults), functional imaging studies have shown reduced activation of the parietal lobe (hyper-or hypo-activated depending on the task), particularly in the left hemisphere (Lobier, Zoubrinetzky, & Valdois, 2012;Peyrin et al., 2012). Since parietal dysfunctioning may underlie dyslexia associated with a reduced visuo-attentional span, some authors propose that such a failure could impede the acquisition of effective word recognition (Peyrin et al., 2012). A decrease in grey matter volume, particularly in the parietal cortex (posterior superior parietal lobule and intraparietal sulcus, respectively) and the frontal cortex was observed in children with dyslexia or dyscalculia (Rotzer et al., 2008(Rotzer et al., , 2009Sun et al., 2010). Furthermore, several authors have observed alterations in white matter (especially at the upper and lower longitudinal beam) in subjects with dyslexia (Sun et al., 2010) as well as lesions of white matter tracts between different cortical areas present in signs of neglect (Bartolomeo, Thiebaut de Schotten, & Doricchi, 2007) and loss of reading skills.

Conclusion
Our results show that spatial perception abilities ameliorate throughout childhood and maintain a dynamic relationship during reading acquisition. The triplet visuo-spatial perception task proved to be a sensitive tool for defining specific space perception processing abilities involved in reading. This task, which is easy to administer, could help with the early detection of such difficulties in neurodevelopmental disorders associated with low academic achievement. This is in accordance with Franceschini et al.'s (2012) longitudinal study, which suggests the predictive value of visual search tasks for future reading acquisition and the early detection of dyslexia. Finally, diffusion tensor magnetic resonance imaging should reveal the structural maturation of white matter in several regions of the frontal and parietal lobes, areas in which white matter maturation is correlated with the development of spatial perception and reading processes. This preliminary study on spatial perception opens a new neuropsychological field on the link between spatial cognition and reading abilities in developmental psychology studies.