Insights into recurrent perseverative errors in aphasia: A case series approach

We would like to thank KCC, KVH, DRS, SMM, and TOW, the participants with aphasia, for their enthusiasm and hours of hard work. Thank you to the War Memorial Hospital and the Royal Rehabilitation Centre Sydney for allowing us to use clients from their centres. Finally, we would also like to thank Professor David Howard for his statistical support. During the preparation of this paper, Dr Lyndsey Nickels was funded by an Australian Research Council QE2 Fellowship.

Abstract

Background: Recurrent perseverative errors are most commonly produced by speakers with aphasia (Albert & Sandson, 1986). While some research has proposed that recurrent perseverative errors are primarily caused by an impaired ability to inhibit a prior response, most recent research proposes that these errors occur as a direct result of underlying language impairment at any specified processing level (e.g., Cohen & Dehaene, 1998).

We would like to thank KCC, KVH, DRS, SMM, and TOW, the participants with aphasia, for their enthusiasm and hours of hard work. Thank you to the War Memorial Hospital and the Royal Rehabilitation Centre Sydney for allowing us to use clients from their centres. Finally, we would also like to thank Professor David Howard for his statistical support. During the preparation of this paper, Dr Lyndsey Nickels was funded by an Australian Research Council QE2 Fellowship.

Aims: This investigation was designed to further explore the hypothesis that the type of recurrent perseverative errors produced reflects the level of language‐processing breakdown. Two specific questions were addressed: (1) Is the distribution of types of perseverative error similar to that of non‐perseverative errors within different language tasks? (2) Do the relative proportions of total and blended perseverative errors reflect an individual's language‐processing profile?

Methods & Procedures: The recurrent perseverative error patterns of five individuals with aphasia with varying levels of language‐processing breakdown were examined across three language tasks (repetition, reading aloud, and picture naming).

Outcomes & Results: The incidence of different types of recurrent perseverative errors was found to be influenced by the processing demands of individual tasks, relative to each participant's language‐processing breakdown. However, certain findings challenged current theories, such as the large proportion of perseverative errors that were unrelated to their targets and some participants' entire reproduction of prior neologisms across different responses. Different theoretical accounts are explored to attempt to understand these error patterns.

Conclusions: This study supports theories proposing that recurrent perseverative errors are a direct result of underlying language‐processing breakdown in aphasia (Cohen & Dehaene, 1998; Martin, Roach, Brecher, & Lowery, 1998). It is argued that examination of both whole word (total) and phoneme (blended) perseverative errors is necessary for a comprehensive analysis of the proposed relationship between perseveration and language‐processing impairment.

Notes

We would like to thank KCC, KVH, DRS, SMM, and TOW, the participants with aphasia, for their enthusiasm and hours of hard work. Thank you to the War Memorial Hospital and the Royal Rehabilitation Centre Sydney for allowing us to use clients from their centres. Finally, we would also like to thank Professor David Howard for his statistical support. During the preparation of this paper, Dr Lyndsey Nickels was funded by an Australian Research Council QE2 Fellowship.

1. From this point forward, we use the term perseverative errors to refer to recurrent perseverative errors.

2. The coding of blended perseverative errors is acknowledged to be arbitrary and hence there is the possibility that false identification of these errors may sometimes occur. To minimise potential false positive miscoding, items sharing initial and final phonemes or main vowels were not placed within five intervening items for initial phonemes, nor within three intervening items for those with the same final phoneme, nor consecutively for those with the same main vowel.

3. We acknowledge that whilst we would have predicted most total perseverations to involve the “classic” reproduction of whole words (as decribed by Cohen & Dehaene, 1998), this was not necessarily the case as our coding criteria for total and blended perseveration did not specify the lexicality of the error. That is, total perseverations need not be lexical and blended perseverations need not be non‐lexical. Hence, although most total perseverations were lexical, a number of neologisms were classified as “total perseverations” as they invoved a complete reproduction of a prior neologism. The main challenge was that these neologisms cannot simply result from retrieval of the same lexical item (see also Hirsh, 1998).

4. Howard, Patterson, Franklin, Orchard‐Lisle, and Morton (1985) proposed that neologisms can reflect distortions of lexical errors.