Spatial abilities in Down syndrome: a comparison of mental rotation and perspective taking

Mental rotation and spatial perspective taking are two key spatial abilities and are integral to one's everyday life. This study compared performance on these two abilities between adolescents and young adults with Down syndrome (DS) and typically developing (TD) children. Twenty-three participants with DS and twenty-three TD children were matched using Raven's progressive matrices ability scores. Two experimental tasks measured each spatial ability, and combined scores were calculated for mental rotation and perspective-taking, respectively. Participants with DS did not differ significantly from their TD counterparts in perspective-taking. However, the groups differed significantly in mental rotation, with participants with DS performing more poorly than TD children. Despite this mean difference, the relation between non-verbal ability and each spatial ability was virtually the same for both groups. Our results supported dissociations between mental rotation and perspective-taking, as one showed a group difference, whereas the other did not. Further, possible strategic processing differences may help explain the group differences in mental rotation. Lastly, the similarities in developmental trajectory slopes between the two groups in mental rotation suggest that people with DS showed a delayed, rather than deviant, developmental pathway of mental rotation relative to TD children.

Down syndrome (DS) is the most common genetic syndrome resulting in intellectual disability (Dykens et al., 2000; Kittler et al., 2008), occurring in approximately 1 in 700 births (Parker et al., 2010). Individuals with DS generally do not exceed a mental age of around seven or eight during adulthood (Pennington and Bennetto, 1998). In addition to intellectual disability, DS is also characterized by speech and language impairments (Abbeduto et al., 2007). Children with DS exhibit delayed cognitive development (Weijerman and De Winter, 2010) and acquire personal/social skills later than most typically developing (TD) children (National Down Syndrome Society, 2007).

More recently, spatial abilities have been identified as delayed in people with DS (Bracken et al., 2024; Yang et al., 2014).

Spatial abilities aid in manipulating, constructing, and navigating the physical world (Newcombe and Shipley, 2015), enabling people to recognize and use relationships among different objects and between themselves and objects in their environment (Morris et al., 2023). These abilities are integral to carrying out daily activities involved in independent living, such as loading a dishwasher, packing a bag, and navigating to the store. Individuals with DS often exhibit stronger spatial abilities than verbal abilities (Davis, 2008). Nevertheless, research has indicated that some spatial abilities may represent a problem for individuals with DS relative to their overall cognitive ability (Morris et al., 2023; see Yang et al., 2014).

Different components of spatial abilities

Spatial abilities are usually identified in terms of a set of sub-domains of ability, rather than a unitary construct (Uttal et al., 2013). For this research, we followed the approach of Newcombe and Shipley (2015), who developed a theoretical framework to classify spatial representations using cognitive, neural, and linguistic approaches (Chatterjee, 2008; Palmer, 1978). In their framework, Newcombe and Shipley argue that spatial representations can be categorized along two dimensions: intrinsic vs. extrinsic and static vs. dynamic. Intrinsic representations reflect properties within-objects such as overall shape and relations between parts, whereas extrinsic representations reflect relations among objects that can involve understanding relative locations and establishing frames of reference (Hodgkiss et al., 2021; Newcombe and Shipley, 2015). Intrinsic and extrinsic spatial relations can further be represented either statically (objects in stable positions) or dynamically (objects in movement) (Chatterjee, 2008; Đokić and Vorkapić, 2024; Kozhevnikov et al., 2002, 2005). Based on these distinctions, the framework identifies four broad spatial skill categories: intrinsic-static, intrinsic-dynamic, extrinsic-static, and extrinsic-dynamic.

In the research reported here, we examined the abilities of people with DS in two sub-domains identified by this framework: intrinsic-dynamic abilities and extrinsic-dynamic abilities. The intrinsic-dynamic category involves transforming object-based spatial relations. One basic skill typical of this category is mental rotation, the ability to rotate a 2- or 3-dimensional object in one's mind (Shepard and Metzler, 1971). Research on mental rotation abilities in TD children suggests that these abilities emerge around 5 years of age (Iachini et al., 2019; Frick et al., 2013b). These abilities develop through middle childhood (ages 5–10), with accuracy of performance on tasks that require mental rotation influenced by the degree of rotation required (Frick et al., 2013a). Essentially, larger rotations generally result in less accurate performance. The extrinsic-dynamic category involves transforming the inter-relations of objects as one or more objects in the environment move, or the observer moves (Newcombe and Shipley, 2015). One extrinsic-dynamic skill is visual perspective taking, reflecting the ability to mentally represent a viewpoint different from one's own (Frick et al., 2014; Tarampi et al., 2016).

We chose to focus on mental rotation and visual perspective taking for two reasons. First, they are among the most frequently researched spatial abilities in typically developing children, and second, they have been identified as important for everyday skills (e.g., navigation, packing a bag). Hence, there is a wealth of data to draw on to understand how they operate and relate to one another in persons with DS. Research has indicated that both are dynamic spatial skills that require encoding spatial images and maintaining these representations in memory (Frick and Pichelmann, 2023; Hegarty and Waller, 2004; Kosslyn, 1994). Yet, although highly correlated (Hegarty and Waller, 2004), the two abilities differ in that mental rotation involves object-based transformations, such as imagining an object moving about an axis, whereas perspective taking involves egocentric transformations, changing one's point of view relative to one or more objects (Hegarty and Waller, 2004; Kozhevnikov et al., 2006). Further, neuroimaging studies provide evidence that perspective-taking and mental rotation engage distinct brain regions: perspective-taking involves the left parietal-temporal-occipital junction, and mental rotation involves bilateral activation of inferior and posterior parietal areas (Zacks et al., 1999, 2003). As a result, identifying similarities and differences between people with DS and TD children in these two abilities could yield important insights into the cognitive abilities related to everyday skills in people with DS.

Mental rotation and perspective taking abilities in Down syndrome

Studies on mental rotation abilities of individuals with DS have produced mixed results. Three studies have reported that people with DS exhibited poorer performance than their TD mental age-matched counterparts (Bracken et al., 2024; Hinnell and Virji-Babul, 2004; Meneghetti et al., 2018). Alternatively, other studies did not find a difference in mental rotation performance between people with DS and TD Children (Morris et al., 2023; Vicari et al., 2006). These discrepancies could be attributed to methodological differences. For example, all of the studies also used different stimuli (i.e., ghost shapes, letters, animals) with different degrees of rotation (i.e., 0–180 °; 0–315 °; 45–270 °), which makes it difficult to compare the assessments and whether or not they were measuring identical variants of the general ability (c.f. Brucato et al., 2023).

The literature on visual perspective taking in individuals with DS is also rather limited, yet the results are more consistent. Using Piaget's Three Mountain Task, studies found no significant difference in perspective taking performance between individuals with DS and their TD mental-age matched counterparts (Bracken et al., 2024; Hershkovich et al., 2023). Morris et al. (2023) used Frick et al.'s (2014) modified version of this task, which also revealed no differences in performance between groups. Despite the relatively consistent findings, more research is needed to better understand the development of visual perspective taking abilities in individuals with DS in conjunction with other spatial abilities. Therefore, our focus in the current study is on the association between mental rotation and visual perspective-taking spatial abilities in people with DS relative to typically developing children.

Current study

The overarching goal of the current study was to compare mental rotation and perspective taking abilities of adolescents and young adults with DS to non-verbal ability-matched TD children. Mental rotation and visual perspective-taking were each assessed using two tasks. To assess mental rotation abilities, a version of the Ghost Task (Frick et al., 2013b) and a Puzzle Rotation task designed by the researchers were used. To assess visual perspective taking, a version of the Three Mountains Task (Piaget and Inhelder, 1956) and a Dog Task adapted from Newcombe and Huttenlocher (1992) were utilized. By using two tasks per ability and non-verbal ability for group matching, our study provided a more stringent and robust test of the spatial profile than prior work that used a single task per ability and verbal ability for matching. Based on the most common results of previous literature, we expect to find that people with DS perform worse in mental rotation than TD children, but not necessarily in visual perspective taking.

In addition to evaluating similarities and differences in group mean performance, our approach includes a cross-sectional developmental trajectory analysis (see Thomas et al., 2009). Using this approach, we examine the relations between non-verbal ability, as measured by Ravens Matrices and the two general spatial abilities for the people with DS and the TD children, treating non-verbal ability level as the developmental variable. This type of analysis can provide insight into whether the development of mental rotation and perspective-taking abilities relates to non-verbal ability in the same way for individuals with DS and TD children. For example, a smaller slope for people with DS may suggest that non-verbal ability and the two spatial abilities are less closely related in people with DS than in TD children. In contrast, finding people with DS and TD children with different overall levels of performance onset (i.e., intercepts) may suggest differentiation early on. This approach of using developmental trajectories is similar to that of Morris et al. (2023), who also included cross-sectional trajectory analyses. However, they examined composites of intrinsic and extrinsic spatial abilities rather than comparing mental rotation and perspective-taking. Furthermore, they used verbal measures as their developmental variable.

The findings of the current study could have practical implications in designing training programs to enhance spatial abilities in individuals with DS. While spatial abilities in TD populations have proven to be malleable and can be improved through training interventions (see Uttal et al., 2013, for review), it is not clear whether the same holds for DS populations. Evaluating differences in multiple spatial abilities in people with DS may help determine whether these abilities can be improved through general spatial training or whether targeted approaches are needed for specific abilities.

Participants

Many participants in this study were part of a larger study examining spatial abilities in individuals with DS. The larger study required 4–5 months of participation (twice per week). Participants who completed the larger study received compensation valued at approximately $175 (monetary compensation at $100–$125 and a fire tablet valued at $50–$75).

The current study included 23 individuals with DS (14 males; Age: M = 18.49 years, SD = 4.02; range: 11–26) and 23 typically developing (TD) children (11 males; Age: M = 6.34 years, SD = 1.27; range: 4–8) who were individually matched on non-verbal ability using the Ability score of the Ravens Progressive Matrices. We over recruited the TD sample to match participants with DS, and here we included all the participants with DS who had completed the tasks. Participants with DS were recruited from various organizations, including the University of Alabama Intellectual Disabilities Participant Registry, DS Alabama, and local service providers. TD children were recruited from Children Helping Science, local schools, and social media groups. As reported by their parents, none of the TD children had any developmental delays. The mean Raven's scores did not differ significantly between the matched DS and TD groups, t (22) = −0.064, p = 0.48, indicating similar levels of performance in both groups.

Consent and assent were obtained before data collection. All the recruiting and testing procedures followed the IRB guidelines of both universities.

General procedure

Participants were individually tested over 2 days via Zoom using an Amazon Fire Tablet provided by the researchers. On the first day, they were administered the Peabody Picture Vocabulary Test, Fourth Edition (PPVT-4), and Ravens Progressive Matrices. On the second day, they completed the perspective taking and mental rotation tasks in the following order: Three Mountain Task, Dog Task, Ghost Task, and Puzzle Assessment. Both sessions lasted approximately 45 min to an hour.

Peabody picture vocabulary test, fourth edition (PPVT-4)

The Peabody Picture Vocabulary Test, Fourth Edition (PPVT-4) measures single-word vocabulary comprehension (Dunn and Dunn, 2007). Participants were shown four colored pictures labeled 1 through 4 and then heard a computer-read vocabulary word. They had to determine the picture that best represented the word's meaning. Participants were asked to verbalize or touch the correct picture. The PPVT Growth Scale Value (GSV) Score was used. GSV scores are not age-adjusted or standard scores; instead, they indicate one's absolute level of performance.

Raven's progressive matrices

The Raven's Progressive Matrices test is a non-verbal measure of general cognitive ability (Raven et al., 1998). Participants were shown a pattern with one piece missing and asked to choose which of the five possible answer choices completes it. Similar to the PPVT, participants were instructed to verbalize or touch correct responses. The ability score was used. Ability scores are not age-adjusted or standard scores; instead, they indicate one's absolute level of performance.

Perspective taking assessments

Two assessments were used to measure perspective taking. The first was a modified Three Mountain Task, adapted from Piaget and Inhelder's (1956) original task (see Hershkovich et al., 2023). This task was created using SuperLab. See Figure 1. The practice phase had three subsections. In subsection 1, participants were shown three cones of varying sizes and colors arranged at the center of a table. They were told that these cones represented “mountains”. Participants were then asked to identify the tallest, second tallest, and smallest mountain by color to ensure visibility of all three mountains. In subsection 2, participants were shown a series of short videos as if they were walking around the cones and stopped at 90 ° intervals. At each stop, the video paused, and then on the next screen, participants were presented with four different views of the mountains at the bottom of the screen and the end of the previous clip at the top (see Figure 1 for a similar layout). Participants were asked to select which of the four views they currently saw. If the participant answered correctly, they were told so and proceeded to the next trial. If they answered incorrectly, they were asked to try again until they answered correctly. This process was repeated until they had completed four trials, reflecting a full 360 ° rotation around the table. In subsection 3, participants were introduced to a stuffed animal (dog) and told to pretend to be the dog and pretend they could see what the dog would see. They were shown a short video clip of the dog being placed directly in front of them and asked to pick which of four views represented what the dog could see. This 0 ° trial with the dog placed directly in front of the participant was repeated to ensure that participants understood the task. Then the experimental phase started. The dog was placed around the mountains in 45 ° increments, ranging between 45 and 315 ° from the participant's perspective, resulting in 8 randomized trials. Each time the dog moved to a new position around the mountains, participants were asked to choose which of 4 answer choices reflected what the dog could see from that position. No response feedback was provided during the experimental phase. Each response was scored as follows: the correct response (3 points), the mirror image of the correct response (2 points), 180 ° (i.e., opposite) from the correct response (1 point) and the child's point of view (i.e., egocentric response; 0 points). There were 8 trials with a maximum possible score of 24 points.

The second perspective-taking assessment was the Dog Task, adapted from Newcombe and Huttenlocher (1992) (see Hershkovich et al., 2023). See Figure 2. Before the task, parents were supplied with red and blue wristbands to place on their child's wrists: red on the right wrist and blue on the left wrist. The wristbands were used to help distinguish between left and right. Upon the start of the experiment, participants were shown a screen displaying four stuffed dogs of different colors (black, white, brown, and spotted). The dogs were arranged at 0, 90, 180, and 270 °, forming a diamond shape. From the child's perspective, the black dog was placed directly in front of them, the white dog on their right, the spotted dog on their left, and the brown dog directly across from them and the black dog. Participants were told that they would pretend to be one of the dogs, referred to here as the “reference perspective.” A practice phase was included to ensure participants understood the task instructions. During this phase, participants were first instructed to pretend to be the black dog (i.e., their perspective) and asked to identify which dog was the furthest away, on the blue side (i.e., left), and on the red side (i.e., right). This was repeated with participants pretending to be the spotted dog (positioned 90 ° to the left from the participant's perspective). If participants responded incorrectly to the first question about the spotted dog (i.e., which dog is the farthest away), a short video clip played starting from the black dog's point of view and moving over to the spotted dog's point of view to demonstrate that the white dog was the farthest away. For all incorrect responses, participants were instructed to try again and, when appropriate, to use their wristbands to help them. Once participants correctly answered all 6 trials of the practice phase, the experimental phase began. The experimental phase mirrored the practice trials, except that no feedback was provided. This phase comprised 12 trials using a factorial design with 4 reference perspectives (black, white, brown and spotted dogs) and 3 direction questions (identifying the dog to the left, right, and farthest away). Once a reference perspective was determined, participants were asked all three direction questions before assuming the role of another dog. The scoring replicated that of the Three Mountain Task for a maximum of 36 possible points.

Although both tasks measure perspective taking, the Three Mountains Task is cognitively more demanding than the Dog Task because the scenes that require perspective taking are more complex in the former than the latter. Two tasks also differed in aspects such as response format (forced choice vs. open-ended) and degree of disparity (45 increments vs. 90 increments). Diversity across these two tasks helps examine the robustness of results on perspective taking.

Mental rotation assessments

Two assessments were used to measure mental rotation. The first assessment was the Ghost Task, adapted from Frick et al. (2013b). Before starting the Ghost Task, participants were given an instructional PowerPoint to ensure they understood how to perform the basic task over Zoom. First, they were trained in how to pick the correct choice. They were shown two objects (A Block Letter L and a Right Triangle) at the bottom of the screen. The shape on the left side was in a red box and the shape on the right was in a blue box. A silhouette of one of the shapes appeared at the top of the screen. They were asked to pick which of the shapes fit the silhouette by naming the color of the box in which it was presented. This was repeated for two trials. Participants were then trained to respond to shapes that were mirror images of each other. For this phase, they were presented the L and its mirror image, one in the red and one in the blue box. They were again asked to identify the color of the box that contained a shape matching a silhouette at the top of the screen once for the L and once for the mirror image. Next, they were trained to respond to rotated letters. The L was rotated at 90 ° in the first trial and 180 ° in the second. For the final part of the practice phase, trials using the 3 different ghosts and their mirror images used in the experiment phase were presented without rotation. Prior to proceeding with the experimental Ghost Task, participants were required to respond correctly on all practice ghost trials. Figure 3 presents the three ghost shapes and one example trial.

The experimental Ghost Task used the three ghost shapes. On each trial, participants saw a silhouette shape at the top of the screen and were asked to determine which of two ghosts (one of the three ghost shapes or its mirror image) matched the silhouette shape. Participants were asked to say “red” for the left response and “blue” for the right response. In the first 6 trials, participants were presented with mirror image ghosts of three different shapes facing either left or right, all rotated at 0 °. Participants needed to respond correctly to five out of the six trials before proceeding to the formal trials. If a participant failed to meet this criterion, they repeated these six trials. After the second round of practice, one participant with DS did not meet the criterion for completing the task and hence discontinued. The 0 ° trials were not included in the data analysis. Then we started the formal trials where each of the three ghost shapes were rotated 30, 60, 90, 120, 150, and 180 °, amounting to a total of 18 trials. All trials at lower angles of disparity were completed before moving to the next higher angle. The color and position of the correct answers were counterbalanced. Participants did not receive feedback once the rotated trials were introduced. Responses were scored as either correct (1 point) or incorrect (0 points), with a maximum of 18 possible points.

The second mental rotation task was the Puzzle Task, which we created and patterned after the Ghost Task. The task was informally piloted with young TD children, which demonstrated that it was sufficiently difficult while engaging. See Figure 4. In this task, participants were shown a puzzle and determined its missing piece (against its mirror image). Two puzzles were used: a Hot Air Balloon puzzle and a Gingerbread House puzzle. Similar to the Ghost Task, the puzzle pieces were presented in red and blue boxes. For each puzzle, there was a 0 ° rotation practice trial. Following this, there were 4 trials in which the pieces were rotated 45, 90, 135, and 180 °. The open space of the puzzle and puzzle pieces remained the same for all trials. Once participants completed all 4 trials for the Hot Air Balloon puzzle, they moved on to the 4 trials for the Gingerbread House puzzle. Responses were scored as either correct (1 point) or incorrect (0 points), resulting in a maximum of 8 possible points. The Puzzle task showed a high Pearson correlation with the Ghost Task, r = 0.59, p < 0.001, hence demonstrating its convergent validity in measuring mental rotation.

Although both tasks measure mental rotation, the Ghost Task might be cognitively more demanding than the Puzzle Task because it uses fewer visual features and has lower realism (Jansen et al., 2020; Snow and Culham, 2021). Two tasks also differed in aspects such as degree disparity (30 increments vs. 45 increments) and the presence of visuospatial context (absent vs. present). Diversity across these two tasks helps examine the robustness of results on mental rotation.

Descriptive statistics for all measures are reported in Table 1. To obtain a single measure for perspective taking and a single measure for mental rotation, we converted earned points to a percentage of the possible points for each task. Then we computed the average of the Three Mountain and Dog tasks to obtain a measure of perspective taking and the average of the Ghost and Puzzle tasks to obtain a measure of mental rotation. Group means and standard deviations based on these calculations are reported in Table 2. These data were used in both the analysis of means and correlations among measures.

The primary analysis was a 2 Group (DS and TD) × 2 Spatial Ability (Mental Rotation × Perspective Taking) ANOVA with Group as a Between-Subjects variable. The analysis revealed a Main Effect of Group, with TD participants outperforming DS participants, F (1, 88) = 8.33, p = 0.005, η 2 p = 0.086. There was also a significant Group × Spatial Ability interaction, F (1, 88) = 6.02, p = 0.005, η 2 p = 0.064. The Main Effect of Spatial Ability was not significant, F < 1.0. One-way ANOVAs indicated a difference between groups in Mental Rotation, F (1, 44) = 12.86, p < 0.001, but not in Perspective Taking, F (1, 44) <1.0. Overall, participants with DS performed significantly worse than TD participants in mental rotation but not in perspective-taking.

We conducted Pearson correlations among our various measures to evaluate possible group differences in how our verbal and non-verbal measures relate to the two spatial abilities. These are reported in Table 3. As can be seen in the table, all of the correlations were positive, with higher verbal and non-verbal scores paired with higher perspective taking and mental rotation performance. Both groups exhibited significant correlations between Ravens Ability scores and the two spatial ability measures. Only the typically developing participants showed significant correlations between PPVT GSV scores and spatial ability measures. Perhaps the typically developing children gained some added or redundant benefit from engaging verbal skills to complete the spatial ability tasks. Even so, the pattern of correlations does not directly help explain the group difference in mental rotation but not in perspective-taking.

To gain a better understanding of spatial ability of our groups, we used the cross-sectional developmental trajectory method described by Thomas et al. (2009) to examine the relation between the Ravens non-verbal ability measure and spatial abilities for the groups. In that approach, the ability measures were first converted to z -scores. Then, the analysis of covariance (ANCOVA) function within SPSS's General Linear Model was used to compare linear regressions across groups. We were specifically interested in whether Raven's ability had a different relationship with the performance variables (MR and PT ability) between the two groups. Thus, ANCOVA was used to test group differences in slope. With Ravens ability entered as the covariate, group differences in slope were indicated by the Group × Ravens Ability interaction. The cross-sectional trajectories of the DS and TD participants showing the relationship between Mental Rotation ability and Ravens ability scores, as well as between Perspective Taking ability and Ravens ability scores, are shown in Figure 5. As is clear from Figure 5 and the failure of the Group × Ravens interaction to yield significant effects, the increase in performance with increasing Ravens ability was essentially the same for both groups when comparing MR ability (p = 0.855) and when comparing PT (p = 0.789). Hence, the group difference in MR ability was maintained across the full range of Ravens ability scores represented in our sample.

The primary goals of this study were to evaluate similarities and differences in the mental rotation and perspective taking performance of people with DS and TD children with whom they were matched on scores achieved on the Ravens Matrices test. We used the average of two measures to assess each ability and found that the participants with DS performed worse than their TD matched counterparts on mental rotation ability but were very similar to them on perspective taking ability. To further investigate the differences between groups, we conducted cross-sectional developmental trajectory analyses relating Ravens' ability (as our developmental variable) with performance on the two abilities. Not surprisingly, as with the means, participants with DS and TD children exhibited a similar increase in perspective taking performance with increases in Ravens performance. Interestingly, even though the DS participants exhibited significantly worse performance on the mental rotation tasks in terms of means, the two groups exhibited a very similar positive relationship between Ravens performance and mental rotation performance.

The observation that group differences are evident in mental rotation ability and not in visual perspective-taking ability is consistent with much of the previous research conducted on people with DS. Visual perspective taking performance of people with DS seems to be consistently similar to that of TD children if they are matched on general non-verbal ability measures such as the Ravens matrices (Bracken et al., 2024; Hershkovich et al., 2023; Morris et al., 2023). In addition, our finding of a difference between the participants with DS and the TD children in performance on our two mental rotation assessments is consistent with several previous results (Bracken et al., 2024; Hinnell and Virji-Babul, 2004; Meneghetti et al., 2018) but not others (Morris et al., 2023; Vicari et al., 2006). However, the results of the new cross-sectional trajectory analyses may suggest a couple of hypotheses that can account for these discrepancies.

The developmental trajectory analyses indicated that, despite observing a group difference in mental rotation but not in perspective-taking, the rate of change associated with increases in non-verbal ability was essentially the same for participants with and without DS. It seems reasonable to consider the possibility that some factor other than general non-verbal ability must account for the observed differences. For example, it is well established that at least a portion of the differences observed between men and women in mental rotation performance can be attributed to cognitive strategies used to perform the task (Bersier et al., 2024; Boone and Hegarty, 2017; Doyle and Voyer, 2018; Goldstein et al., 1990; Hirnstein et al., 2009; Peters, 2005; Voyer, 2011). To the extent that this characterization applies to differences between people with and without DS, we might observe group differences in spatial tasks where strategies can be effective (e.g., mental rotation) but not in those where they are less effective. However, future studies would need to empirically test whether strategic differences help explain the results of our study.

With respect to visual perspective taking, it is important to note that finding similar levels of performance by people with DS and TD children on the perspective taking tasks is, at best, a relatively modest positive result. In this study, we found that adolescents and young adults with DS performed at the level of 4- to 9-year-old children (an average of approximately 6.5 years, using the Raven's Age Equivalent score) on basic tasks of visual perspective-taking. Given that perspective taking is important in many important skilled spatial activities involved in navigating and orienting oneself in the environment, performing at this level would clearly hinder performing a range of everyday behaviors that most of us take for granted. Other researchers have suggested a fundamental difference between understanding which object in one's frame of reference can be seen by another individual and knowing that two observers often have differing viewpoints of the same object (Flavell et al., 1981; Masangkay et al., 1974; Moll and Meltzoff, 2011; Moll and Tomasello, 2006). Hence, there is both a great deal to learn about perspective taking skills of people with DS and a great deal of room for improving those skills. Indeed, given that spatial abilities receive very little formal training in a typical school curriculum (National Research Council, 2006), we would anticipate that some type of formal training might be extremely beneficial for TD children as well as people with DS.

We believe that this study and the one reported by Morris et al. (2023) have demonstrated the value of examining both mean differences and cross-sectional developmental trajectories when evaluating similarities and differences in spatial abilities between people with and without DS. In both studies, it was observed that similarities in one measure do not ensure similarities in the other. In fact, training studies may benefit the most from this dual approach. Studying whether and how the relationship between the performance variable and the developmental variable changes may yield important information on the mechanism of training responsible for any observed improvement.

Also consistent with Morris et al. (2023), we found Newcombe and Shipley's (2015) framework for classifying spatial abilities valuable for evaluating the performance of typically developing children and people with DS. In this study, we found that performance differences were observed in one classification of abilities, intrinsic dynamic, but not in a second classification, extrinsic dynamic. Even though it is likely that the pattern of group differences is more complex than being in one category vs. a second, we think the Newcombe and Shipley framework provides a practical starting point for understanding those differences. Their double-dimension framework can be used to categorize basic spatial abilities and complex spatial tasks.

We compared the performance of Ravens ability matched people with DS and TD children on two fundamental spatial abilities: Mental Rotation and Visual Perspective Taking. We found a relatively large mean difference in Mental Rotation performance, with the TD children performing better than the people with DS. Interestingly, we found that trajectory analyses relating performance on both spatial abilities to Ravens ability scores yielded a similar slope for the people with DS and the TD children, despite the significant difference in mean performance observed for Mental Rotation.