TalkBank

AphasiaBank

This page provides information about approaches to discourse analysis with specific focus on analyzing AphasiaBank protocol tasks. It also provides some results of analyses that have been conducted on AphasiaBank transcripts. We are grateful to investigators who have shared their work.

Some basic approaches to discourse analysis that have been applied to the AphasiaBank discourse data:

• Coherence
• Core Lexicon
• Correct Information Unit (CIU)
• C-NNLA -- automated Northwestern Narrative Language Analysis
• C-QPA -- automated Quantitative Production Analysis
• Main Concept Analysis
• Story Grammar
• Systemic Functional Linguistic Analysis

Some additional publications with discourse norms on tasks from the AphasiaBank standard discourse protocol:

• Main events, Cat in the Tree picture description, Wright, Capilouto, Wagovich, Cranfill, & Davis, 2005
• Content units, Cookie Theft picture description, Yorkston & Beukelman, 1980

Discourse analysis results from AphasiaBank transcripts:

• Fergadiotis-2018 -- The Excel spreadsheet includes the following data for 118 PWA, as described in Fergadiotis, Kapantzoglou, Kintz, and Harris (2018): CIUs and Total Number of Words for Free Speech, Cinderella, and Refused Umbrella portions of the AphasiaBank standard discourse protocol; scores from the BNT, WAB-R Object Naming subtest, and VNT.
• Cunningham-2020 -- The Excel spreadsheet includes the following data for 225 PWA, 27 participants with aphasia who tested above the WAB-R AQ cutoff, and 225 controls, as described in Cunningham & Haley (2020): word count, Word Information Measure (WIM), Moving-Average Type-Token Ratio (MATTR-5) for the Cinderella task from the AphasiaBank standard discourse protocol; scores from the WAB-R AQ subtests. See "Codes" tab in spreadsheet for github link to scripts for calculation of WIM and MATTR-5.