Neural correlates of spelling difficulties in Alzheimer’s disease
Introduction
The cognitive profile of Alzheimer’s disease (AD) patients is mainly characterized by an episodic memory impairment that complicates the acquisition of new information (Fox et al., 1998). However, a linguistic deficit is also present in many of the cases. The verbal output of AD patients, even in the early stages of the disease, lacks lexical diversity and presents semantic errors and circumlocutions (Cuetos et al., 2009, Cuetos et al., 2012, González Nosti et al., 2008, Rodríguez-Ferreiro et al., 2009). This linguistic pattern has been extensively documented in the oral domain, but there is much less evidence regarding writing abilities of these patients (for a review see Neils-Strunjas et al., 2006).
Writing difficulties associated to the disease were observed by Alzheimer (1907) himself in the patient Auguste D. Since then, a variety of deficits, with origins ranging from central to peripheral, have been described in the AD population. This line of research has gained certain popularity in the last years with the case-studies of the renowned writers Iris Murdoch (Garrard et al., 2005, Pakhomov et al., 2011) and Agatha Christie (Lancashire, 2010).
The most common pattern of deficiency includes lexical dysgraphia (Croisile et al., 1996, Forbes et al., 2004, Hughes et al., 1997, Rapcsak et al., 1989) which may be concurrent with a phonological or sublexical deficit (Aarsland et al., 1996, Luzzatti et al., 2003, Neils and Roeltgen, 1994, Pestell et al., 2000, Platel et al., 1993) in later stages of the disease. In some cases, peripheral impairments are also present (Forbes et al., 2004, Horner et al., 1988, Neils-Strunjas et al., 1998, Venneri et al., 2002).
Following the influential model by Ellis (1982), lexical dysgraphia is characterized by errors in real word, as opposed to pseudoword, writing, and, more specifically, a diminished capacity to correctly spell irregular words with arbitrary orthography due to damage in the orthographic lexicon. On the other hand, phonological dysgraphia is present when the patient is incapable of writing pseudowords or rare unfamiliar words, and is caused by damage to the route that allows conversion of phonemes into their corresponding graphemes. Finally, peripheral dysgraphias involve a deficit in the mechanics of written output, including inappropriate allograph selection, inaccurate graphomotor patterns, awkward letter construction, poor letter spacing or spatial misalignment, among others.
Hence, the profile of spelling impairment present in AD patients is congruent with the deficit described in the oral domain and could be caused by degradation of specific linguistic subcomponents (Croisile et al., 1996, Forbes et al., 2004, Rapcsak et al., 1989). Nonetheless, some authors ascribe it to the general cognitive decline associated to dementia (Aarsland et al., 1996, Glosser and Kaplan, 1989, Silveri et al., 2007), especially in the cases of peripheral disorders (Neils-Strunjas et al., 1998).
The cognitive deficit associated to AD is grounded in the degradation of neural structures responsible of the different capacities affected. Histopathological studies have shown that neuronal changes associated to AD initiate in the hippocampus and the trans-enthorrinal region and, later on, spread to the neocortex (Braak et al., 1999, Braak and Braak, 1995). More recently, automatic analyses of structural MRI scans through voxel-based morphometry (VBM) have confirmed that brain atrophy in AD patients starts in medial temporal structures and gradually extends to more lateral temporal regions as well as to the parieto-occipital area and the frontal lobe (Derflinger et al., 2011). In contrast, regions like the occipital pole, the sensoriomotor cortices and the cerebellum appear to be spared (Karas et al., 2003). In the last years, several studies have used VBM to examine the relationship between cortical changes and different symptoms of AD such as depression (Hyun Son et al., 2013), delusions (Bruen et al., 2008) or semantic deficits (Rodríguez-Ferreiro et al., 2012). However, the link between the spelling impairment of AD patients and their pattern of brain atrophy remains unstudied.
Historically, peripheral writing mechanisms have been situated in the Exner’s area, just above Broca’s area in the left frontal lobe (Exner, 1881), while central components have been associated to the angular gyrus in the parietal lobe (Dejerine, 1891). Thus, the extrasylvian temporo-parietal cortex has been linked with the processing of orthographic representations (Patterson and Kay, 1982, Rapcsak et al., 1990) and lesions in the posterior temporal area appear to be linked to lexical agraphia (Croisile et al., 1989) as well as to tasks involving the retrieval of orthographic word forms (Beeson et al., 2003). Moreover, activity in the angular gyrus has been specifically associated with lexical subprocesses of writing in AD participants (Penniello et al., 1995).
In this study we aimed to obtain a broad profile of the spelling capacities of our participants and investigate its relation with brain atrophy. Given its preeminence in the AD population, we decided to focus in the more common deterioration of central, rather than peripheral, components of written language processing. Thus, we measured the amount of correctly spelled words in each subtest while ignoring errors related to peripheral subcomponents of written language, such as case misuse or misalignment. We assessed the capacity of a group of AD patients to write in different tasks, including spontaneous narrative, dictation and written naming, with both sentence and single word stimuli. In order to differentiate between lexical and sublexical contributions to spelling, we also included a pseudoword dictation task, and we distinguished between different types of real word stimuli depending on their degree of spelling ambiguity in the word writing task. Due to the high amount of one-to-one correspondences between sounds and letters in the Spanish orthographic system, most of the words can be spelled correctly following the basic, sublexical, phoneme-to-grapheme correspondences. On the other hand, the correct spelling of words containing non-univocal correspondences (i.e. /b/ to “b” or “v”; /y/ to “y” or “ll”) depends entirely on lexical knowledge, although orthographic rules based on letter groupings might be applied to assist accurate spelling in some cases (i.e. all words ending in /–iyo/ are spelled with “ll”, and not “y”; all words starting with /ue/ are spelled with an initial mute “h”).
The results of the AD patients in those tasks sensitive enough to differentiate between healthy and impaired seniors were then introduced into VBM analyses along with their brain volumes in order to investigate the relationship between cognitive capacity and the degradation of neural structures.
Section snippets
Participants
Twenty-two outpatients with probable AD and the same number of healthy seniors took part in the study. They all were native Spanish speakers and came from similar socio-economic backgrounds. None of them had a history of alcohol abuse, or neurological or psychiatric disorders other than AD. Other possible sources of cognitive impairment such as focal lesions or microbleeds were ruled out by neuroimaging tests. Participants with sensory impairments were also ruled out from the sample. All the
Results
The behavioural results of the AD and control groups are summarized in Table 2. Scores obtained by the two groups in all the tasks and subtests were compared by means of nonparametric Mann–Whitney tests due to violations of the normality assumption (Shapiro–Wilk p<.05). We used a Bonferroni correction to adjust for multiple comparisons, so all effects are interpreted at a.005 level of significance.
Significant differences appeared in six of the measures, including the oral (U=46, z=−4.635, p
Discussion
In this study we aimed to assess the writing abilities of a group of AD patients by means of a test battery that included a selection of tasks designed to establish a wide profile of their spelling capacity. Moreover, we intended to determine the relationship between writing dysfunction and grey matter loss in different brain areas by entering the participants’ behavioural results into multiple voxel-based regression analyses with cortical volume values extracted from high resolution
Acknowledgements
This study was supported by grant PSI2012-31913 from the Spanish Government.
References (56)
- et al.
Lexical and non-lexical spelling deficits in dementia of the Alzheimer’s type
Brain Lang.
(1996) A fast diffeomorphic image registration algorithm
NeuroImage
(2007)- et al.
Unified segmentation
NeuroImage
(2005) - et al.
Comparison between oral and written spelling in Alzheimer’s disease
Brain Lang.
(1996) - et al.
Mini-mental state: a practical method fro grading the cognitive state of patients for the clinician
J. Psychiatr. Res.
(1975) - et al.
The evolution of dysgraphia in Alzheimer’s disease
Brain Res. Bull.
(2004) - et al.
Linguistic and nonlinguistic impairments in writing: a comparison of patients with focal and multifocal CNS disorders
Brain Lang.
(1989) - et al.
Impaired spelling in Alzheimer’s disease: a linguistic deficit?
Neuropsychologia
(1999) - et al.
Dysgraphia in mild dementia of Alzheimer’s type
Neuropsychologia
(1997) - et al.
A comprehensive study of gray matter loss in patients with Alzheimer’s disease using optimized voxel-based morphometry
NeuroImage
(2003)