A Scoping Review on the Effects of Emotional Stimuli on Language Processing in People With Aphasia

Abstract

Purpose:

Emotional stimuli have been shown to influence language processing (both language comprehension and production) in people with aphasia (PWA); however, this finding is not universally reported. Effects of emotional stimuli on language performance in PWA could have clinical and theoretical implications, yet the sparsity of studies and variability among them make it difficult to appraise the significance of this effect. The purpose of this scoping review was to (a) determine the extent and range of research examining the effect of emotional stimuli on language processing in PWA, (b) summarize and evaluate research findings, and (c) identify gaps in the literature that may warrant future study.

Method:

PsycINFO, PubMed, Web of Science, and Cumulative Index to Nursing and Allied Health Literature databases were systematically searched for articles that compared performance in response to emotional and nonemotional stimuli on at least one language measure in one or more adults with aphasia. Data related to methods and results were extracted from each article and charted in Excel.

Results:

Five hundred forty unique articles were found, and 18 articles, consisting of 19 studies, met inclusion/exclusion criteria for this review. Of the 19 studies included, 11 studies reported enhanced performance on a language task for emotional compared to nonemotional stimuli, seven reported no difference, and one reported worse performance for emotional compared to nonemotional stimuli. Possible modulating variables such as task type, measurement, stimulus characteristics, and sample characteristics are discussed along with gaps in the literature.

Conclusion:

The extent of research in this area is sparse; however, there does appear to be some early evidence for better performance in response to emotional over nonemotional stimuli in PWA for some, but not all, language processes investigated.

The observation that people with aphasia (PWA), particularly those with severe aphasia, have a more preserved ability to produce emotional content dates back to the English neurologist Hughlings-Jackson (1874, 1879) in his descriptions of more preserved nonpropositional (more automatic) compared to propositional (more intentional) speech in aphasia, particularly in severe aphasia. Nonpropositional speech was thought to include “recurring utterances” and utterances that have affective or emotional value. Since Hughlings-Jackson's observations, there has been a body of literature investigating the effect of emotional stimuli on language processing abilities. Some have shown that PWA demonstrate superior language performance for emotional stimuli compared to nonemotional on a range of tasks (Boller et al., 1979; Borod et al., 2000; Landis et al., 1982; Ramsberger, 1996), whereas others have found no difference in performance between emotional and nonemotional stimuli (Bakhtiyari et al., 2015; Wallace & Canter, 1985). Differences in results may be related to variability in methods used—for example, type of language process investigated, type of language task, type of emotional stimuli (e.g., word-level, pictures, visual, or auditory), and type of emotional manipulation (e.g., prosody, semantics, or context). There also appear to be differences regarding the way researchers conceptualize or account for emotional variables.

It is important to understand whether and how stimulus emotionality could affect language processing because an “emotion effect” could have clinical implications when working with PWA. For example, effects of emotion on language could have implications for the use of emotional stimuli in language assessment when evaluating PWA or other aging populations. If emotional words or pictures are named differently than nonemotional words or pictures, then naming assessments that do not account for this difference may lack reliable internal consistency or, worse, validity. Communication abilities are associated with quality of life and rates of depression among stroke survivors with aphasia (Ashaie et al., 2019; Cruice et al., 2003; Hilari et al., 2003, 2012; Mitchell et al., 2017), thus improving the precision of diagnostic tools by considering emotion is needed.

Effects of emotion on language could also have implications for aphasia treatment. Although aphasia therapy has been shown to be effective (Brady et al., 2016), treatment response is quite variable across patients, with some showing large improvements and some showing minimal (Lazar & Antoniello, 2008). Thus, finding avenues to augment language outcomes in poststroke aphasia is critical. For instance, if emotional stimuli are shown to enhance language processing in PWA, then these materials may be used to support language performance during intervention. Although, to our knowledge, there have been no empirical studies that have assessed the use of emotional stimuli in language intervention, some posit that emotional stimuli have the potential to enhance performance (Cicero et al., 1999). Borod (1992) suggested that emotional stimuli may tap into emotional processing in the intact right hemisphere to support language performance. Targeting neural substrates, such as intact right-hemisphere structures during therapy, has been vetted using other treatment techniques, the most notable of which is melodic intonation therapy (Albert et al., 1973; Schlaug et al., 2008), which aims to take advantage of the right-hemisphere dominance for melody and rhythm during language production to aid in relateralization of language. Similarly, intention therapy (Benjamin et al., 2014; Crosson et al., 2009), developed by Crosson and colleagues, aims to behaviorally target neural substrates in the right hemisphere through complex movement of the left hand during language therapy to enhance treatment outcomes. Results from these studies investigating melodic intonation therapy and intention treatment suggest that, in the chronic stage, behaviorally targeting the right hemisphere during language treatment can improve treatment effectiveness by relateralizing language function to the right hemisphere, at least to some extent. To advance understanding of whether emotional stimuli could elicit this same effect of enhancing treatment outcomes by engaging the right hemisphere, a thorough review of the current literature examining the effect of emotional stimuli on language performance in PWA is necessary.

Emotional Valence and Emotional Arousal

Two dimensions on which emotional stimuli are commonly rated are valence and arousal (Lang et al., 1993; Russell, 1980). Emotional valence is the degree to which a stimulus is negative or positive, and emotional arousal refers to the extent to which a stimulus is calming or arousing. Emotional stimuli can also be defined by the discrete emotion that it is meant to elicit (e.g., happy, sad). There remains disagreement regarding the number of basic emotions that humans exhibit, but Ekman's (1992) six basic emotions are commonly referenced: fear, anger, joy, sadness, disgust, and surprise.

The type of emotional stimulus may also modulate behavioral responses, as neurotypical adults have demonstrated greater attention toward emotional pictures compared to emotional words (Lees et al., 2005; Pool et al., 2016). Similarly, valence and arousal may have different effects on attention, depending on whether the stimulus is a picture or a word (Sutton & Lutz, 2019). In a dot probe task, Sutton and Lutz (2019) found that after viewing negative words of both low and high arousal, participants responded faster to probes that appeared on the same side as the negative words (i.e., a congruency effect) compared to neutral words, suggesting that the negative words captured more attention than neutral words. The authors did not find this congruency effect with positive words of either low or high arousal. Participants also demonstrated this congruency effect with negative pictures of high and low arousal and positive pictures with high (not low) arousal. The authors concluded that (a) valence and arousal interact with stimulus type in their effects on attention and (b) empirical findings from studies that use words as stimuli may not always generalize to pictures.

In summary, there are methodological differences across studies investigating the effects of emotion on language processing in PWA that may be contributing to variability in conclusions. Moreover, to our knowledge, the existing studies examining the effect of emotionality of stimuli on language processing in PWA have not yet been reviewed, making it difficult to quantify and expose these differences across studies.

The objective of the current study was to conduct a scoping review ¹ of the literature that examines the effect of emotional stimuli on language processing in PWA. Following the scoping review framework put forward by Arksey and O'Malley (2005), the purpose of this scoping review was to (a) determine the extent and range of research in this area, (b) summarize research findings, and (c) identify gaps in the literature that warrant future study. We addressed the specific research question: Are there differences in language performance for emotional versus nonemotional stimuli in PWA?

Method

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for scoping reviews (Tricco et al., 2018) reporting guidelines were used in conjunction with reporting recommendations from Arksey and O'Malley (2005), Levac et al. (2010), and Peters et al. (2015, 2020). The PRISMA scoping review checklist is in the Appendix. Four major research databases, including PsycINFO, PubMed, Web of Science, and the Cumulative Index to Nursing and Allied Health Literature, were searched for articles about the effect of emotion on language processing in PWA. The systematic search strategy was designed and carried out by the first author and concluded in early July 2019; however, articles that were identified as meeting inclusion criteria after that time from other sources and nonsystematic searches were included as they were discovered. The search strategy used for PsycINFO is shown in Table 1, which includes the search terms and filters used in all searches.

Table 1.

Search strategy for the PsycINFO database.

Search box	Terms
Search Term 1	valence* OR arous* OR emotion*
AND Search Term 2	language OR produc* OR comprehen* OR lexical OR word OR sentence OR discourse OR repetition
AND Search Term 3	aphasi*
Other criteria	English, age groups (adulthood, 18 years and older)

Inclusion and exclusion criteria for this review were determined in an iterative process through a series of meetings between the two coauthors. Articles were included if

1. the participants included a single case or group of PWA;

2. the participants were adults (18 years or older);

3. the study included an emotional manipulation of stimuli as an independent variable; that is, a numeric or statistical comparison was made between emotional and nonemotional stimuli or among affective categories (e.g., positive, negative, and neutral);

4. the participants completed a language processing task using the aforementioned stimuli, and a linguistic dependent variable was used; and

5. the article was in English (non-English articles were excluded due to feasibility constraints related to time and translation costs).

If PWA were included in a larger group of participants with left-hemisphere damage (i.e., not all participants in the group of participants with left-hemisphere damage had aphasia), these studies were still included, but this was noted in the results. Articles were excluded if

1. the participants had primary progressive aphasia or frontotemporal dementia;

2. the article was a review, comment, or book chapter that did not report new data (though these texts were used to identify additional studies to include);

3. the language processing task was to categorize a linguistic stimulus by emotional category (e.g., participants are asked to listen to a sentence produced with certain emotional prosody and then identify the emotion being portrayed) as we considered this a metalinguistic task rather than a direct measure of linguistic performance; or

4. the dependent variable was an acoustic or perceptual variable related to production or perception of emotional prosody, as this was considered a paralinguistic rather than linguistic measure.

If data reported in a dissertation were subsequently published in a peer-reviewed journal, we extracted data from the peer-reviewed paper rather than the dissertation. Database search results were exported into Zotero, and duplicates were removed. Initially, titles and abstracts were screened by the first author for inclusion and exclusion criteria. Then, full-text articles of abstracts that appeared to meet inclusion/exclusion criteria or had uncertain eligibility were reviewed by the first author. Reference lists of included full-text articles were also reviewed to identify additional studies. Twenty-two articles (21.6% of full-text articles assessed for eligibility) were provided to the second author to obtain reliability on whether to include/exclude them in the review. The second author was blind to the first author's decision prior to judgment. There was 73% agreement between the first and second authors (16 of 22 articles). Consensus was reached by discussion on the remaining six articles.

While reaching consensus, there were discussions on whether a language task or dependent measure was, in fact, linguistic rather than a measure of another cognitive or emotional variable. We decided to exclude studies that used tasks that primarily taxed the memory system rather than the language system but used linguistic stimuli (e.g., word recall task). We also excluded studies that used dependent measures that evaluated the presence of emotional content within a participant's response (e.g., ratings of emotional intensity in a narrative) because this measure would inherently differ for emotional and nonemotional conditions. That is, although emotional intensity would rely somewhat on language use, emotional and nonemotional conditions are expected to be different in their emotional intensity or content regardless of participant performance; thus, the differences observed may not be due to a difference in linguistic performance but rather reflect the inherent differences in condition.

After final inclusion and exclusion decisions were made, data from each included article were charted in Excel. The following information was extracted:

1. article characteristics (i.e., authors, title, year of publication, country of origin, whether it was peer-reviewed or a dissertation),

2. participant sample characteristics (i.e., sample size, number of left-hemisphere stroke participants, number of PWA, age of sample, aphasia type and diagnostic criteria, time poststroke),

3. stimulus characteristics (i.e., stimuli used, stimulus type and modality, type of emotional manipulation, levels of emotional variable, emotional stimulus validation approach, number of items),

4. language task characteristics (i.e., type of task, whether it was a production or comprehension task, language process targeted),

5. the dependent measure(s),

6. main findings, and

7. noteworthy study limitations.

Results are organized first by methodological features of the studies and then by the language task investigated. This organization was chosen to identify differences and gaps in methodology across studies (aligned with the third purpose of this scoping review), examine the extent of language processes investigated (first purpose), and summarize/compare findings of studies that employed similar task methodology (second purpose).

Results

There were a total of 540 unique search results combined across the four databases. Of these, 13 articles were identified as meeting inclusion/exclusion criteria. An additional five articles were found to have met inclusion/exclusion criteria from other sources (e.g., references, other searches), resulting in a total of 18 articles. See the PRISMA diagram (see Figure 1) for a flow chart of the article selection process. Sixteen of the included articles were publications from peer-reviewed journals, and two were dissertations (Alasseri, 2007; Pick, 2002). One article (Landis et al., 1982) included two different studies examining the effects of emotion on language—one on reading and one on writing—so these studies were separated in the analysis, resulting in 19 different studies across the 18 articles. Publication year ranged from 1975 to 2022. Fourteen studies were conducted by groups in the United States, and the other five studies were conducted by groups in Romania, Greece, Canada, Sweden, and Iran.

Figure 1.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagram (Moher et al., 2009).

Participants

Two-hundred sixty-five patients with left-hemisphere stroke were included across the 19 studies in the review. One article (Kreindler et al., 1980) did not report the hemisphere of brain damage of their included participants, but they were considered to have aphasia, so these 20 patients were included in the overall number of patients with left-hemisphere brain damage. Of these 265 patients, 247 were considered to have aphasia based on varied diagnostic criteria. There were three articles by Bloom et al. (1992, 1993, 1996) that appeared to use the same set of participants across these three studies; thus, there were an estimated 241 unique participants with left-hemisphere brain damage and 223 unique patients with aphasia included across all studies. The number of participants with left-hemisphere damage included in an individual study ranged from four to 41. The average age of participants ranged from 42.0 to 64.4 years. Overall, female participants were underrepresented in the included studies. Studies also varied in the type, severity, and months postonset of aphasia of included participants. Almost all studies controlled for handedness by only including right-handed participants. Table 2 summarizes participant demographics across all studies ordered by publication year.

Table 2.

Participant demographics.

Study	N	Age in years	Gender	Aphasia type	Aphasia severity	Diagnostic test(s)	MPO	Etiology	H
Heilman et al. (1975)	6	54.5 (37–64)	3F, 3M	5 conduction, 1 anomia	NR	NR	NR	3 CVA, 2 glioma, 1 hematoma	R
Boller et al. (1979)	8	58.9 (38–77)	3F, 5M	2 Wernicke, 6 global	7 severe, 1 moderate	10-item auditory comprehension screening test, needed 3 or fewer correct to qualify	NR	CVA	NR
Kreindler et al. (1980)	20	42 (16–66)	NR	8 Broca, 12 Wernicke or 10 nonfluent, 9 fluent, 1 unclassifiable	NR	“Standard examination of expressive and receptive speech”	a) 1–6 days post b) 1–3	8 closed trauma, 12 open trauma	NR
Landis et al. (1982) Study 1	22	NR	0F, 22M	14 nonfluent, 8 fluent	Mild–severe (numbers NR)	BDAE	NR	NR	19 R, 3 L a
Landis et al. (1982) Study 2	18 b	NR	0F, 18M	12 nonfluent, 6 fluent	Mild–severe (numbers NR)	BDAE	NR	NR	17 R, 1 L
Roeltgen et al. (1983)	3	49 (30–62)	1F, 2M	1 Broca, 1 Wernicke/ conduction, 1 global	NR	WAB + clinical observation	2 acute, 1 chronicc	2 CVA, 1 demyelination	R
Wallace & Canter (1985)	7	59.9 (52–74)	NR	1 Broca, 2 Wernicke, 1 conduction, 1 anomic, 1 mixed anterior, 1 global	NR	BDAE + clinical observation + neurological data	1 week–108	6 CVA, 1 CVA with bilateral damage	R
Reuterskiöld (1991)	12	64.4 (44–77)	0F, 12M	3 Wernicke, 1 transcortical mixed, 8 global	NR	BDAE	1–50	NR	R
Bloom et al. (1992)	12	63.3 (NR)	3F, 9M	NR	Mild–moderate	BDAE + MTDDA	At least 6, M = 44.7	CVA	R
Bloom et al. (1993) d	12	63.3 (NR)	3F, 9M	NR	Mild–moderate	BDAE + MTDDA	At least 6	CVA	R
Bloom et al. (1996) d	12	63.3 (NR)	3F, 9M	5 nonfluent, 7 fluent	Mild–moderate	BDAE + MTDDA	At least 6, M = 44.7	CVA	R
Ramsberger (1996)	20	61 (50–81)	0F, 20M	4 Broca, 2 Wernicke, 2 conduction, 2 global	BDAE severity ratings ranged from 0.0 to 4.0 (M = 1.20)	BDAE	3–106	CVA	R
Borod et al. (2000)	13 e	62.2 (39–78)	5F, 8M	5 nonfluent, 2 fluent, 6 mixed	3 mild, 1 mild–moderate, 2 moderate, 1 mod–severe (only 7 with severity reported)	BDAE	At least 1, M = 18.3	CVA	R
Pick (2002)	7 e	63.7 (48–79)	2F, 5M	NR	Mild	BNT + BDAE	2–49	CVA	R
Alasseri (2007)	9	50.9 (38–62)	5F, 4M	5 Broca, 1 fluent, 3 anomic	1 mild, 7 moderate	JADIT	4–48	CVA	R
Bakhtiyari et al. (2015)	15	57.9 (45–65)	4F, 11M	NR	NR; excluded for severe auditory comprehension deficits	PAT + informal assessment	At least 6	CVA	R
Efthymiopoulou et al. (2017)	41	60.4 (24–86)	12F, 29M	NR	NR	BDAE short form adapted to Greek	0–69.3	CVA	R
Newton et al. (2020)	20	61.4 (36–83)	10F, 10M	1 Broca, 1 transcortical, f 3 conduction, 15 anomic	WAB-R AQ ranged from 46 to 96.8 g	WAB-R + PALPA	15–201	CVA	R
Harmon et al. (2022)	8	51.8 (34–76)	3F, 5M	4 Broca, 1 Wernicke, 3 anomic	3 mild, 5 moderate: WAB-R AQ ranged from 60.3 to 85.9	BNT + WAB-R	13–196	CVA	NR

Note. N represents the number of people with aphasia (PWA) included in the data analysis for each study, not the number of total participants in the study. Demographic variables are reported for only the aphasia group if reported. Em dashes indicate data not reported. Age = mean (range); MPO = months postonset of stroke/brain injury; H = handedness; F = female; M = male; NR = not reported; CVA = cerebrovascular accident; R = right; L = left; BDAE = Boston Diagnostic Aphasia Evaluation; MTDDA = Minnesota Test for the Differential Diagnosis of Aphasia; BNT = Boston Naming Test; JADIT = JISH Aphasia Diagnostic Test; PAT = Persian Aphasia Test; WAB-R AQ = Western Aphasia Battery–Revised Aphasia Quotient; PALPA = Psycholinguistic Assessments of Language Processing in Aphasia.

^a One of the left-handed PWA had damage to only the right hemisphere. All other PWA had damage to the left hemisphere only.

^b These participants were from the same sample of participants in Study 1. In both Studies 1 and 2, participants who reached floor or ceiling on the tasks were excluded, explaining the unequal numbers included in the analysis.

^c One participant was tested 2–4 days postonset, one participant was tested 1 week postonset, and the third was tested at either 6 or 12 MPO (authors did not report which tests occurred at which time point).

^d Appears to be the same sample reported in Bloom et al. (1992).

^e These PWA were grouped within a larger group of people with left-hemisphere damage (LHD). Borod et al. (2000): out of 15 with LHD. Pick (2002): out of 13 with LHD.

^f The authors do not report whether this participant had transcortical motor, transcortical sensory, or transcortical mixed aphasia.

^g The authors state that although the WAB-R AQ cutoff for a diagnosis of aphasia is 93.8, the three participants that scored above this criterion had difficulty “producing narrative and when abstract or less frequent words were required.”

Language Tasks

Of the 19 studies included in the review, 14 examined the effect of emotion on language production tasks, and five examined the effect of emotion on language comprehension tasks. Discourse production was the most common language process examined (eight of 19), ² followed by sentence-level auditory comprehension (three of 19), word writing (two of 19), word repetition (two of 19), word retrieval (one of 19), oral word reading (one of 19), word recognition (one of 19), and word-level auditory comprehension (one of 19). See Table 3 for a summary of the language processes, type and complexity of tasks, and stimulus type included in the scoping review.

Table 3.

Language processes and tasks included in the scoping review.

Language process/task	Complexity level	Stimulus type (mode)	No. of studies
Production			14
Writing		Word (auditory)	2
Repetition	Word	Word (auditory)	2
Reading aloud		Word (orthographic)	1
Speaking	Word	Pictures (visual)	1
		Prompt (auditory)	1
		Pictures (visual)	3
	Discourse	Prompt (auditory) + word (orthographic)	2
		Video	1
		Prompt (auditory) & picture (visual)	1
Comprehension			5
Word recognition	Word	Word (orthographic)	1
Auditory comprehension		Word (auditory)	1
Auditory comprehension	Sentence	Sentence (auditory)*	3

Note. “+” indicates that both stimulus types were used for each item. “&” indicates that different stimulus types were used for different items.

^*All stimulus items across studies were manipulated for emotion based on their semantic meaning except two of these studies, which manipulated their sentences by emotional prosody only.

Stimuli

There was a wide range of emotional stimulus types used in the studies (see Table 3). Most studies used linguistic stimuli (13 of 19), and fewer used pictures or video (six of 19). Seventeen studies used stimuli that were semantically or contextually emotional, and two used linguistic stimuli that were prosodically emotional but semantically neutral. A variety of methods were used to define, categorize, and identify emotional stimuli. Most studies divided stimuli into a set of emotional and nonemotional stimuli. Others categorized stimulus sets by emotional valence (positive, negative, and neutral) or by emotional category (e.g., angry, neutral, happy, and sad). Eleven of the 19 studies reported their method for selecting the included emotional stimuli, which most commonly consisted of having a separate group of raters rate a larger set of stimuli based on emotional variables, which experimenters then used to narrow down and identify their emotional stimulus set (Alasseri, 2007; Bakhtiyari et al., 2015; Boller et al., 1979; Borod et al., 2000; Landis et al., 1982; Newton et al., 2020; Ramsberger, 1996; Reuterskiöld, 1991; Wallace & Canter, 1985). For example, Reuterskiöld (1991) asked a group of five neurotypical men to judge 40 words on a 7-point scale, ranging from “neutral” to “highly emotional,” and the author then chose the 14 words that were rated as most emotional and 14 words that were rated as least emotional for their final stimuli. Four other studies lacked detail in their descriptions of how they designed their emotional stimuli (Bloom et al., 1992, 1993, 1996; Roeltgen et al., 1983). For instance, Bloom et al. (1992, 1993, 1996) reported that they had designed three sets of three-picture sequences in which one set was designed to elicit seven emotional elements that were reliably produced by neurologically intact controls, whereas the other two sets were designed to include visuospatial and procedural elements, respectively. However, they do not specify the method they used to select or design these pictures or how they arrived at the final elements of interest. Four studies reported no method for arriving at their emotional stimulus set (Efthymiopoulou et al., 2017; Heilman et al., 1975; Kreindler et al., 1980; Pick, 2002). Another source of variability related to stimulus sets was the number of items within each condition. Studies ranged from including one item per condition (e.g., one emotional stimulus and one nonemotional stimulus) to including 60 items per condition. Table 4 summarizes stimulus and task variables for each study along with central study findings.

Table 4.

Methods and results.

Study	Stimulus type and mode	Levels of emotion	No. of items per condition	Language process (task)	Dependent measure	Performance better, worse, or no different for E compared to NE?
Language production studies
Ramsberger (1996)	Words, auditory	1. E (abstract) 2. NE (abstract & concrete)	22 (10 of the NE words were abstract, and 12 were concrete)	Rep (word repetition)	Accuracy	Better when comparing E abstract to NE abstract
Bakhtiyari et al. (2015)	Words, auditory	1. E 2. NE	20	Rep (word repetition)	Accuracy	No difference
Landis et al. (1982) Study 1	Words, orthographic	1. E abstract 2. NE abstract 3. NE concrete	12	R (oral word reading)	Score of 0–2 per item: accurate (2), semantic substitution (1), incorrect (0)	Better
Landis et al. (1982) Study 2	Words, auditory	1. E abstract 2. NE abstract 3. NE concrete	12 a	W (word writing to dictation)	Score of 0–2 per item: accurate (2), semantic substitution (1), incorrect (0)	Better
Roeltgen et al. (1983)	Words, auditory	1. E 2. Neutral	Patient 1: 145 Patient 2: 25 Patient 4: 40	W (word writing to dictation)	Accuracy	2/3 better, 1/3 no difference
Harmon et al. (2022)	Pictures, visual	1. Positive 2. Negative 3. Neutral	20 positive, 20 negative, 60 neutral	WR (object picture naming)	1. Accuracy 2. Reaction time	Worse for negative, no difference for positive and neutral (for both accuracy and reaction time)
Kreindler et al. (1980)	Prompt, auditory	1. Emotional description 2. Free description	1	SDP (narrative production)	1. Speaking duration 2. Number of utterances 3. Speaking rate	No difference
Bloom et al. (1992)	Pictures, visual	1. E 2. Visuospatial 3. Procedural/ neutral	1	SDP (picture set story description)	Score of 0–7 per item: 1 point for each content element included in the story	No difference
Bloom et al. (1993) b	Pictures, visual	1. E 2. Visuospatial 3. Procedural/neutral	1	SDP (picture set story description)	Score of 0–7 per item: 1 point for each of 7 pragmatic features	Better
Bloom et al. (1996) c	Pictures, visual	1. E 2. Visuospatial 3. Procedural/neutral	1	SDP (picture set story description)	1. Score of 0–2 on each of 7 coherence features for each item 2. Cohesion score	Better on majority of coherence variables, but not cohesion variables
Borod et al. (2000)	Prompt, auditory + emotion word, orthographic	1. E (positive & negative) 2. NE (positive & negative) d	3	SDP (narrative production)	Score of 1–5 on each of 6 pragmatic features for each item	Better for positive valence
Pick (2002)	Prompt, auditory + emotion word, orthographic/auditory	1. E 2. NE	8	SDP (narrative production)	Score of 0–6 on accuracy (how well the content reflected prompt/cue)	No difference
Alasseri (2007)	Video, visual (no audio)	1. Positive 2. Negative 3. Neutral	3	SDP (video description)	Scores on 6 different measures of pragmatic discourse performance	Better on either positive or negative compared to neutral on 5/6 pragmatic variables
Efthymiopoulou et al. (2017)	1. Prompt, auditory 2. Picture, visual	1. Narrative about stroke (emotional content) 2. Picture description (no emotional content)	1	SDP (narrative production + picture description)	Speech rate	No difference
Language comprehension studies
Heilman et al. (1975)	Sentences, auditory	1. Angry 2. Happy 3. Sad 4. Indifferent	4 (same 4 sentences repeated in each prosodic condition)	AC (sentence-to-picture matching)	Accuracy	No difference
Boller et al. (1979)	Sentences, auditory (live and recorded)	1. High E 2. Low E 3. Neutral	10 high E, 20 low E, 30 neutral (all presented twice in each auditory condition)	AC (answering questions and following commands)	Score of 0–3 per item based on accuracy, appropriateness, and behavior change	Better scores for high emotional sentences
Wallace & Canter (1985)	Sentences, auditory	1. Angry 2. Neutral 3. Melodically intoned	21 (same 21 sentences in each prosodic condition)	AC (sentence-to-picture matching)	Accuracy	No difference
Reuterskiöld (1991)	Words, auditory	1. E 2. NE	14	AC (word-to-picture matching)	Score of 0–2 per item: accurate within 5 s (2), accurate within 30 s (1), inaccurate	Better
Newton et al. (2020)	Words, orthographic	1. Positive 2. Negative 3. Neutral	38	WRec (lexical decision)	1. Accuracy 2. Reaction time	Better accuracy for positive valence; faster for positive and negative valence

Note. Only tasks and dependent measures relevant to the research question of interest of this scoping review were included. That is, some studies may have included additional tasks or measures that did not meet inclusion criteria, so they were not listed in this chart or included as part of the review. For stimulus type and mode, “+” indicates that both stimulus types were used for each item, and “&” indicates that different stimulus types were used for different items. E = emotional; NE = nonemotional. For language processes: Rep = repetition; R = reading; W = writing; WR = word retrieval; SDP = spoken discourse production; AC = auditory comprehension; WRec = word recognition.

^a These were the same 12 words used in Landis et al. (1982) Study 1.

^b The stimuli and task used in these two studies are the same as those used in Bloom et al. (1992).

^c The authors did not complete statistical tests to compare within-group differences across conditions for people with aphasia, so only a mean comparison was reported.

^d The authors included a set of positive and negative stimuli in both the E and NE conditions. They report that although the NE trials had some emotional connotation (positive and negative), the relative intensity of these trials “should be much less than for the emotional categories.”

Dependent Measures and Main Findings

A variety of dependent measures were used across studies, the most common of which was accuracy (seven studies). Other dependent measures included reaction time (Harmon et al., 2022; Newton et al., 2020), a graded accuracy scale based on time (i.e., 0 points for incorrect, 1 point for correct in 30 s, and 2 points for correct in 5 s; Reuterskiöld, 1991), a graded scale based on accuracy and quality (Boller et al., 1979; Landis et al., 1982), speaking duration (Kreindler et al., 1980), number of utterances (Kreindler et al., 1980), speaking rate (Efthymiopoulou et al., 2017; Kreindler et al., 1980), and scores based on the inclusion of specific elements or qualitative features identified by the authors (Alasseri, 2007; Bloom et al., 1992, 1993, 1996; Borod et al., 2000; Pick, 2002).

There was evidence that linguistic performance in response to emotional stimuli was better than to nonemotional stimuli in eight studies (42.1%), partial evidence for this effect in three studies (15.8%), seven articles that found no difference between the set of emotional stimuli and nonemotional stimuli (36.8%), and one that found worse performance for emotional compared to nonemotional stimuli (5.3%).

Language Production Studies

Of the 14 studies that examined the effect of emotionality on language production, five found evidence of a performance-enhancing effect of emotion (Bloom et al., 1993; Borod et al., 2000; Landis et al., 1982; Ramsberger, 1996); three found evidence of a performance-enhancing effect of emotion on several, but not all, dependent measures (Alasseri, 2007; Bloom et al., 1996) or for certain individuals when there was no group-level analysis completed (Roeltgen et al., 1983); and five studies found no difference in performance between emotional and nonemotional stimulus sets (Bakhtiyari et al., 2015; Bloom et al., 1992; Efthymiopoulou et al., 2017; Kreindler et al., 1980; Pick, 2002). One study found an interference effect of emotion on language production (Harmon et al., 2022). The following is a summary of results by language task.

Word repetition. Two studies examined the effect of emotional stimuli on word repetition (Bakhtiyari et al., 2015; Ramsberger, 1996). Ramsberger (1996) found that participants repeated emotional abstract words significantly better than nonemotional abstract words; however, Bakhtiyari et al. (2015) found no difference in repetition performance for emotional versus nonemotional words. It is worth noting that Ramsberger's study was conducted in English whereas Bakhtiyari et al.'s study was conducted in Persian. Also, Ramsberger controlled for concreteness in her stimulus set and found the facilitation effect of emotion only when comparing repetition performance on abstract emotional words with performance on abstract nonemotional words. However, there was no difference in repetition performance when comparing abstract emotional words with concrete and abstract nonemotional words combined. Bakhtiyari et al. did not control for concreteness and compared emotional abstract words with both concrete and abstract nonemotional words combined—the same comparison in which Ramsberger also did not find a performance-enhancing effect of emotion. These limited findings suggest that concreteness may interact with an emotion effect on repetition, and/or there may be some cultural–linguistic differences in the effects of emotion on language.

Reading aloud. One study examined the effect of emotional stimuli on reading aloud at the word level (Landis et al., 1982). These authors found significantly more accurate performance on reading emotional compared to nonemotional words. In contrast to Ramsberger (1996), Landis et al. (1982) found better performance for abstract emotional words compared to both nonemotional abstract and concrete words. Therefore, the emotion facilitation effect on oral word reading may not be as sensitive to concreteness as repetition. However, given that there was only one study examining this effect on oral word reading, results should be interpreted cautiously.

Writing to dictation. Landis et al. (1982) found an emotion facilitation effect for word writing to dictation. PWA wrote abstract emotional words more accurately than nonemotional abstract and concrete words. Roeltgen et al. (1983) compared word writing to dictation performance on emotional nouns and adjectives compared to nonemotional nouns and adjectives in three participants. Two of the three participants with aphasia made more errors writing neutral words to dictation than emotional words, though the differences were only statistically significant for one participant (p < .05 and p = .055). The third participant demonstrated no difference in writing emotional compared to nonemotional words. In summary, there is preliminary evidence that emotional words may improve writing to dictation performance in PWA with some interparticipant variability.

Word retrieval. The one study that examined effects of emotion on word retrieval found an interference effect for word retrieval for negative items (Harmon et al., 2022). PWA named images depicting emotional nouns and named negative words less accurately and slower than positive and neutral words. Their performance did not significantly differ for positive and neutral items.

Discourse production. The remaining eight studies that examined language production investigated the effects of emotional stimuli on discourse. Four of these studies concluded that emotional stimuli resulted in better performance than nonemotional stimuli on all or most of their dependent measures (Alasseri, 2007; Bloom et al., 1993, 1996; Borod et al., 2000), and four of eight studies found no difference in discourse production between emotional and nonemotional stimuli (Bloom et al., 1992; Efthymiopoulou et al., 2017; Kreindler et al., 1980; Pick, 2002). Of the four studies that found better performance for emotional than nonemotional stimuli, three found benefits of emotion on pragmatic performance (Alasseri, 2007; Bloom et al., 1993; Borod et al., 2000). Borod et al. (2000) and Alasseri (2007) found that advantages for emotional stimuli varied as a function of emotional valence, whereas Bloom et al. (1993) did not examine valence effects. Alasseri (2007) showed that, for certain pragmatic measures, negative stimuli resulted in better pragmatic performance than positive and neutral stimuli (number of words and quality); for other measures, both negative and positive stimuli resulted in better pragmatic performance than neutral stimuli (discourse efficiency and coherence); and positive stimuli resulted in better pragmatic performance than neutral and negative stimuli for lexical selection. Borod et al. (2000) found that positively valenced narratives were rated better for pragmatic performance than neutral and negative narratives. Finally, Bloom et al. (1996) found higher means for emotional stimuli compared to nonemotional stimuli on six of seven coherence variables and two of six cohesion variables; however, without a statistical comparison, we are unable to evaluate whether these differences were statistically significant. In summary, emotional stimuli may result in better pragmatic performance, discourse coherence, and possibly cohesion compared to nonemotional stimuli.

Of the four studies that found no differences between emotional and nonemotional stimuli on discourse, each found no difference in content elements recalled (Bloom et al., 1992); number of utterances, duration, and rate (Kreindler et al., 1980); accuracy of conveying the desired emotional category (Pick, 2002); and speech rate between conditions (Efthymiopoulou et al., 2017). Kreindler et al. (1980) note, however, that, at the individual level, some PWA did produce more utterances and speak longer for emotional stimuli compared to nonemotional, whereas others were more fluent on nonemotional stimuli compared to emotional; therefore, it may be that emotion effects are individual specific. Overall, the findings are split as to whether emotional stimuli are facilitative for discourse production. However, taken together, the results suggest that it likely depends on the discourse measure (e.g., pragmatics, coherence, cohesion, productivity, and quality). For instance, of the three studies investigating the effects of emotion on pragmatics, all three found evidence of performance-enhancing effects of emotion, whereas the other studies found limited or no effects of emotion on cohesion, speech rate, and linguistic productivity. However, discourse elicitation techniques (e.g., picture description, video description, and narrative) and performance measures used across these studies varied, complicating comparison.

Language Comprehension Studies

Of the five studies examining the effects of emotional stimuli on language comprehension, three found evidence of better performance for emotional versus nonemotional stimuli (Boller et al., 1979; Newton et al., 2020; Reuterskiöld, 1991), and two studies found no differences between emotional and nonemotional stimuli (Heilman et al., 1975; Wallace & Canter, 1985). Interestingly, the two studies that found no differences were the only two studies that manipulated their linguistic stimuli by emotional prosody rather than emotional meaning (sentences were semantically neutral). Therefore, it may be that emotional stimuli result in better auditory comprehension only when the stimuli are semantically emotional rather than only prosodically emotional.

At a more basic level, Newton et al. (2020) found that positive words resulted in more accurate lexical decisions for PWA compared to neutral words and that both positive and negative words resulted in faster lexical decision times than neutral words. They were also the only study included in this review that statistically examined the impact of severity on the emotion effects. They found that accuracy for recognition of positive words was slightly more preserved in people with more severe aphasia compared to negative and neutral words. Aphasia severity was not a significant covariate in their reaction time model, however.

Discussion

To address the extent and range of research in this area, our findings suggest that the extent of this research is quite limited, but a range of language processes have been investigated. There is a dearth of replication studies, making it difficult to determine the degree to which these effects or lack thereof were sample specific. The second purpose of this review was to summarize research. Based on the literature included in this review, there does appear to be some preliminary evidence for better performance for emotional stimuli over nonemotional stimuli in PWA for repetition of abstract words (at least in English), reading words aloud, writing words to dictation, pragmatic performance in discourse production, auditory comprehension when stimuli are semantically emotional (rather than just prosodically emotional), and word recognition. In contrast, there may be an interference effect of emotion on word retrieval. Findings appear to be mixed regarding this performance-enhancing effect of emotion on other discourse measures of coherence, cohesion, productivity, and quality. However, given that there are so few articles investigating each language process, in some cases only one, each with its limitations, these findings should be interpreted cautiously. To address our review's third purpose, research gaps and future directions are discussed below.

Limitations and Gaps in the Current Evidence

Though quality appraisal of the evidence is not traditionally a part of scoping reviews, we felt that some limitations in the literature were noteworthy and may influence interpretation of the results. First, many studies lacked clear procedures for validating their stimuli to ensure that emotional stimuli actually contained emotional qualities perceived by their participants and nonemotional stimuli did not. Though the slight majority of studies reported sufficient detail explaining their method of choosing stimuli (11 of 19), these methods typically lacked rigor. Additionally, only one of the included studies evaluated the participants' perceptions (Harmon et al., 2022), leaving it uncertain as to whether the participants interpreted the stimuli as intended in the rest of the studies. Future studies should make sure to rigorously validate their stimuli and report their methods for doing so. Second, several studies included a limited number of items per condition. Five studies only included one item for the emotional condition and one item for the nonemotional condition, and all five of these studies were discourse production studies (Bloom et al., 1992, 1993, 1996; Efthymiopoulou et al., 2017; Kreindler et al., 1980). Although discourse analysis can be time-consuming, conclusions are limited when comparing one stimulus item with another, particularly if the stimulus items have not been validated. Therefore, these studies' results, especially on their own, should be interpreted cautiously.

Similarly, to examine differences in auditory comprehension, Heilman et al. (1975) used four sentences presented in each of four emotional prosodic conditions; however, they used the same four sentences for each condition, introducing a possible confound of repeated exposure. Wallace and Canter (1985) also used the same 21 sentences for each of three prosodic conditions to examine auditory comprehension differences. Given that, with repeated exposure, performance can be impacted (e.g., repetition effects; Salasoo et al., 1985) and emotional reactions can be dampened through habituation (Bradley et al., 1993; Thompson, 2009), this practice could impact results. Though repeating stimuli can control for linguistic variables, such as word frequency and concreteness, this benefit should be weighed against the potential costs of repeated exposure. Finally, several studies did not report important participant demographic information, though the more recent studies did a better job of this. Overall, the limitations of individual studies further complicate the ability to draw sound conclusions when summarizing findings.

Despite the breadth of language processes that were examined among the included studies, the emotion facilitation effect on discourse comprehension has not been investigated. Also, research studies varied widely in how they defined emotion as a property of the stimulus (e.g., by emotional valence or category). Some researchers examined differences across emotional valence categories (positivity vs. negativity), whereas others combined these valence categories and only examined emotional versus nonemotional stimuli. Emotional valence appears to be an emotional dimension that is important to consider as some studies found differences in the emotion effect based on whether the stimulus was positive or negative (Alasseri, 2007; Borod et al., 2000; Harmon et al., 2022; Newton et al., 2020).

Studies included in this scoping review were also inconsistent in emotional arousal of stimuli (excitability vs. calmness), with some including highly arousing stimuli, others including a range of arousal levels, and others simply not reporting any information on stimulus arousal. There is some limited evidence that degree of emotional arousal is a salient factor in language processing, specifically auditory comprehension. Boller et al. (1979) found that auditory comprehension for highly emotional sentences was better compared to neutral sentences; however, when combining sentences that were rated high for emotion (likely higher in arousal, though authors did not use this term) with those that were rated low for emotion (likely lower in arousal), no significant difference in auditory comprehension between the combined emotional stimuli and neutral stimuli was found. There has been some debate regarding the level to which emotional arousal and valence independently or interactively affect language processing in neurotypical participants. However, both emotional dimensions appear to be important considerations when examining the effect of emotional stimuli on language and other cognitive processes.

In summary, to answer the question of whether the emotionality of stimuli impacts the language processing of PWA, we must be clear about what we mean by “emotionality of stimuli,” so there can be a level of consistency across studies that will allow for an adequate comparison. Dimensional models of emotion appear to be the most commonly used in studies examining the effects of emotion on cognitive processes; however, there are several other emotion models. We suggest future researchers consciously adopt a theoretical model on which to base the manipulation of their stimuli for emotion before validating their stimuli and then clearly report their approaches.

Additionally, among the studies included in this review, only a limited subset of participant and stimulus factors that have been shown to modulate emotion effects have been studied or controlled. For example, emotion perception appears to change as we age (Addis et al., 2010; Carstensen et al., 2003), and there is some evidence that these differences in emotion perception may impact language processing in neurotypical adults (Harmon et al., 2022; Schwen Blackett et al., 2017). Thus, age should be considered in the design and analysis of these studies. There also appear to be differences in emotion perception by gender in neurotypical adults (see special issue on gender differences in emotion introduced by Hess, 2015; Kring & Gordon, 1998), which may also influence language processing (Glenberg et al., 2009). Some studies attempted to control for gender/sex ³ differences by restricting their sample to male participants. This solution is not adequate as it leaves out at least half of the general population. In this sample of studies, female participants are underrepresented, and as is often the case, gender and sexual minorities are absent; thus, efforts should be made to recruit from these populations to examine differences in the effects of emotion on language. Furthermore, stimulus factors, such as word frequency, concreteness, and visual complexity of images, were not consistently controlled for or investigated. These factors are likely also important to consider when determining whether there is a performance-enhancing effect of emotion among PWA, as we have seen evidence that concreteness may modulate effects of emotional stimuli on word repetition (Ramsberger, 1996).

Future Directions

Although not the focus of the current study, individual emotional factors, such as mood, stress, and emotional reactivity, may play a role in how emotions are perceived and affect language performance in PWA. Indeed, some of our findings suggest that there are likely individual differences in the effects of emotion on language (Kreindler et al., 1980; Roeltgen et al., 1983). For instance, depression may lead to poorer rehabilitation outcomes poststroke (Aron et al., 2015), and people with depression have also shown differences in emotional perception of faces (Bourke et al., 2010; Dalili et al., 2015) and vocal prosody (Kan et al., 2004; Young et al., 2017). Since depression is not uncommon among PWA (Ashaie et al., 2019; Åström et al., 1993; Kauhanen et al., 2000; Mitchell et al., 2017), examining how it and other mood disorders affect the perception of emotional stimuli and its effect on language processing could be an avenue for future research. Acute stress and chronic stress elicited by linguistic contexts in PWA may also interact with the perception of emotional stimuli and modulate the effects of emotion on language performance (Cahana-Amitay et al., 2011; Laures-Gore et al., 2007). That is, the stress that some PWA experience, when forced to face their language impairment in a communicative context, may negatively affect performance (Torres-Prioris et al., 2019), rendering any language-enhancing effects of emotion nonexistent. In addition, emotional reactivity, which has been characterized by the activation, duration, and intensity of positive and negative affect (Davidson, 1998; Nock et al., 2008), to our knowledge, has not been studied in PWA. Given that brain damage and having aphasia have been shown to cause changes in emotional functioning (Åström et al., 1993; Campbell Burton et al., 2013; Code & Herrmann, 2003; Goldstein, 1952; Hunting Pompon et al., 2019; Kauhanen et al., 2000), it may be worth investigating whether PWA tend to have changes in emotional reactivity also. Based on the limitations and gaps of the extant literature and our suggestions for future research, a list of recommendations for researchers pursuing this line of research is presented in Table 5.

Table 5.

List of recommendations for researchers examining effects of emotionality of stimuli on the language processing of people with aphasia.

Number	Recommendation	Explanation
1	Define and validate emotional value of stimuli	Researchers should explicitly define emotion and identify a theoretical model on which to base their manipulation of emotional stimuli (e.g., dimensional model). Researchers should report how they validated emotional and nonemotional materials (if not already validated) and obtain data from study participants on their perceptions of stimuli to ensure they were perceived as intended.
2	Consider dimensions of valence and arousal in manipulation of emotion	Researchers should avoid using broad categories of emotional versus nonemotional stimuli given that emotional valence and arousal have been shown to modulate effects of emotion on language processing.
3	Control for other task/stimulus variables so conditions only differ by emotional value	Researchers should make sure that emotional and nonemotional conditions are matched for other task (e.g., elicitation technique) and stimulus (e.g., stimulus type, word frequency, concreteness, and visual complexity) variables that are known to affect language processing to eliminate confounding variables.
4	Limit repeated exposure of a stimulus	Given that emotional reactions and general response to stimuli can change with repeated exposure, researchers should limit the number of times a stimulus is presented.
5	Include multiple items/trials per condition	Researchers should use more than one item/trial per condition to improve power, generalizability of results, and validity of conclusions. The ideal number of items/trials per condition will vary depending on other variables like the type of task, number of participants, etc. Researchers should also strive to include approximately equal items/trials in each condition.
6	Consider other participant variables that may interact with effects of emotion	Researchers should control for, consider in their design, or at least report data on other variables that could affect participant perceptions of emotional stimuli, such as age, gender/sex, neurological variables, mood, stress, and emotional affect/reactivity.
7	Report characteristics of participants with aphasia	Researchers should report method of diagnosis of aphasia, aphasia type/severity, months poststroke, etiology of aphasia, and premorbid handedness, as these variables could affect results.

Limitations of the Current Review

Through this systematic scoping review, we revealed the state of current evidence relating to the effects of emotional stimuli on language processing in PWA and identified important gaps in the literature; however, our review has some limitations. First, several studies included in this review were chosen based on their methodology rather than the authors' specific research question, which may explain some limitations. For example, many of these studies focused on differences in hemispheric dominance for emotional processing in relation to language, so participants were recruited based on left-hemisphere damage rather than their aphasia profile, which would explain why their language profiles were not measured or clearly reported. Studies were chosen in this review based on their methodology because we found that our main research question—how emotional stimuli affect language processing in PWA—could be undertaken with the data these studies were reporting. Nonetheless, this approach may have exposed more limitations in the literature than would have been otherwise had the search only included articles that asked our same research question. In addition, although we consulted a librarian early on in our project when determining which type of review to conduct, we did not use a librarian when developing our search strategy. Though we carefully chose our search terms, using a librarian's help to develop the search strategy may have resulted in identifying more relevant literature; however, identifying additional articles through reference lists of included and some excluded articles likely helped in finding articles that were not captured by our search terms. Future groups that have the resources to dedicate to translation should also include non-English studies, as we likely had to exclude relevant data due to this exclusion criterion. Finally, although both authors reviewed 20% of the articles in the full-text review stage of the article selection process, only the first author reviewed all articles. Future work should include multiple authors at all stages of the selection process to further ensure valid and reliable results.

Conclusions

Results suggest that the extent of research on the effects of emotion on language processing in PWA is sparse as there were only 19 articles that met inclusion/exclusion criteria, and findings appear to vary based on the language process that is being investigated, among other variables. The wide range in research methodology and quality makes it difficult to compare articles and draw conclusions about the effect of emotional stimuli on language processing in PWA. Based on the existing literature, initial data show that, in comparison to nonemotional stimuli, emotional stimuli may result in better linguistic performance by PWA for some language processes and possibly worse for word retrieval (Harmon et al., 2022). However, there is still not enough evidence to support firm conclusions regarding the effect of emotional stimuli on language processing in PWA. More research is needed to clarify whether emotion could be facilitative or inhibitory for language processing in PWA and, if so, for which language processes (e.g., repetition, discourse production, auditory comprehension, and word retrieval) and under which circumstances (e.g., highly arousing vs. less arousing stimuli, positive vs. negative stimuli, and abstract vs. concrete stimuli). As discussed in the introduction, continuing to pursue this line of research is worthwhile because of the importance of communication in poststroke quality of life (Cruice et al., 2003; Hilari et al., 2003, 2012) and the potential clinical implications of this work for language assessment and treatment.

Appendix

Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) Checklist

SECTION	ITEM	PRISMA-ScR CHECKLIST ITEM	REPORTED ON PAGE #
TITLE
Title	1	Identify the report as a scoping review.	Title page (p. 4327)
ABSTRACT
Structured summary	2	Provide a structured summary that includes (as applicable): background, objectives, eligibility criteria, sources of evidence, charting methods, results, and conclusions that relate to the review questions and objectives.	Abstract (p. 4327)
INTRODUCTION
Rationale	3	Describe the rationale for the review in the context of what is already known. Explain why the review questions/objectives lend themselves to a scoping review approach.	Introduction (pp. 4327–4329)
Objectives	4	Provide an explicit statement of the questions and objectives being addressed with reference to their key elements (e.g., population or participants, concepts, and context) or other relevant key elements used to conceptualize the review questions and/or objectives.	Introduction (p. 4329)
METHODS
Protocol and registration	5	Indicate whether a review protocol exists; state if and where it can be accessed (e.g., a web address); and if available, provide registration information, including the registration number.	n/a
Eligibility criteria	6	Specify characteristics of the sources of evidence used as eligibility criteria (e.g., years considered, language, and publication status), and provide a rationale.	Methods (p. 4329)
Information sources*	7	Describe all information sources in the search (e.g., databases with dates of coverage and contact with authors to identify additional sources), as well as the date the most recent search was executed.	Methods (p. 4329)
Search	8	Present the full electronic search strategy for at least 1 database, including any limits used, such that it could be repeated.	Methods (Table 1)
Selection of sources of evidence†	9	State the process for selecting sources of evidence (i.e., screening and eligibility) included in the scoping review.	Methods (pp. 4329–4330)
Data charting process‡	10	Describe the methods of charting data from the included sources of evidence (e.g., calibrated forms or forms that have been tested by the team before their use, and whether data charting was done independently or in duplicate) and any processes for obtaining and confirming data from investigators.	Methods (p. 4330)
Data items	11	List and define all variables for which data were sought and any assumptions and simplifications made.	Methods (p. 4330)
Critical appraisal of individual sources of evidence§	12	If done, provide a rationale for conducting a critical appraisal of included sources of evidence; describe the methods used and how this information was used in any data synthesis (if appropriate).	n/a
Synthesis of results	13	Describe the methods of handling and summarizing the data that were charted.	Methods (p. 4330)
RESULTS
Selection of sources of evidence	14	Give numbers of sources of evidence screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally using a flow diagram.	Results (p. 4330, Figure 1)
Characteristics of sources of evidence	15	For each source of evidence, present characteristics for which data were charted and provide the citations.	Results (pp. 4330–4331, Tables 2 and 3)
Critical appraisal within sources of evidence	16	If done, present data on critical appraisal of included sources of evidence (see Item 12).	n/a
Results of individual sources of evidence	17	For each included source of evidence, present the relevant data that were charted that relate to the review questions and objectives.	Results (Table 4)
Synthesis of results	18	Summarize and/or present the charting results as they relate to the review questions and objectives.	Results (pp. 4334, 4337–4338)
DISCUSSION
Summary of evidence	19	Summarize the main results (including an overview of concepts, themes, and types of evidence available), link to the review questions and objectives, and consider the relevance to key groups.	Discussion (pp. 4338–4339)
Limitations	20	Discuss the limitations of the scoping review process.	Discussion (pp. 4340–4341)
Conclusions	21	Provide a general interpretation of the results with respect to the review questions and objectives, as well as potential implications and/or next steps.	Discussion (pp. 4339–4340, 4341)
FUNDING
Funding	22	Describe sources of funding for the included sources of evidence, as well as sources of funding for the scoping review. Describe the role of the funders of the scoping review.	p. 4327

Note. From Tricco et al. (2018).

^*Where sources of evidence (see second footnote) are compiled from, such as bibliographic databases, social media platforms, and websites.

^† A more inclusive/heterogeneous term used to account for the different types of evidence or data sources (e.g., quantitative and/or qualitative research, expert opinion, and policy documents) that may be eligible in a scoping review as opposed to only studies. This is not to be confused with information sources (see first footnote).

^‡ The frameworks by Arksey and O'Malley (6) and Levac and colleagues (7) and the Joanna Briggs Institute guidance (4, 5) refer to the process of data extraction in a scoping review as data charting.

^§ The process of systematically examining research evidence to assess its validity, results, and relevance before using it to inform a decision. This term is used for Items 12 and 19 instead of “risk of bias” (which is more applicable to systematic reviews of interventions) to include and acknowledge the various sources of evidence that may be used in a scoping review (e.g., quantitative and/or qualitative research, expert opinion, and policy document).

Funding Statement

This work was supported by National Institute on Deafness and Other Communication Disorders Grants F31DC017367 and T32DC014435, awarded to Deena Schwen Blackett, and Grant R01DC017711, awarded to Stacy M. Harnish. The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institutes of Health.