A systematic Review: Idiom Comprehension in Aphasia: The Effects of Stimuli and Task Type

Introduction and Aims
Idioms differ in their semantic dimensions of familiarity (frequency of encounter), ambiguity(possibility of a literal interpretation), decomposability (possibility of the idiom’s words to assist in its figurative interpretation) and transparency (possibility to deduce the original metaphorical motivation of an idiomatic phrase). Studying clinical populations such as subjects suffering from aphasia, often caused by Left Hemisphere (LH) brain lesions, can help elucidate idiom processing in the brain (e.g., 1, 2, 3, 5). Patients with aphasia sometimes show less impaired comprehension for specific semantic dimensions, suggesting that some idiom semantic dimensions may not depend solely on the LH. However, recent literature (e.g., 1, 3,4) examines whether the types of idioms and tasks employed affect how different brain areas are involved.This study investigates idiom comprehension in aphasia and explores the potential effect of1. the idiom semantic dimensions, and 2. the tasks employed, on the patients’ idiom comprehension.


Methods
A systematic review was done following the PRISMA approach. Starting from an initial find of n =451, screening and assessment for inclusion and exclusion criteria, 15 articles were retained for further analysis. Extracted information included patient characteristics, types of idiomatic stimuli, types of tasks employed and patients’ performance in idiom comprehension. For the idiomatic stimuli, we extracted information about the combinations of their semantic dimensions, and how these dimensions were defined. For the task employed, we extracted information about the type of the task, the modality of presentation and response, context manipulation as a variable of interest, and whether accuracy and/or reaction times were measured. For the patients’ performance, we extracted information about the errors arising 1. considering different combinations of idiom semantic dimensions, and/or 2. considering the task design.

Results
Results indicated that studies in idiom comprehension in aphasia are characterized by heterogeneity regarding the patient characteristics, the idiomatic stimuli and the experimental task employed. Patient ages ranged between 35-82 years old. Out of a total of15 studies, 4 were case studies including Wernicke’s, Broca’s, Global, Anomic orTranscortical Sensory aphasia. The rest considered all types of aphasia as one group. For the idiomatic stimuli, 10 studies used only familiar idioms, 4 studies did not control ambiguity and 6 studies did not measure transparency. Only 1 study mentioned measurements for both transparency and decomposability. Last, only 2 studies included combinations of idiom semantic dimensions. In addition, 4 studies used the term“transparency” but they measured decomposability.For the experimental task, 4 studies conducted only a string-to-word matching task(where participants were presented with an idiomatic string and were provided with different target words), 4 studies conducted only a string-to-picture matching task (where participants were presented with an idiomatic string and were provided with different pictures) and 1 study used only an oral definition task. 4 studies used a combination of the tasks mentioned. 1 study used a reading task, where they asked participants to simply read the idiomatic string they saw and 1 study employed a lexical decision task, where participants were asked to judge whether a target word is related to the idiomatic string hat they have been previously presented with. For the matching tasks, 7 studies included options (either pictures or words) representing (1) the figurative meaning of the idiom, (2) a literally related meaning of the idiom and (3) an unrelated situation, while 4 studies used other options with either excluding the literal meaning or replacing it with a semantic associate. 2 studies included context manipulation. In terms of modality, most studies provided participants with both an oral and a written presentation of the stimuli.Considering participant response, 6 studies required their response by pointing, 1 study by using the keypad, 1 study orally and 6 studies did not specify this. Last, 13 studies focus on accuracy, only 1 measured both reaction times and accuracy as well as reading times and 1reported only on reaction times.


Discussion
Considering the idiomatic stimuli employed, patients with aphasia show more deteriorated comprehension when presented with unambiguous/opaque and ambiguous/decomposable idioms. In the former case, transparency, i.e., the illustrative motivation of an idiom, may be facilitating comprehension, while in the latter case, decomposability, i.e., the existing link between the idiom’s content words and its figurative meaning, may be hindering comprehension. In addition, idiom types were linked to the involvement of specific brain areas. Current literature can’t unanimously account for what types of idioms involve the left and/or right temporal gyrus, but there are indications that the left Middle Temporal Gyrus(MTG) is predominantly involved in idiom comprehension irrespective of the idioms’ semantic dimensions, while the role of the right MTG is still controversial. Also, patients with frontal lesions show impaired comprehension in ambiguous idioms, but it is uncertain how an intact right temporal lobe would affect (perhaps facilitate) comprehension patterns.Considering the experimental task employed, the string-to-word matching task appears more appropriate for testing patients with aphasia since it does not obfuscate their abilities.However, difficulties arise specifically relating to ambiguous idioms, irrespective of the language. The choice words or pictures for the string-to-word and string-to-picture matching tasks respectively is also crucial. Specifically, more literal errors are observed inmost of the studies, irrespective of the type of task employed. More literal errors compared to unrelated errors are observed for ambiguous and decomposable idioms. Next, context displayed facilitatory outcomes. Patients were more accurate when the context was preceding the idiomatic stimulus rather than when it was following it. Last, modality may be a factor co-influencing the outcomes. Considering the case studies, only the one including the anomic aphasics referred to the error patterns, showing more literal errors with ambiguous/transparent idiomsFinally, this review underlines some limitations. First, the idiom semantic dimensions were insufficiently described and often accompanied by terminological inconsistencies. Second, patient profiles were incomplete, and results were not stratified according to type of aphasia. Future studies should give more information on patient profiles and consider giving information per type of aphasia.