A Swedish Cookie-Theft Corpus

A Swedish Cookie-Theft Corpus Dimitrios Kokkinakis1, Kristina Lundholm Fors1, Kathleen Fraser1, Arto Nordlund2† Department of Swedish1; Department of Psychiatry and Neurochemistry2 University of Gothenburg, Sweden {dimitrios.kokkinakis; kristina.lundholm; kathleen.fraser}@svenska.gu.se {arto.nordlund}@neuro.gu.se Abstract Language disturbances can be a diagnostic marker for neurodegenerative diseases, such as Alzheimer’s disease, at earlier stages. Connected speech analysis provides a non-invasive and easy-to-assess measure for determining aspects of the severity of language impairment. In this paper we focus on the development of a new corpus consisting of audio recordings of picture descriptions (including transcriptions) of the Cookie-theft, produced by Swedish speakers. The speech elicitation procedure provides an established method of obtaining highly constrained samples of connected speech that can allow us to study the intricate interactions between various linguistic levels and cognition. We chose the Cookie-theft picture since it’s a standardized test that has been used in various studies in the past, and therefore comparisons can be made based on previous research. This type of picture description task might be useful for detecting subtle language deficits in patients with subjective and mild cognitive impairment. The resulting corpus is a new, rich and multi-faceted resource for the investigation of linguistic characteristics of connected speech and a unique dataset that provides a rich resource for (future) research and experimentation in many areas, and of language impairment in particular. The information in the corpus can also be combined and correlated with other collected data about the speakers, such as neuropsychological tests, brain physiology and cerebrospinal fluid markers as well as imaging. Keywords: Speech elicitation corpus, Cookie-theft stimuli, dementia, linguistic features, language impairment 1. Introduction and Background Elicitation of spontaneous speech and narrative discourse samples from (young and/or adult) individuals both healthy as well as with a variety of mental and cognitive impairments, at various stages of the pathology and with a variety of stimuli, have been developed and studied for several decades (e.g., Bottenberg et al., 1987; Bryant et al., 2016). Such systematic form of language production can be a valuable procedure for describing and even quantifying the degree and severity of e.g. cognitive decline, which most often1 incorporate a single picture description, a picture sequence description (e.g., the Dog Story; de Lira et al., 2011) or a personal opinion on an event or idea. Even work on normative data has been reported for use in various narrative discourse tasks. For instance, based on a large control sample, Richardson & Dalton (2015) created main concept checklists (information content units) for three semi-spontaneous discourse tasks, a picture sequence narrative Broken Window, a storytelling Cinderella, and a procedure Peanut Butter and Jelly. For a similar purpose, Catricalà et al. (2017) created their own original picture Summer Time used as normative data on 134 Italian subjects pooled across homogeneous subgroups for age, sex, and education. Nonetheless, the single black and white picture of the Cookie-theft from the Boston Diagnostic Aphasia Examination Battery (BDAE; Goodglass et al., 2001) is one of the most used stimuli for the assessment of language production. Other images, such as …

Content truncated to 500 words