Short-term Memory Assessment in Patients With Poststroke Aphasia

1. Introduction 1.1. Background The literature has witnessed a steady increase in the number of studies investigating the role of short-term memory (STM) and working memory (WM) impairments in aphasia patients, which clearly shows that aphasia patients have impairments in verbal STM and WM. Throughout the years, views on the nature of aphasia have changed from the loss of linguistic representations (e.g., semantic and phonological representations) into impairments in the processing (i.e., access and retrieval) of language representations. Therefore, the emphasis of research has been refined in identifying the component operations of language processing, such as the mechanisms behind the access and retrieval of words or the temporal course of language processing. These operations rely on other cognitive functions, such as the STM. Psycholinguistic models describe the role of the STM in language as the maintenance of activation of linguistic representations over the time course of processing words. Particularly, the interactive activation model takes levels of linguistic representations (e.g., semantic, lexical and phonological) and its connections into account. Specifically, according to this influential model, word retrieval is mediated by interactive spreading activation that activates the target word and semantically or phonologically related words. Two important concepts (i.e., connection strength and decay rate) play an important role in spreading and maintaining activation over time in the semantic-lexical-phonological network. In aphasia, the impairments are thought to occur through deficits in linguistic representations, and moreover, through deficits in connection weight and decay rate (i.e., processing impairment). Errors occur if connection strength is weak (activation spreads too slowly) or if the decay rate is too rapid (representations are activated but decay too fast). Therefore, the interactive activation model can explain different patterns of word retrieval errors (i.e., semantical or phonological) in aphasia. In order to examine the role of the STM in language processing, there is an excessive need for reliable assessment tools. However, the psychometric properties of tests commonly used to assess STM problems in individuals with aphasia have been called into question. A common issue across these tests is that small standardized samples are used, which lead to validity and reliability problems of normative data. This raises some concerns since there is growing awareness that impairments in STM functions may be predictive of certain aspects of language function, recovery and rehabilitation in aphasia patients. Therefore, it is of crucial importance that STM assessment tools are reliable and valid in order to replicate, compare and further examine these findings. A recent study developed a new English assessment tool (i.e., The Temple Assessment of Language and (Verbal) Short-term Memory in Aphasia; TALSA) that evaluates language and verbal STM/WM abilities in individuals with poststroke aphasia in an interactive activation framework. In other words, to evaluate the nature of aphasia as a processing disorder. The TALSA battery is unique compared to other tests (e.g. Psycholinguistic assessments of language processing in aphasia (PALPA)) since it takes into account features related to temporal aspects of language processing and effects of memory load on production and comprehension of language. Therefore, the TALSA battery is a specific diagnostic assessment tool and is sensitive to language impairments at all levels of severity, including mild aphasia. The complete TALSA battery consists of three parts among which the last part involves verbal STM tasks. Note that the authors of the TALSA battery are still working on a clinical version that is not published yet. 1.2. Rationale of the study To our knowledge, the existence of a Dutch assessment tool that examines verbal STM in individuals with poststroke aphasia is lacking. In other words, a Dutch STM test specifically developed for people with phatic disorders is not available. However, assessing non-linguistic influences in aphasia patients, such as the role of STM on language components, is of crucial importance as communication in everyday life fluctuates in processing time. Moreover, an overwhelming amount of STM measures (standardised and experimental) are used in aphasia research. However, diversity in STM tests makes it harder to compare discrepant findings and makes it more difficult for clinicians to select appropriate assessment tools. Furthermore, the validity and reliability of used STM tools are questioned in the aphasic population. Thus, the establishment of a Dutch verbal STM assessment tool is crucial to better comprehend the nature of aphasia as a processing disorder, and has the purpose to serve as a practical tool to help Dutch clinicians and speech pathologists diagnose STM deficits more specifically, which can lead to optimized speech therapy. 1.3. Objective of the study The first aim of the present study is to develop a Dutch clinical verbal STM assessment tool, based on the third part of the TALSA battery, to assess STM deficits at all levels of severity. With this intention, the investigators hope to better comprehend the nature of aphasia and the impact of STM deficits which may ameliorate treatment techniques. Importantly, this test focuses on clinical use and administration, in which despite the patient's language problems, the test must be understandable and applicable (e.g., test duration and response modality). The second aim is to run a pilot test on the aphasic population and the healthy population. Running a pilot test will validate the stimuli and the wording of the tasks. Confounded items (e.g., ambiguous stimuli, poor wording) will be revised or eliminated if necessary. The third aim is to evaluate the verbal STM assessment tool with respect to important psychometric properties, as the methodological quality of a test affects the interpretation of the results.
2. Method

2.1. Study population

The investigators will include a maximum of 100 patients with aphasia in the acute to subacute phase after an ischemic or hemorrhagic stroke. The patients will be directly recruited in the Ghent University hospital at the Stroke unit. Furthermore, 20 healthy controls will be recruited via social media platforms. Healthy participants will be matched to the aphasic group for gender, age and education. All participants will voluntarily participate in this research project and sign an informed consent.

2.2. Verbal STM assessment tool development (first aim)

The first step towards the development of a Dutch clinical verbal STM assessment tool is to carefully select the most crucial verbal STM subtests of the original English TALSA battery. This is because the original TALSA battery is quite long and is not yet developed for administration in clinical settings. In particular, the Dutch verbal STM assessment tool will be based on the third part of the TALSA battery which involves verbal STM span tasks. The third part of the TALSA battery, moreover, evaluates phonological and semantic language representations by varying word characteristics of the items. Thus, the Dutch verbal STM assessment tool will consist of verbal STM span tasks with linguistic variations (phonological and semantic representations).

The stimuli in the original TALSA battery consist of pictures and recordings of words, nonwords, sentences, and sequences of words, nonwords, and digits. However, literally translating the TALSA test-battery is inappropriate because there is not a one-to-one match between words and syntactic structures across any two languages. In particular, it is important to take into account word characteristics, such as word frequency, syllables, imageability and age of acquisition, that influence the degree of success in word retrieval in aphasia. Therefore, the verbal STM assessment tool will rely on the same subset tasks of part 3 of the TALSA battery, but with the careful selection of new Dutch stimuli with respect to important word characteristics.

In the original TALSA battery, the authors took word frequency, syllables and imageability into account. However, research showed that age of acquisition (AoA) is also a crucial word characteristic in success of word retrieval in aphasia, thus AoA will be included during the stimuli selection procedure. In order to collect stimuli that meet these general word characteristics, word databases will be consulted, such as the Dutch Lexicon Project 2.

2.3. Pilot test (second aim)

Once the test-battery is developed, pilot testing will be of crucial importance in the clinical setting and healthy population to adapt and improve the test-battery where necessary.

Particularly in the clinical setting, patients with aphasia will be recruited according to the inclusion and exclusion criteria. Specifically, all patients who are eligible will be asked for informed consent to participate in this trial. Subsequently, the Oxford Cognitive screen, the Token Test and the Dutch verbal STM assessment tool will be administered.

The Token Test determines the severity of aphasia and investigates sentence comprehension. The Oxford Cognitive screen is a short cognitive screening tool that is administered in the acute phase of stroke. The outcomes on the Token Test and Oxford Cognitive Screen can serve as a support to interpret the results of the Dutch verbal STM assessment tool.

It is worth noting that the Dutch verbal STM assessment tool, the Token Test and the Oxford Cognitive screen will be administered over a time span of maximum two consecutive days depending on how fatigued the patient is during testing. The total test duration on patients is estimated to take approximately two hours. The total test duration on healthy controls is estimated to take approximately 1 hour. During administration of the Dutch verbal STM assessment tool, audio recordings will be made of the patient. In this way, it is possible to reevaluate misunderstandings during the scoring of the subtests.

In order to verify or adapt the difficulty of the items, it is critical that the Dutch verbal STM assessment tool will also be tested on healthy controls. Results of the original TALSA battery revealed that healthy controls performed near ceiling on most of the subtests. However, in some of the subtests (e.g., nonword span), healthy controls showed a similar results pattern as aphasia patients, which questions the contribution of those subtests.

To conclude, running a pilot test provides the opportunity to reevaluate items that stand out from the distribution of scores (e.g., sufficient number of healthy controls respond incorrectly to an item). Specifically, these items may be confounded and need to be eliminated or adapted (e.g., ambiguous stimuli).

2.4. Psychometric properties (third aim)

Reliability and validity are important psychometric properties. Reliability refers to the consistency, stability and accuracy of an assessment tool. Validity of a test refers to whether the assessment tool measures what it intends to measure.

• Internal consistency: To verify the reliability of the Dutch verbal STM assessment tool, the internal consistency of the test will be estimated using the split-half reliability procedure. In this way, the constancy of results across items within a subtest can be analyzed.
• Face validity: The TALSA battery is developed to approach aphasia as a processing disorder, supported by cognitive models of language processing. Therefore, the test battery focuses on component stages of word processing (e.g., semantic or phonological) and, moreover, with an emphasis on processes related to access, retrieval and short-term maintenance of language representations.
• Concurrent validity: The results of the Dutch verbal STM assessment tool will be compared with the results of the original TALSA battery to verify the concurrent validity. High correlations between the subtests of the original TALSA battery and Dutch verbal STM assessment tool suggest a good concurrent validity.

To conclude, throughout the process of pilot testing, the Dutch verbal STM assessment tool will be modified and improved where necessary and will be psychometrically evaluated (e.g., reliability, validity and item analysis).