Mental rotation and motor performance in children with developmental dyslexia

Highlights

• Dyslexic children had a slower overall reaction time than non-dyslexic children.

• Dyslexics show higher reaction times in “letters” and “pseudo-letters”, but not in “animals” compared to the controls.

• The increase of reaction times with increasing angles is more pronounced in dyslexic children than in the control group.

• Dyslexic children do not differ from non-dyslexic children in their motor performance.

Abstract

We compared the performance of normal-reading (N = 14) and dyslexic children (N = 14) in a chronometric mental rotation task (cMRT) using letters, animals and pseudo-letters, which are objects that look like letters. In a typical chronometric mental rotation task two items are presented simultaneously on a screen whereby the right item is a rotated version of the left item and could be the same or a mirror version of the left item. The mental rotation paradigm is an appropriate method to test predictions of two different approaches trying to explain the problems for dyslexics when reading. According to the functional coordination deficit (FCD) model dyslexics show a failure in suppression of symmetry in the representation of graphemic material and therefore cannot decide whether the letter is normal or mirrored because of an ambiguous mapping between phoneme and grapheme representations. Therefore, the deficits of dyslexic children regarding mental rotation performance are restricted to the stimulus “letters”. According to findings that propose the involvement of the cerebellum in mental rotation tasks and a cerebellar deficit in dyslexia, an impaired mental rotation is expected affecting all types of stimuli. To investigate the involvement of the cerebellum, motor performance was additionally assessed because the cerebellum plays an important role in motor functions and motor imagery. For the cMRT we found that the dyslexic children show both slower reaction times regarding the stimulus “letters” and “pseudo-letters” and increased overall reaction times compared to non-dyslexic children. The mental rotation effect was more pronounced in dyslexic children than in normal readers. In contrast to previous approaches, the results of our study support the idea that poor results in mental rotation result from deficits in mental rotation itself rather than from a decision problem after mental rotation which supports the predictions of the cerebellar deficit hypothesis. However, since the impairment of dyslexics regarding mental rotation performance is letter-specific and motor results show no differences between dyslexic and non-dyslexic children, further approaches next to the cerebellar deficit hypothesis must be taken into account, especially in consideration of the fact that there are a number of causes for the failure in reading.

Introduction

According to the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM–IV–TR; American Psychiatric Association, 2000) developmental dyslexia is defined as a specific and persistent failure to acquire efficient reading skills despite normal intelligence, sufficient motivation, adequate sensory abilities and appropriate reading instruction. Dyslexia is the most common of the learning disabilities; its clinical prevalence is ranging from 5% up to 17.5% (Shaywitz, 1998). Results concerning the sex ratio of this specific reading disorder remain controversial. Whereas Flannery, Liederman, Daly, and Schultz (2000) reported a clear sex bias toward males irrespective of severity of disability or comorbidities like attentional disorders, Shaywitz, Shaywitz, Fletcher, and Escobar (1990) interpreted the imbalance as referral bias. Dyslexic children show clinical features such as impaired reading comprehension skills, problems in word recognition and poor oral reading skills performance compared to normal reading children of a comparable age, intelligence and education level (4th ed., text rev.; DSM–IV–TR; American Psychiatric Association, 2000).

One characteristic symptom of dyslexia is the reversal error: a bias to reverse both the order of letters within a word (for example “was” instead of “saw”) and the orientation of single letters (for example “b” vs. “d”) (Lachmann, Schumacher, & van Leeuwen, 2009). The functional-coordination deficit (FCD) model (Lachmann, 2002) concentrates on the explanation of this type of error. According to this theory, reversal errors result from a failure to suppress symmetry generalization in reading. Symmetry generalization is a tendency to generate mirror images as well as other orientations of an object and to store them in the same category as the original. Neurologically, mirror images and other orientations activate similar patterns of neural activity as the upright original (Lachmann, 2002). This facilitates the recognition of an object in different orientations (object constancy). In learning to read, however, such a mechanism is a hindrance. Problems occur, for instance, when graphemes in different orientation or mirror images like “b” and “d” have different phonology, but are stored under the same category. As a consequence, two different letters share one common phonological representation. However, this one-to-one relation between the grapheme and the phoneme representation is essential in learning to read. Normal readers suppress the symmetry generalization while learning to read, children with developmental dyslexia do not. The FCD approach assumes that symmetry generalization problems are not restricted to letters that have symmetrical counterparts, like “b”, “d” and “p” (Rusiak, Lachmann, Jaskowski, & van Leeuwen, 2007).

Dyslexics show different patterns of reading and writing problems which led to the distinction between three subgroups: (1) dysphonetics; (2) dyseidetics (3) dysphoneidetics, a mixture of both deficits. Dysphonetics show phonological processing deficits, whereas in dyseidetics visual deficits lead to problems in recognizing the visual gestalt and therefore to a slowed direct access to the lexicon (Boder, 1970). According to the Dual Route Cascade model (DRC; Coltheart, Curtis, Atkins, & Haller, 1993), words follow two different routes: regarding frequent words there's a direct route from the visual gestalt of the word to its phonology and meaning (lexical semantic route), whereas the second route is restricted to non-words, which are irregular words, where a grapheme-to-phoneme conversion of individual letters takes place (grapheme-to-phoneme route; Coltheart et al., 1993). Thus, problems in the first route lead to deficits in word-reading, whereas the second route is involved in non-word reading (Lachmann, Berti, Kujala, & Schröger, 2005). Thus, Coltheart (1996) suggested that performance in word- vs. non-word reading helps to differentiate between the mentioned subtypes. In line with this, problems in frequent word reading argue for visual problems which are prominent in the dyseidetic subtype, whereas dysphonetics who fail in tasks requiring phonological skills are expected to show problems in non-word reading (Lachmann et al., 2005). However, even if the conclusion that a failure in non-word reading stems from problems in grapheme-to-phoneme mapping is plausible, problems in frequent word reading must not necessarily characterize visual deficits, but can also represent phonological impairments (Lachmann et al., 2005). Since some studies (Borstig et al., 1996, Vellutino et al., 1972) showed, that dyseidetics don’t differ from the other subtypes regarding their visual information processing. However, it remains unclear to which extent phonological deficits are responsible for these results (Lachmann et al., 2005).

Within our study we used the mental rotation paradigm to clarify where reading problems result from because it offers an opportunity to test the predictions of two specific approaches concerning this issue, namely (1) The functional-coordination deficit model (Lachmann, 2002) and (2) The cerebellar deficit hypothesis (Nicolson, Fawcett, & Dean, 2001). According to the FCD-approach reading problems result from an ambiguous mapping between graphemic and phonemic representations. The cerebellar deficit hypothesis argues for difficulties in automatization of both mild motor and basic articulatory skills. This in turn would cause problems in learning to read.

Mental rotation is a certain visuo-spatial ability which involves the process of imagining how a two- or three-dimensional object would look if rotated away from its original upright position (Shepard & Metzler, 1971). In the classic paradigm of Cooper and Shepard (1973) two stimuli are presented simultaneously and the participants have to decide as fast and accurately as possible if the right stimulus, presented under a certain angle of rotation, is the same or a mirror image of the left stimulus, the so called “comparison figure”. While angular disparities are varied systematically, response times and accuracy rate are assessed as dependent variables.

Results of the mental rotation performance should be interpreted with focus on the traditional theory of mental rotation, which differentiates five independent information-processing stages of mental rotation (Shepard & Cooper, 1982). These are: (1) perceptual preprocessing, (2) identification/discrimination of the character and identification of its orientation, (3) mental rotation, (4) judgment of the parity, and (5) response selection and execution (Heil & Rolke, 2002).

Heil and Rolke (2002) showed that the subprocesses of mental rotation are executed successively. Based on this independency Cooper and Shepard (1973) concluded that only mental rotation itself is influenced by the angular disparity. Therefore the increase of the function reflects mental rotation itself whereas the slope represents the other four subprocesses (see Kail, Pellegrino, & Carter, 1980) which leads to the conclusion that a steeper increase of the function in dyslexic children argues for impairments restricted to the mental rotation process itself.

According to Lachmann et al. (2009) the failure of dyslexic individuals to suppress symmetry generalization should not lead to a problem in the mental rotation itself, but to a deficit in a later stage of processing, namely the decision if the letter is “normal” or “mirror-reversed”. For instance, dyslexics associate both the mirror-reversed grapheme “

” and the normal grapheme “R” with the phonemic “R”. This could be an explanation, why dyslexic participants show no impairments in naming the letters shown in different orientations (Corballis, Macadie, Crotty, & Beale, 1985). The problem occurs, when dyslexic participants have to judge the orientation, because both the normal and the mirror version represent the phonemic “R”, but in normal readers only the normal “R” is connected with the phonemic representation of this letter, which facilitates the decision. Rusiak et al. (2007) supported this idea by showing slower reaction times for dyslexic participants compared to controls, but no differences regarding the accuracy rate of mental rotation. Besides, the mental rotation effect, the effect of the linear increase of reaction time with increasing angular disparity, occurred in both groups. The researchers concluded that the problems of people with dyslexia are restricted to the deciding process and are not related to the visuo-spatial information processing itself. However, it was concluded that in tasks requiring the participants to detect symmetry, this failure of suppression might be an advantage. Lachmann and van Leeuwen (2007) tested this assumption by using a same-different task. The stimulus material consisted of letters and dot patterns which were either symmetric (for example: A, D, T) or asymmetric (for example: F, R, S). Participants had to press the “same” button, when both normal and mirror images of the letters and the dot patterns were presented. Asymmetric stimuli were judged as “different”. According to the prediction, dyslexic children showed faster reaction times, in particular with letters.

As soon as the failure of suppression is adverse like in classic mental rotation tasks, slower reaction times and error rates are expected to be stimulus-specific.

Impairments regarding mental rotation performance in developmental dyslexia have been found to be restricted to letter-specific stimulus material (Rusiak et al., 2007) which would support the notion of an affected judgment of the parity or the perceptual preprocessing. However, Rüsseler, Scholz, Jordan, and Quaiser-Pohl (2005) showed a decreased accuracy rate of the dyslexic group for all three kinds of stimuli (letters, three-dimensional figures, and pictures of animals or humans) used in the mental rotation task. This underlines the notion that mental rotation itself is impaired which is predicted by the cerebellar deficit hypothesis.

However, earlier stages can also be involved in developmental dyslexia since the failure to suppress symmetry generalization in letters not only affects processes after mental rotation, but also occurs in earlier stages, namely in visual encoding. In a series of studies, dyslexic participants showed no deficits in standard visual processing tasks (Vellutino, Steger, Moyer, Harding, & Niles, 1977). Thus, it was concluded that visual encoding deficits are not involved in developmental dyslexia. However, according to Lachmann and van Leeuwen (2007) this reasoning is premature. Dyslexic individuals may use abnormal visual information processing strategies resulting from an underlying deficit. However, the visual task used may be insensitive to this deficient strategy. Lachmann and van Leeuwen (2007) assumed that dyslexic individuals show an automatization deficit. According to the FCD approach (Lachmann, 2002) normal readers learn to suppress the tendency to generate mirror images of letters, but children with developmental dyslexia do not acquire this symmetry-suppression strategy. They treat graphemes like objects, whereas normal reading adults perceive letters and non-letter material with different strategies. This notion of differentiation in normal-readers between letter and non-letter processing was confirmed by the findings of van Leeuwen and Lachmann (2004). Non-letter stimuli surrounded by a congruent shape facilitated object recognition (congruence effect), letters surrounded by a congruent shape led to impaired recognition (negative congruence effect). This was explained by the theory that a congruent surrounding emphasizes the symmetry of a configuration, which facilitates the processing of shapes by the use of symmetry generalization, but complicates the encoding of letters in which symmetry generalization needs to be suppressed. In more recent research, Lachmann, Khera, Srinivasan, and van Leeuwen (2012) came to the conclusion that the congruence effect indicated holistic grouping; the negative congruence effect, which means a preference for incongruent surroundings, was associated with an analytic encoding strategy. Therefore, the researchers argued that normal readers differentiate between holistic non-letter processing and analytic letter processing. Based on this finding, Lachmann et al. (2012) addressed whether this original differentiation also occurs in illiterates. Their results revealed that the dyslexic participants showed analytic visual perception for both letters and non-letters.

The cerebellar deficit hypothesis (Nicolson et al., 2001) proposes that cerebellar abnormality causes the characteristic impairments of developmental dyslexia. The authors tried to outline a hypothetical causal chain declaring that cerebellar deficits would lead to difficulties in automatization of both mild motor and basic articulatory skills. This in turn would cause problems in learning to read. There is empirical evidence confirming this hypothetical chain: (1) skill automatization corresponds to the role of the cerebellum (Jenkins, Brooks, Nixon, Frackowiak, & Passingham, 1994). (2) Dyslexics show difficulties in the automatization of skills, which has been revealed in the dual-task paradigm of Nicolson and Fawcett (1990). Whereas dyslexic children showed no deficits in the single-task condition (balancing) they were significantly impaired in the dual-task condition (balancing and counting backwards), which implicates that dyslexic children require conscious resources to solve the balance task, and thus their performance is impaired by adding a second task which distracts attention from the primary task.

According to findings that propose the involvement of the cerebellum in mental rotation tasks (Podzebenko, Egan, & Watson, 2002) and a cerebellar deficit in dyslexia (Bishop, 2002, Nicolson et al., 1999), an impaired mental rotation performance is expected. If reversal errors result from problems related to spatial operations which is predicted by the cerebellar deficit hypothesis, there will be group differences in the increase of reaction time with rotational angle (Lachmann et al., 2009, Rusiak et al., 2003). Moreover, poor results should be present for all kinds of stimuli used in the task (Rüsseler et al., 2005). This prediction relies on a functional magnetic resonance imaging (fMRI) study (Jordan, Heinze, Lutz, Kanowski, & Jäncke, 2001) revealing that the same neural structures (bilateral superior and inferior parietal lobes) are activated when a mental rotation task was performed with different types of stimuli. In line with this approach, Rüsseler et al. (2005) revealed a decreased accuracy rate for dyslexic participants compared to controls, regarding all three kinds of stimuli (letter, pictures, abstract figures).

Since the cerebellum plays an important role in mental rotation tasks we analyzed motor performance of dyslexic children to investigate the meaning of this hypothesis for mental rotation performance and in the end for the reading problems of dyslexic children.

Besides these visuo-spatial problems, dyslexic children also show impairments in motor performance (Nicolson and Fawcett, 1994, Viholainen et al., 2002, Viholainen et al., 2006). The co-occurrence of motor and language difficulties in children with developmental dyslexia is fairly common and reflected in the high prevalence of about 80% of the affected people (Nicolson & Fawcett, 1994). Findings of Reynolds, Nicolson, and Hambly (2003) support the relationship between motor and language development by revealing improvements of dyslexic children in both motor (balance, manual dexterity and eye movement) and language skills (phonological skill, reading, verbal fluency and semantic fluency) after a six month exercise-based treatment. The longitudinal study of Viholainen et al. (2002) also analyses this co-occurrence by comparing children with and without familial risk for dyslexia regarding their motor development. Their results showed both a slower gross and fine motor development in the at-risk group compared to the control group.

There are several assumptions trying to explain why motor and language impairments overlap so often. One of these views argues for a common underlying neurocognitive correlate. Several studies hypothesized that the cerebellum is involved in both language (Silveri & Misciagna, 2000) and motor skills (Marr, 1969, Stein, 1986). This finding is supported by neuroanatomical work revealing cerebellar connections to Broca's area and to premotor areas (Leiner, Leiner, & Dow, 1993). Silveri, Leggio, and Molinari (1994) support the idea of cerebellar involvement in language by showing agrammatic speech without other cognitive impairments caused by right hemi-cerebellar syndrome. In a series of studies Nicolson and Fawcett analyzed the overlap between cerebellar deficits and the symptoms of dyslexia. Compared to age matched controls dyslexic children showed behavioral evidence of abnormal cerebellar function in the following skills: balance (Nicolson & Fawcett, 1994), time estimation (Nicolson, Fawcett, & Dean, 1995), muscle tone and coordination (Fawcett, Nicolson, & Dean, 1996). This finding is reinforced by the functional imaging study of Nicolson et al. (1999) showing less brain activation in the right cerebellum of dyslexic participants during motor tasks compared to controls.

These results from behavioral tests and the functional imaging study support the cerebellar deficit hypothesis (Nicolson et al., 2001) proposing that cerebellar abnormality causes the characteristic impairments of developmental dyslexia.

We used mental rotation tasks to test predictions of the FCD approach against those of the cerebellar deficit hypothesis. According to the FCD approach, mental rotation itself is not affected in developmental dyslexia, but rather later stages of processing. This notion results from the finding of significantly increased reaction times for dyslexic children and the null-finding of an interaction between “group” and “angular disparity” in the previous literature (Lachmann et al., 2009). According to the FCD model (Lachmann, 2002), differences in reaction times were explained by a decision problem which occurred after the mental rotation process. Here, participants have to distinguish whether the letter is normal or mirror-reversed (Lachmann et al., 2009). This decision difficulty results from an ambiguous mapping between phonemes and graphemes. Therefore, the analysis is conducted separately for “same” (rotated and not mirror-reversed) and “different” (rotated and mirror-reversed) trials to examine how the failure to suppress symmetry generalization affects mental rotation performance. Dyslexics are expected to be impaired in judging the orientation only of “different” (mirror-reversed) images because in dyslexics both the normal and the mirror version represent the phonemic letter, whereas in normal readers only the normal upright orientation of the letter is connected with the phonemic representation of this letter. Since this problem is restricted to letters, dyslexic children should be impaired only in graphemes.

However, according to the cerebellar deficit hypothesis, dyslexic children are impaired for all three types of stimuli which is confirmed by the finding of Rüsseler et al. (2005) who found increased error rates in a psychometric mental rotation task, a task with no time pressure, in dyslexic children compared to normal-reading children. To confirm this approach, it is expected that mental rotation itself is impaired which leads to a steeper increase of reaction times with increasing angles in the dyslexic group compared to the controls. Regarding stimulus type, dyslexic children should not differ in graphemic material from non-dyslexic children both in “same” and “different” trials, but should show stimulus independent impairments manifested in increased overall reaction times. Furthermore, poor motor results are expected (Nicolson et al., 2001) because the cerebellum plays an important role in motor functions and motor imagery (Calhoun et al., 2001). Due to the large variety of motor tasks in the previous literature we concentrated on basic motor skills, like manual dexterity, ball skills and balance.