INTRODUCTION
Aphasia is a debilitated condition that hampers the patients’ communication. This condition can be due to stroke, and approximately 30% of stroke survivors may be prone to develop aphasia following a stroke (Sekine et al., 2013). A large pool of linguistic research is conducted to understand language performance in persons with aphasia (PWAs), which may aid speech-language pathologists (SLPs)/researchers to get better insight during their rehabilitation. Previous linguistic aphasia research studies mainly focused on processing language components at the word or phrase level in the absence of the context (Armstrong, 2000). However, in recent times, there has been a paradigm shift, and researchers have appraised language at the sentence level (Bryant et al., 2016). Also, they aim to understand the cohesion and coherence of language via discourse tasks in stroke survivors with language impairment to improve aphasia assessment in clinical practices (Richardson & Dalton, 2020). Richardson & Dalton (2020) conducted a study focusing on creating a checklist for analysing the main concepts in narratives elicited by pictures prompts involving a cat in a tree and a person refusing an umbrella elicited narrative tasks in neurotypical healthy adults across age groups 20~39, 40~59, 60~79, and >80. The findings of the study evinced the normative range for each age group for each picture individually. These results aid SLPs and researchers in ascertaining the extent of deviation in discourse skills from normalcy in PWAs.
Discourse can be elucidated as one of the units of language that goes beyond the sentence level. This encompasses the use of various linguistic structures such as phonology, lexicon, syntax, semantics, and discourse markers. These linguistic structures aid individuals to convey appropriate meaning of the context (Cavanaugh & Haley, 2020; Stemmer, 1999; Wright, 2011). Discourse-based studies in PWAs were based on structural and functional aspects of discourse per se. Former aspects of discourse were elucidated as it is solely based on the measures of syntax and lexical production during picture description or while describing personal narrative. The latter discourse measure was focused on how meaning is organized at the pragmatic interactive level (Armstrong, 2000; Pritchard et al., 2017).
One of the widely exploited methods to obtain representative speech samples in PWAs is the picture description task which is frequently included in language standardized tests (Vandenborre et al., 2018). It offers several advantages over conversational analysis, such as obtaining an object reference to the conversational topic (Fergadiotis et al., 2011), and memory and sustained attentional demands are minimized (McNeil et al., 2004). In comparison to picture description, storytelling is deemed to increase lexical productivity and propositional density, which gives rise to what we might call expository or descriptive discourse (Alyahya et al., 2022; Stark, 2019). Despite storytelling being the pivotal task to measure discourse, this task could be challenging in PWAs owing to the involvement of memory and planning per se. In this line, a study by Stark et al.(2023) ascertained PWAs had a reduced type-token ratio and reduced words per minute during narrative tasks relative to other forms of discourse genre.
Discourse in aphasia
Discourse varies with different elicitation methods (Doyle et al., 1995; Roberts & Wertz, 1989). Roberts & Wertz(1989) aimed to compare the structured and unstructured language tasks in PWAs. A total of 20 PWAs were enrolled in the study and their language sample were audio-video recorded. In specific to the elicitation of language samples, conversation tasks and verbal descriptions of the object were used and the same tasks were compared across 1, 3, 6, 9, and 12 months post-stroke onset. The study findings showed there is a difference between the two different elicitation methods.
Understanding the relationship between different discourse elicitation methods and their influence on language is deemed pivotal. These aid researchers/SLPs in designing the effective treatment plan for PWAs (Shadden et al., 1991). In a similar line, a study by McCullough et al.(2017) aimed to compare the language demand imposed for structured and conversation discourse tasks in PWAs. Ten PWAs inclusive of both fluent and non-fluent aphasia were recruited for the study and discourse was analysed using the correct information unit. The findings of the study showed that there is a significant difference in discourse elicitation tasks (structured versus conversation tasks).
In addition, McCullough et al.(2017) ascertained fluent aphasia outperformed non-fluent aphasia in the conversation task, but no significant difference was evinced between fluent and non-fluent aphasia for the structured task. The above findings shed light on differences in the task demand and its influence on the language. This finding acknowledges the fact that each discourse elicitation method yields distinct discourse abilities. Thus, there arises a need to conduct research in this line to a large extent.
Discourse analysis is an important component of aphasia assessment because it can provide insight into functional communication abilities. The purpose of the current study is to explore the possible differences in discourse-level communication in persons with fluent aphasia (PwFA). Discourse analysis provides a rich data source to gauge the linguistic content/structure of language use in a more naturalistic way (Hazamy & Obermeyer, 2020).
Previous studies on fluent aphasia posit that lexical and semantic impairment is predominantly noted in PwFA (Edwards, 2005). Also, researchers noted they manifest a larger amount of grammatical impairment during discourse production (Bastiaanse et al., 1996). This necessitates the researcher to understand the potency of different discourse genres specific to fluent aphasia. Further, Andreetta et al.(2012) compared the cohesion of anomic aphasia with their age-matched control group. The findings of the study portrayed a large number of cohesion errors during a narrative discourse task in persons with anomic aphasia compared with the age-matched control group.
On similar lines, Manning & Franklin(2016) compared the narrative discourse abilities in persons with fluent and nonfluent aphasia. The findings of the study unveil that there is a difference in macro and microstructural aspects of discourse. In specific, non-fluent aphasia are more prone to have higher rates of subject/object omissions and article errors than fluent aphasia. Fluent aphasia produced errors concerning information being presented in illogical order with more pronoun difficulty compared to non-fluent aphasia. All these findings alluded to earlier were unable to show statistical differences. The study results shed light on the need for establishing comparative norms across different types of aphasia. Ascertaining the commonalities and differences among the variant types of aphasia is paramount, owing to the identification and selection of appropriate assessment and intervention approaches in PWAs.
Need for the study
Previous literature on discourse-based seems inconclusive, owing to different methods employed to analyse the discourse, single case description, and distinct linguistic profile (Zhang et al., 2021; Marangolo et al., 2014). Discourse samples can be analyzed using different elicitation procedures. Out of the discrete discourse elicitation procedures, the picture description task stands out as a common and frequently used procedure. Storytelling activities are carried out in a way by narrating the individual’s personal experience or by narrating familiar stories. This activity is employed more in recent studies than in regular clinical practice (Andreetta et al., 2012). Storytelling/narration is also known to be a more naturalist way of eliciting discourse. Personal narratives result in inconsistencies both across and within participants, due to heterogeneity of experience of each individual. Procedural description (e.g., preparation of oats/coffee) involves a natural mode of connected speech. However, in most of the aphasia literature or clinical practice, the procedural description task was under-utilized.
Furthermore, the difference across the distinct genres of discourse is more explicitly elucidated in the theoretical framework of discourse production (Sherratt, 2007). Sherratt(2007) postulates the fact that discourse takes place at different stages, and it varies with different genres. The literature on discourse genre tries to understand the potent of each genre in isolation, only a handful of studies tried to compare the superiority of genres in PWAs (e.g., Stark, 2019; Armstrong et al., 2011). In addition, most of the studies evinced paradoxical findings in establishing the relationship between discourse and language impairment. (e.g., Olness & Ulatowska, 2011). In addition, understanding the insight of discourse becomes pivotal, as it is the most rudimentary way in which individuals have their conversation in their daily day-to-day routines. (Andreetta et al., 2012). Thus, to bridge these alluded research gaps the present study aimed to uncover the superiority of different discourse elicitation tasks in PwFA.
MATERIALS AND METHODS
Participants
In the current study, a total of six PwFA were recruited, who were recruited from in and around Mangalore district, India via a convenient sampling (Table 1). All the participants gave their consent to be part of the study. The study received ethical clearance from Father Muller Charitable Institutions (FMCI) (FMMCIEC/CCM/611/2022).
The current study followed stringent criteria to recruit the participants. 1) The PwFA should have a minimum of 6 months post-stroke onset. 2) PwFA should have at least cleared their matriculation with fairly good reading skills in Kannada. 3) PwFA should possess normal hearing/good vision/corrected vision. 4) PwFA should have a left middle cerebral artery lesion. 5) PwFA should be free from cognitive impairment were ruled out using Montreal Cognitive Assessment (Nasreddine et al., 2005). And 6) PwFA with psychiatric issues were excluded.
Materials
In the current study pictures of ‘broken windows’, ‘cat rescue’, and ‘refused umbrella’ were used for the picture description task. Western Aphasia Battery in Kannada was used to gauge the Aphasia Quotient and type of aphasia. Audio-video recorder (QIWA® Wireless HD 960P; AUSHA overseas, Bangalore, India) was used to record the discourse sample.
Procedure
PWAs were assessed across three tasks namely; picture description, conversation, and narration. In the picture description, the PwFAs were asked to describe ‘broken windows’, ‘refuse umbrella’, and ‘cat in the tree’ pictures (Nicholas & Brookshire, 1993) (Appendix 1). The description of pictures was solely based on PWA’s discretion. In this task, verbal prompts or cueing were not encouraged. In conversation task, PwFAs, were probed questions about their hobbies/daily routine/current affairs. Here, PWAs were given adequate time (60~80 seconds) to respond to each question. If they were unable to maintain the conversation then the researcher tried having a conversation with a different topic. If still PwFA fails then the researcher moves to the next task without giving any feedback. In the narration task, PwFAs were asked to narrate the events of their interest namely, stories, explaining their job profile, and explaining their favorite book/books they have read. All these discourse samples were audio-video recorded and verbatim of each sample were transcribed and later used for discourse analysis.
Computing discourse via the aforementioned task was further subjected to analysis using the discourse analysis scale (Hema & Shyamala, 2008) (Appendix 2). In all these tasks, the propositional and non-propositional aspects of communication were evaluated. Discourse structure, communication intent, topic management, and so on are some of the propositional aspects of communication. Repair strategies, revision behaviour, turn-taking, etc., are some of the non-propositional aspects of communication. After recording the responses, the scores of each task (picture description, conversation and narration) were consolidated to compute the discourse quotient (DQ).
For the conversation analysis there are total of ‘38’ parameters, inclusive of both propositional and nonpropositional aspects of discourse. Each parameters receive maximum score of ‘2’, thus maximum score participants can obtained from this task is ‘76’. Subsequently, for the narration task, there are total of ‘26’ parameters, hence, the participants can score maximum of ‘52’ in this task. In picture description task, there were total of ‘28’ parameters, hence the participants can score maximum of ‘56’ in this task. For instance, if a PWA scores 48/56 in the picture description task, then DQ for the picture description task would be 61.53% (48/56 × 100 = 85.71%).
Scoring
Each discourse sample across the tasks was scored based on the discourse analysis scale (Hema & Shyamala, 2008). It is a three-point rating scale, 0 is poor, 1 is fair, and 2 is good. Tasks alluded to earlier were scored based on a three-point rating scale (Appendix 1B).
Data and statistical analysis
The computed data was fed into SPSS software for statistical analysis analysis version 18 (IBM, Chicago, IL, USA). The Friedman test was used to compare the statistical differences across different discourse genres, and further Wilcoxon signed ranked test was used to compare the pair wise differences among the discourse tasks. Data was also subjected for reliability assessment using intra-class correlation coefficient (ICC) to obtain inter-rater and intrarater reliability for the discourse tasks (narration, picture description, and conversation).
Reliability measures
5% of the discourse sample was given to the ‘2’ experienced speech language pathologists for assessing the effectiveness of discourse quotient. The findings revealed ICC of 0.85 (95% confidence interval [CI], 0.65 to 0.95), indicating excellent reliability. Further, intra-rated reliability revealed an ICC of 0.90 (95% CI, 0.70 to 0.98), indicating higher consistency in the rater’s assessments.
RESULTS
The data was fed into SPSS software version 18 (IBM) to compute descriptive and quantitative analysis. The mean and standard deviation for the general conversation, narration, and picture description tasks were computed. Pronounced scores were evinced for general conversation compared to narration and picture description tasks. On comparing narration and picture description, narration scores were higher compared to picture description (Figure 1). Subsequent to group analysis, the data was analyzed individually via qualitative description. This analysis aimed to understand the subtle changes within each participant.
The data were subjected to statistical analysis using the Friedman test, this test was applied to gauge the overall difference across the tasks. The findings of the study evinced significant differences across the general conversation, narration, and picture description tasks (χ2(2) = 9.47; p = 0.09). Further, the Wilcoxon sign ranked test was applied to carry out a pair-wise comparison. The results of this test revealed a significant difference across general conversation versus narration, and general conversation versus picture description. However, the researcher ascertained no significant difference while comparing picture description and narration tasks (Table 2).
Individual data analysis
Individual analysis revealed that the partcipant (P)1, P2, and P6 scores for general conversation were higher followed by narration and picture description tasks (Figure 2). The participants P4 and P5 evinced pronounced scores for general conversation, but the equivalent scores were noted for narration and picture description tasks. Participant P3 scores were higher in conversation; however, the robust score was noted for picture description than narration. Overall, all these participants outperformed in the conversation task compared to the narration and picture description task.
DISCUSSIONS
The study aims to shed light on the superiority of three distinct discourse tasks (conversation, narration, and picture description) in PwFA. The findings of the study partially supported the hypothesis of the study with no significant difference between narration and picture description tasks. On the other hand, the findings of the study rejected the hypothesis, where significant difference was manifested across conversation versus narration tasks; and conversation versus picture description tasks.
The experimental paradigm conceived in the study seems to be novel. The literature in this line of research is scanty. Hence, the current study findings are discussed merely based on discourse-based studies in PWAs. Conversation tasks manifested more robustness than other tasks (narration and picture description) which could be asserted to lack of density of language requirement (Stark, 2019). That is, during conversation PWAs respond merely to the questions posed by the researcher. Hence, they would be able to use only limited vocabulary, limited phrases, and reduced nouns/verbs compared to narration tasks. Owing, to which conversation tasks impose lesser taxes on the linguistic abilities of PWAs, resulting in better discourse scores relative to other tasks.
Conversation tasks utilize lesser processes of memory relative to narrative tasks and picture description tasks. In narrative tasks, there is the involvement of long-term and working memory to a larger extent. In addition, executive functioning skills such as planning and programming would play a pivotal role in narration compared to conversation tasks. Owing to the aforementioned elucidations, it is perceptible that linguistic complexity varies with conversation, narration, and picture description tasks (Stark, 2019). This might enable PWAs to perform better at conversation tasks than narration and picture description tasks.
According to cognitive load theory Sweller et al.(2011) cognitive load varies across tasks like conversation, narration, and picture description. Conversation usually entail processing information in real time with a focus on interaction. Narration task requires individuals to construct and recall events in a coherent fashion, often triggering working memory. Picture description task may require recalling details from long-term memory and organizing those into speech, this could be more demanding than simple exchange of information in conversation.
Robust performance in conversation tasks is also in accordance with the studies (Beeke et al., 2003 & 2007). Beeke et al.(2003 & 2007) have also manifested differences in the use of particular grammatical forms during monologue discourse and conversation. Also, they asserted the fact that PWAs use more complete grammatical forms during monologue discourse than conversation. Conversation tasks may aid PWAs to have assistance from a partner and enable participants to get contextual cues for effective communication. The priming effect can manifest during the conversation which in turn, may excrete lesser linguistic demand than other discourse-based tasks (Armstrong et al., 2011). Moreover, in the current study majority of the participants recruited exhibited anomic aphasia, as this type of aphasia generally performs well in the context of cues (Abhishek et al., 2018). Thus, when a conversation task is rendered to persons with anomic aphasia this might have facilitated contextual cues. Owing to this the superiority in performance of conversation tasks can be noted compared to narration and picture description tasks.
On similar lines, Bryant et al.(2016) aimed to investigate the discourse abilities of PWAs with their age-matched control group. In this study, authors used picture description tasks and sequence pictures to assess the discourse abilities. The findings of the study displayed the difference in discourse skill across picture description and sequence picture tasks. The sequence picture task resulted in more words per minute, whereas the picture description task elicited denser language with a pronounced noun-verb ratio. These results showed that discourse varies across the tasks and the type of stimuli presented. Also, the authors concluded that the picture description task is more suitable for moderate to severe forms of aphasia, and also for individuals who heavily rely on visual modality for lexical retrieval.
On the other hand, the current study findings are contraindicated with few other studies (e.g., Beeke et al., 2003; Heeschen & Schegloff, 2003). They asserted the fact that PWAs tend to produce more language with greater linguistic load during monologues than dialogues.
Further, the data was analysed individually and authors noted different trends of results. Wherein, participants P1, P2, and P6 evinced increased scores for conversation tasks followed by narration and picture description tasks. This could be alluded to earlier elucidation of how conversation yielded better discourse scores compared to the others.
Participants P4 and P5 also discerned higher scores in conversation, but scores for narration and picture description remain equipotent. This could be due to participants’ word retrieval deficits, lexical deficits, and syntactic deficits; these would have hampered their performance in specific narration and picture description tasks rather than conversation tasks. That is, narration and picture description are deemed to be constraint-based tasks, where PwFA should be able to use words and sentences specific to the topic/pictures. Thus, discourse scores for narration and picture description are impoverished. Participant P3 showed an atypical pattern compared to other participants in the study, i.e., P3 showed higher scores in conversation followed by picture description and narration tasks. This finding can be explained based on the visual cues that the participant is exposed to his lexicon activation is confounded to that particular picture/context alone without activation of neighbourhood lexicons. Thus, performance may be ameliorated by the usage of correct words and sentences. In addition, it can be posited that P3 might have better lexical retrieval in the presence of visual cues. P3 result findings should be interpreted with caution owing to anecdotal reports.
Discourse analysis is essential to the assessment of aphasia due to its ability to illuminate functional communication abilities in PWAs. The findings of this study suggest conversation scores outperformed narration and picture description tasks. The study findings would serve as important insight for both clinical and research aspects in understanding the potency and setbacks of different discourse genres. The different tasks used to assess discourse in PWAs, ascertained about the complexity of the task. This information aids SLPs to choose the appropriate discourse task based on the severity of aphasia. The study also highlights the importance of using multiple discourse tasks in order to gain a comprehensive insight on the discourse skills of the participants. Findings of the study highlight the fact that different discourse task should poses different cutoff points/different analytical approach for accurate measurement of discourse skills per se in PWAs.
Also, this study serves to be a preliminary study, which may pave way for the future studies to be conducted in a more extended and elaborated manner. Despite the robust scores for conversation tasks than other discourse genres in PwFA, the findings of the study should be interpreted meticulously owing to the small sample size; and due to usage of a qualitative scale to appraise discourse skills. Furthermore, selecting suitable discourse elicitation tasks specific to PwFA is deemed paramount, when identifying the generalization of impaired-based training to functional skills per se.
Despite the study findings that shed light on differences among different discourse genres. The current study poses limitations concerning sample size. The study design restricted the preliminary assessment merely to identify the type of aphasia and word retrieval impairment. However, it is also important to assess working memory in PwFA, owing to its reciprocal relationship with discourse abilities.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
