Pseudoword spelling ability predicts response to word spelling treatment in acquired dysgraphia

ABSTRACT

Although rehabilitation of acquired dysgraphia can be quite effective, identifying predictors of responsiveness to treatment is useful for prognosis and individualization of treatment protocols. This study examined whether various features of treatment response were predicted by the integrity of one or more of the central cognitive components of spelling: orthographic long-term memory, orthographic working memory, and phoneme-grapheme conversion. Twenty dysgraphic individuals received 12 weeks of bi-weekly, individualized, lexically-based spelling rehabilitation using a spell-study-spell paradigm. Linear multiple regression modelling examined whether the type and severity of the dysgraphic deficit, assessed before rehabilitation, predicted the magnitude and rate of improvement, generalization to untrained items and maintenance of treatment gains. The results revealed that pseudoword spelling accuracy – indexing the integrity of the phoneme-grapheme conversion system – was the only factor examined that significantly predicted the rate of accuracy gains for trained words as well as the extent of generalization to untrained words. Pre-treatment pseudoword spelling accuracy also predicted retention of gains for trained and untrained words at 3-month follow-up. These findings reveal that the integrity of the phoneme-grapheme conversion system prior to dysgraphia rehabilitation may play a key role in rehabilitation-driven recovery, even when the treatment approach targets lexical rather than pseudoword spelling processes.

KEYWORDS:

• Spelling
• Dysgraphia
• Phoneme-grapheme conversion
• Sublexical
• Pseudowords

Acknowledgements

We extend our gratitude to the study participants for their contributions to this work, as well as to Donna Gotsch for her dedication to testing and treatment during the course of this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Except for participant RFZ, who only attempted enough untrained words to select a set of 30 approximately matching the 40 trained words.

2 Low probability segments had phoneme-grapheme mappings that accounted for, on average, fewer than 5% of the observed mappings for a given phoneme in a given syllable position in American English (Hanna et al., 1966.) For example, Hanna et al. documented that the sound “ch” in syllable initial position mapped to the graphemes TI (e.g., “QUESTION”) in 4.08% of words. Therefore, we considered QUESTION to be irregular because it contains at least one low probability segment.

3 Note that timepoint was coded as a categorical variable, using simple coding with pre-treatment as the reference level.

4 With the exception of the Model 1B5 for Irregular Words, which used whole-word accuracy instead of percent letter accuracy.

5 Note that subsets of the current participant cohort and behavioral dataset have been reported previously (Purcell & Rapp, 2018; Tao & Rapp, 2019; Wiley & Rapp, 2019; Rapp & Wiley, 2019).

6 Predictive R² is a measure of variance-explained that uses the Predicted Residual Sum of Squares (PRESS; Allen, 1974; see also Schüürmann et al., 2008). Conceptually equivalent to a leave-one-out approach, the Predictive R² analysis generates the proportion of variance that is expected to be explained by the predictors on a new sample of participants. Therefore, in this case, predictive R² provides a more stringent evaluation of a model’s ability to provide a prognosis of response to treatment.

7 Note that the models assessing performance on Training Words at the pre-timepoint and during training had n = 20 (the full set of participants); all other models had either n = 19 (excluding participant JRE, who did not complete the post or follow-up timepoints) or n = 18 (additionally excluding participant AES, who did not complete the follow-up timepoint).

8 With regard to interpretation of R² values, we note that the commonly agreed upon standards are based on Cohen (1988) who suggests 0.01 as a small effect, 0.09 as a medium effect, and 0.25 as a large effect.

9 The pre-treatment accuracies on the specific word lists evaluated in each of the analyses were always significant predictors as well. Note that the direction of the relationship was always negative, indicating that lower pre-treatment scores on a list predicted greater improvement. However, this is due simply to the fact that lower scores mathematically allow for greater improvement, due to ceiling effects.

10 In that model, Irregular Words: Pre vs. Post, the usefulness of the lesion size predictor (the variable’s contribution to R² when included last in the model) was 7%, compared to 31% for the pseudoword accuracy predictor.

11 As indicated in Figure 1, another possible locus of spelling impairment are the motor planning or execution processes. These were not the source of the participants’ spelling difficulties because: (a) all participants had legible writing, (b) for the majority, spelling difficulties were not equally affected for words and pseudowords which is what would be expected from motor deficits and (c) the 8 participants who completed the copy-transcoding task specifically designed to evaluate letter form selection/planning, scored between 99–100% accuracy.

12 Note that the activation of both systems and their interaction has also been proposed for reading (Miceli et al., 1994).

Additional information

Funding

This work is part of a multi-site, National Institute on Deafness and Other Communication Disorders (NIDCD)-supported project examining the neurobiology of language recovery in aphasia [grant number DC006740].