No effect of orthographic neighbourhood in treatment with two cases of acquired dysgraphia

ABSTRACT

Background: Studies on the nature of processing within the spelling system have provided evidence for interactivity, where activation between levels of processing flows bidirectionally. In particular, activated letters at the level of the graphemic buffer feed activation back to the lexicon. As a consequence lexical neighbours are activated, which in turn reactivate the graphemic buffer, supporting target selection and leading to an effect of orthographic neighbourhood size in spelling. As a consequence of this feedback, treating words with many neighbours may be more beneficial, and generalisation may be more likely to occur.

Aims: The current study aimed to further examine this interactivity, and in particular the role of orthographic neighbourhood size in both treatment effects and generalisation, in the treatment of two individuals with acquired dysgraphia. Two phases of copy and recall treatment were conducted to investigate if orthographic neighbourhood size modulates treatment effects: in the first phase, treated words had no orthographic neighbours, in the second phase, treated words had many neighbours. Untreated control sets were used to investigate the influence of neighbourhood size on potential generalisation across items.

Main Contribution: Results showed that treated items improved, however neighbourhood size did not significantly influence the size of the treatment effect for either participant, and no clear evidence was found for generalisation to untreated items.

Conclusions: It is argued that (1) the amount of feedback from the buffer to the lexicon was reduced in the participants in this study, and consequently the treated items provide insufficient activation to orthographic neighbours or (2) the effect of feedback from neighbours is small compared to effects of treatment, resulting in similar results for treated words with and without neighbours.

This study informs the relationship between spelling impairment and effects of treatment. Furthermore, the absence of generalisation emphasises the importance of choosing functionally relevant items for treatment.

KEYWORDS:

• Acquired dysgraphia
• treatment
• orthographic neighbourhood
• spelling

Acknowledgements

During the preparation of this paper, Trudy Krajenbrink was funded by an International Macquarie University Research Excellence Scholarship, Lyndsey Nickels by an Australian Research Council Future Fellowship (FT120100102) and Saskia Kohnen by a Macquarie University Research Fellowship. We would like to thank David Howard for valuable assistance with the analyses.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Some words were repeated on the different lists that were administered. For the regression analysis, we only included the response from the first time the participants were asked to spell the word.

2. Regularity was not included in the regression; however, GEC showed no effects of regularity on any of the word lists (see Appendix D).

3. See the Treatment Results section for a detailed description of our letter accuracy measure.

4. We defined an error as orthographically related when either at least 50% of target letters were in the response (task: trash) or at least 50% of response letters were target letters (hatred: hit; based on Nickels’ (1995) analysis of phonological errors in spoken production).

5. Regularity was not included in the regression; however, JOD showed inconsistent effects of regularity on matched lists (see Appendix D): he showed better spelling for regular compared with irregular words on one of the word lists (JHU—PGC) but no significant difference was found on two other lists (Krajenbrink et al. (2016) and PALPA 44).

6. From this point on, when we refer to neighbours, we mean orthographic substitution neighbours.

7. Initially, we included an extra set of untreated words for JOD. With those items included JOD showed a significant improvement between Baseline 1 and Baseline 2 (McNemar’s test exact, p = .047). We therefore decided to include a third baseline to better determine the trajectory of performance. For simplicity and congruence with GEC, these items are not reported; however, these untreated items also showed no significant improvement after treatment.

8. Relative position indicates the position of a letter in relation to the other letters in the word, as opposed to absolute position in the target. For example, as the result of a deletion of the letter r from the target dress in the response dess, the e is no longer the third letter in the word, however it is still in the correct position in relation to the d and s. Therefore, the e is scored correct for position. In the case of a transposition (e.g., widonw for window), half a point is subtracted from the transposed letter (in this example, from the n). An addition is scored by subtracting half a point from each of the flanking letters. For example, the response diarty to the item diary receives a total score of four (five correct letters minus 0.5 for r and y) out of five target letters, which is a score of 0.8. Letters that are deleted or substituted receive a score of 0. For example, aro for the target arena receives a score of two letters correct (a and r) out of the five target letters, which results in a score of 0.4 for this response. (Note: o is assumed to be a substitution for e, and not an addition. Therefore, no scores are subtracted from the r.)

9. While our treatment conditions were confounded with order, in practice, the lack of difference between the treatment conditions means this is not an issue for interpretation of our results. However, it is worth acknowledging that had we found different effects in the two phases of treatment, it would have been hard to disentangle the effects of treatment type and order. While a third phase of treatment using the same treatment type as in the first phase could have been one way to possibly address this issue (had it been necessary) it is worth noting that even here we could not have excluded order effects completely: Effects of order are complicated and may interact with effects of, for example, motivation over time.

Additional information

Funding

Trudy Krajenbrink was funded by an International Macquarie University Research Excellence Scholarship, Lyndsey Nickels by an Australian Research Council Future Fellowship (FT120100102) and Saskia Kohnen by a Macquarie University Research Fellowship.