A quantitative production analysis of picture description

I would like to express my gratitude to my tireless research assistants—Kristin Lilly, Victoria Tumanova, and Kari Norman—for their hard work; to Rita Berndt for her patience with endless questions; and to the participants and staff at the North York Aphasia Centre for their generous cooperation. I am also grateful to the American Speech‐Language‐Hearing Foundation and to Division 2 (Neurophysiology and Neurogenic Speech and Language Disorders) of the American Speech‐Language‐Hearing Association for funding to support this research.

Abstract

Background: The Quantitative Production Analysis (QPA) (Berndt et al., 2000; Saffran et al., 1989) was developed to provide an objective means of characterising the grammaticality of spontaneous speech in aphasia. However, there is still no consensus as to the profile of performance, on this or any other measure, that identifies agrammatism.

I would like to express my gratitude to my tireless research assistants—Kristin Lilly, Victoria Tumanova, and Kari Norman—for their hard work; to Rita Berndt for her patience with endless questions; and to the participants and staff at the North York Aphasia Centre for their generous cooperation. I am also grateful to the American Speech‐Language‐Hearing Foundation and to Division 2 (Neurophysiology and Neurogenic Speech and Language Disorders) of the American Speech‐Language‐Hearing Association for funding to support this research.

Aim: The aim of this study is to advance the effort to characterise agrammatic speech production, by examining the influence of aphasia severity and fluency on patterns of QPA scores obtained from a broad range of individuals with aphasia. The goal of this process is to identify the quantitative variables most predictive of agrammatism.

Methods & Procedures: Spontaneous speech samples were elicited from non‐brain‐damaged (NBD) older subjects and a group of participants with aphasia unselected by diagnosis, and analysed using the QPA. Data from the NBD and aphasia groups were statistically compared and correlated with ratings of severity. Fluent and non‐fluent subjects within the group with aphasia were then compared. Last, the non‐fluent group was explored for evidence of agrammatism.

Outcomes & Results: Many of the QPA's measures were found to reflect the severity of aphasia and/or the fluency of speech production, suggesting that results on these measures may not be indicative of agrammatism per se. In particular, measures of output productivity (the amount and efficiency of production) were largely determined by aphasia severity, and measures of lexical selection were related to the fluency of production. Profiles of the non‐fluent subjects also varied widely, reinforcing the dissociability of different grammatical functions.

Conclusions: None of the QPA's measures could be interpreted to reflect agrammatism alone; in fact, many appeared to be reflective of aphasia severity or fluency. Diagnoses of agrammatism still depend on clinical judgement; however, this judgement is aided greatly by the array of objective quantitative information provided by the QPA.

Notes

I would like to express my gratitude to my tireless research assistants—Kristin Lilly, Victoria Tumanova, and Kari Norman—for their hard work; to Rita Berndt for her patience with endless questions; and to the participants and staff at the North York Aphasia Centre for their generous cooperation. I am also grateful to the American Speech‐Language‐Hearing Foundation and to Division 2 (Neurophysiology and Neurogenic Speech and Language Disorders) of the American Speech‐Language‐Hearing Association for funding to support this research.