Impaired linguistic processing and atypical brain laterality in adults with ADHD

Abstract

Introduction: cognitive/behavioral testing, structural imaging, and functional imaging, has demonstrated atypical cerebral asymmetries in patients with attention-deficit/hyperactivity disorder (ADHD). However, few studies directly examined the nature of hemispheric specialization and interaction in this population. Methods: the present experiment applied techniques from behavioral laterality research to assess directly left/right brain dynamics in unmedicated adults with ADHD (n=21) and controls (n=22). We used a lateralized lexical decision task to assess hemispheric differences in word recognition and cross-callosal interhemispheric transfer of linguistic information. Results: analysis of variance indicated that ADHD subjects were impaired relative to controls in identifying words in both hemispheres (P=0.001). Furthermore, ADHD subjects exhibited decreased effects for ‘word regularity’ (P=004), enhanced effects of ‘word frequency’ (P=007), and an increased bias for ‘nonword’ responses overall (P=03), as well as during left visual field trials in particular (P=01). Conclusions: adult subjects with ADHD demonstrated poor linguistic processing. Group differences in sensitivity to semantic and phonological linguistic variables, along with differences in response biases, suggested that ADHD subjects had reduced left hemisphere and enhanced right hemisphere involvement during our task. These findings are relevant to current research investigating ‘endophenotypes’ in ADHD, as laterality indices may prove useful in etiological research, particularly molecular genetic investigations, and highlights the relevance of brain laterality research in clinical psychiatry.

Introduction

The validity of the diagnosis of attention-deficit/hyperactivity (ADHD) in adults has gained considerable strength with recent reports [1], [2], [3], [4], [5], [6], [7], [8], [9]. However, relative to studies of children with ADHD, few have compared basic cognitive processes between adults with ADHD and controls, and what core deficits in cognitive processing exist, if any, remains unclear. Therefore, more studies of adult ADHD are needed to clarify the basic processes which underlie the disorder and a focus on adults with ADHD may actually have an advantage of being less confounded by developmental effects.

A variety of cognitive studies have suggested that childhood ADHD is associated with differences in brain lateralization [10], [11], [12], [13], [14]. However, few have been designed explicitly to address this issue. The covert orienting of spatial attention paradigm adapted from Posner et al. [15] has been used to study brain laterality with respect to the neural processes used for shifting attention. Work with this paradigm in children with ADHD has generated interesting but conflicting results [10], [13], [14], and has not consistently incorporated controls relevant to brain laterality research.

Sheppard et al. [12], using a line bisection task, showed that children with ADHD bisect lines with a rightward bias indicating a ‘neglect’ of the left half of space. They inferred that ADHD might involve abnormal right hemisphere processing. A similar deficit was observed in patients with right-sided parietal lesions [16]. In this vein, Heilman and colleagues have highlighted numerous findings showing that ADHD symptoms are, in many ways, reminiscent of those associated with right sided brain damage [17]. Furthermore, Garcia-Sanchez et al. [11] found ADHD teenagers to be impaired on a battery of ‘right hemisphere’ tasks.

In contrast to the above studies, some recent findings have implicated left hemisphere dysfunction in child and adolescent ADHD samples showing that linguistic impairments, including difficulty with semantic processing, are contributing factors [8], [18], [19], [20], [21], [22], [23], [24].

Manassis and colleagues [25], [26] conducted two studies utilizing versions of Bryden and MacRae's Dichotic listening for emotions and words [27] to investigate brain laterality differences between children with ADHD, anxious children, and normal controls. They reported no group differences in measures of laterality. However, these experiments utilized verbal responses which might have confounded the results given that speech is left lateralized. Also, these experiments utilized the signal detection measure ‘d prime’ as a primary dependant measure in spite of the inability to assess laterality of false alarms with this paradigm.

In summary, neurocognitive testing in ADHD, primarily with child samples, has yielded conflicting results in relation to lateralized differences. Not all studies have incorporated important controls for laterality and no consensus exists as to the exact nature of associated cognitive impairments.

Imaging studies, both structural and functional, have reported possible lateralized differences in both adult and child ADHD samples (for review see: [28], [29]). For instance, structural studies in ADHD children and adolescents have found reduced caudate volumes and reversed asymmetry as well as decreased right-sided globus pallidus and anterior grey and white matter compared with controls [3], [5]. Furthermore, although inconsistent with regard to the exact location, several reports suggest that ADHD children have smaller regions of the corpus callosum compared with controls [30], [31], [32]. A number of resting-state functional activation studies indicate abnormal R>L asymmetries in adults and children with ADHD [33], [34], [35], [36], while others looking at activation during cognitive challenges have indicated decreased right-sided activation [7], [37], [38].

Importantly, lateralized differences in brain structure and function do not clearly inform us about the functional impact on interhemispheric interaction. ADHD-associated grey and white matter volume changes are inconclusive as to their impact on cognition and behavior while functional imaging differences in a single hemisphere are difficult to wholly attribute to a specific lateralized deficit given the complexity and diffuse networking of the cerebral cortex. Nonetheless, taken together, there is a broad range of imaging studies suggesting that ADHD involves atypical cerebral asymmetries of both structure and function.

In the current experiment, we used a well-studied lateralized lexical decision paradigm to directly investigate the nature of hemispheric specialization and interaction in adults with ADHD. This paradigm has been studied in normal control, split brain, dyslexic, and schizophrenic subjects [39], [40], [41], [42], [43]. The lateralized lexical decision task measures a subject's ability to process words and orthographically regular (pronounceable) nonwords presented in either the left visual field (LVF), targeting the right hemisphere (RH), or in the right visual field (RVF) targeting the left hemisphere (LH).

During this task, right-handed subjects typically show an overall RVF advantage in accuracy and latency, which is taken to reflect left hemisphere (LH) specialization for word recognition. Also, it is common to find that words are processed better in the LH while nonwords are processed equally well in either hemisphere. Considering each hemisphere independently, words are processed more accurately than nonwords in the LH, while nonwords are often processed better than words in the RH. However, the size of the latter difference is variable across studies.

A previous study by Iacoboni and Zaidel [41] sought to determine to what extent each hemisphere can process words and nonwords independently. They reasoned that if task processing took place in a single hemisphere there should be no difference between unilateral and bilateral presentations where a distracter ‘word’ or ‘nonword’ adds ‘noise’ to the non-targeted hemisphere. Alternatively, if task processing required interhemispheric interaction then the addition of ‘distracter noise’ to the non-targeted, but ultimately contributing hemisphere, should impair performance. They found that bilateral presentations only significantly impaired performance for word targets projecting to the RH. This suggests that RH word processing is dependent upon some form of LH involvement during this task.

With lateralized lexical decisions, interhemispheric interaction is assessed by measuring lexicality priming. This measures how distracters projected to one hemisphere affect the subject's ability to resolve targets in the opposite hemisphere. Normal right-handed subjects typically perform better when the distracter is the same category as the target (i.e. target and distracter are either both words or both nonwords). If interhemispheric processing is impaired this effect should be attenuated or absent [44].

Lateralized lexical decision also informs about the degree to which left or right hemisphere ‘stimulus processing’ and/or ‘response’ strategies are expressed. Stimulus processing strategies are explored within the framework of the hemispheric version of the dual route model of word recognition [45]. This model distinguishes two types of reading: sounding words out (phonological), or recognizing them as visual patterns (lexical). When using a sounding out strategy, subjects typically process words that sound like they are spelled (regular words) better than words with irregular spellings. When using the ‘visual word’ strategy, subjects typically process common or ‘high frequency’ words better than uncommon or ‘low frequency’ words. Results from acquired dyslexia [46], [47] and from split brain patients [48] suggest that both hemispheres use the ‘visual word’ strategy but only the LH uses a sounding out strategy. Also, some data suggests that the RH may be especially sensitive to word frequency [49]. Thus, group's sensitivity to ‘word’ frequency and regularity can potentially inform us about the degree to which left and/or right hemisphere ‘stimulus processing’ strategies are emphasized.

Response strategies can be explored with the signal detection measure ‘beta’ or ‘bias’. Beta measures response strategies knowingly or unknowingly adopted by subjects during a task [50]. In lateralized lexical decisions, normal right handed subjects typically show a bias to respond ‘word’ during RVF-LH trials and ‘nonword’ during LVF-RH trials [41], [51]. Also, subjects typically show a ‘word’ bias overall which likely indicates LH dominance for this task. Thus, by assessing bias we can potentially gain insight about the relative use of left or right hemisphere based ‘response strategies’.

In the current study, we assess lateralized word and nonword processing along with interhemsipheric interaction in adults with ADHD. We also explored the degree to which ADHD subjects emphasize left or right hemisphere ‘stimulus processing’ and/or ‘response’ strategies.