Search

There is a lack of modern quantitative language assessment tests in Russian, integrating neuropsychological and psychometric traditions, and allowing to specify the type and severity of linguistic deficits in individuals with different aphasia profiles. In response to these clinical and research needs, a novel standardized aphasia test – the Russian Aphasia Test (RAT) – is currently being developed. The principal novelty of this test is that each subtest corresponds to a specific level of linguistic processing in one of the four language domains: auditory comprehension, oral production, reading, and writing. In selection of specific tasks for each level of processing we took into account the structure and materials of contemporary standardized aphasia tests in other languages (e.g., Comprehension Aphasia Test, Howard et al., 2010; Psycholinguistic Assessment of Language Processing in Aphasia, Kay et al., 1996), as well as modern (psycho)linguistics theories, and the structural and phonetic specifics of the Russian language. Here we present data on preliminary standardization of auditory comprehension subtests that included the following tasks: • Phonemic – judgment of minimal pairs of nonwords (n=100) and words (n=72). The following factors were taken into account in the design of the subtest: phoneme manner and place of articulation, syllabic structure and position, frequency and imageability (for words). • Lexical – lexical decision (n=120). Lexical frequency, word length, and the degree of similarity of non-words to real words were taken into account. • Lexical-semantic processing – word to picture matching for objects (n=67) and actions (n=68). Items were selected based on naming and image agreement. • Syntactic – sentence to picture matching for various syntactic constructions (n=68), including reversible passives, subject and object clefts, and prepositional phrases. • Discourse – comprehension of orally presented stories of varying lexical-semantic and syntactic difficulty indexed by response accuracy to a set of 16 yes-no questions on explicit and implicit content of the stories. Each subtest was completed by at least 20 individuals with aphasia and 20 healthy age-matched controls. To maximize validity and reliability of the subtests, “poor” items were removed according to the following principles. First, items that were answered erroneously by two or more healthy participants were eliminated. Second, based on item difficulty and corrected-item-total correlation derived for the aphasia group, most sensitive and valid items were retained. Also, it was ensured that each influential psychometric property was represented in the refined set by a wide range of values. Thus, a final smaller set of items for each subtest was selected for further norming and standardization. We will discuss in detail the various conceptual and psychometric considerations that went into the design of the subtests and affected item selection. This project is supported by Russian Scientific Foundation for Humanities, grant №14-04-00596.

Introduction Deficits in working memory (WM) are amongst the most widely acknowledged cognitive impairments in aphasia.However, there is still on-going debate what tasks should be used to assess WM in aphasia (Wright & Fergadiotos, 2012). The two main alternatives for this purpose are simplified complex span tasks and N-back tasks. In a typical complex span task, a processing task (e.g., sentence reading), is given along with a set of stimuli (e.g., words) to be remembered for later recall. In N-back tasks, participants are instructed to judge whether an item matches a previous one presented n items before. Proponents of complex span tasks state that these tasks are the gold standard for assessing WM capacity in cognitive psychology and that variations of these tasksare endorsed as valid means of indexing WM capacity within different theoretical frameworks. On the other hand, researchers using N-back tasks state that since these tasks are more language-free in nature, they are more appropriate for indexing cognitive non-linguistic abilities with language-impaired populations. Several investigations comparing the two tasks in healthy controls have demonstrated no relationship between the two tasks (Jaeggi et al., 2010; Kane et al., 2007). Although, limited conflicting findings indicating a significant relationship between the two tasks have also been reported (Schmiedek et al., 2009). Additionally, performance on complex span task in aphasia studies have repeatedly been related to performance on standardized language tests (Sung et al. 2009; Wright & Fergadiotos, 2012), while similar correlations have never found for N-back tasks (Christensen, & Wright, 2010; Mayer & Murray, 2012). The aim of the present study was to directly investigate the relationship between performance on complex span tasks and N-back tasks in aphasia. Methods 32native speakers of Russian with aphasia following left hemisphere stroke participated. Complex span task (modified listening span task, Ivanova & Hallowell, 2014), 2-back and 0-back tasks with words, a test of auditory language comprehension– Quantitative Assessment of Speech in Aphasia (QASA; Tsvetkova et al., 1981) – were administered. Results and discussion No significant correlations were observed between performance on complex span task and N-back tasks.Furthermore, performance on the modified listening span was related to performance on the comprehension subtest of the QASA, while no relationship was found for 2-back and 0-back tasks.Our results mirror studies in healthy controls that demonstrated no relationship between performance on the two tasks(Jaeggi et al., 2010; Kane et al., 2007). Thus although N-back tasks seem similar to traditional complex span measures and may also index abilities related to cognitive processing, the evidence to date does not warrant their direct association with the construct of WM. Implications for future investigation of cognitive deficits in aphasia will be discussed.