Assessment of Language Impairment and Function

Abstract

This chapter is written for the qualified neurologist or related professional working with persons who have had a stroke or other sudden brain injury. It is critical that the presence of aphasia is detected, no matter how mild the presentation, and to support that assertion, this chapter highlights the plight of persons with latent aphasia. At the individual level, the impact of aphasia is devastating, with overwhelming evidence that aphasia negatively impacts psychosocial outcomes. At the global level, sensitive detection and accurate diagnosis of aphasia is critical for accurate characterization and quantification of the global burden of aphasia. The word “LANGUAGE” is leveraged as an acronym to create a useful and memorable checklist to guide navigation of aphasia screening and assessment: it begins with the definition of language (L), followed by the definition and diagnostic criteria for aphasia (A). Then language abilities and characteristics to be considered in assessment are presented: naming (N); grammar and syntax (G); unintelligible words, jargon, and paraphasias (U); auditory comprehension and repetition (A); graphemic abilities – reading and writing (G); and everyday communication and discourse (E). Recommendations for improving procedural adherence are provided, and a list of potential brief assessment measures are introduced.

1. Purposes of aphasia assessment

This chapter is written for the qualified neurologist, neuroradiologist, neurosurgeon, neuroscientist, physiatrist, or other clinical neuro-professional working with persons who have had a stroke or other sudden brain injury. For those clinicians, the primary purposes of aphasia assessment are detection, preliminary prognosis, and for some professionals with further specialization, differential diagnosis. In all cases of suspected or confirmed aphasia, the neuro-professional will relay the information to the qualified speech-language pathologist (SLP) for more in-depth assessment that will determine the presence or absence of aphasia, and if present, the type and severity of aphasia. The SLP will then be able to establish the present level of performance with regard to patient strengths/weaknesses and functional communication abilities, and this information will be used to guide treatment planning and patient (and family, if present) education and/or counseling. The assessment information obtained by the SLP should then be shared with the entire patient care team to assist with prognosis refinement and disorder management.

1.1. Impact of aphasia and importance of detection

It is critical that the presence of aphasia is detected, no matter how mild the presentation. At the individual level, the impact of aphasia is devastating, with evidence that aphasia leads to significantly worse psychosocial outcomes than those experienced by typically aging peers (e.g., Cruice, Worrall, & Hickson, 2006; Ross & Wertz, 2003) as well as peers without aphasia with otherwise similar profiles (e.g., social support, physical limitations) (e.g., Hilari et al., 2012; Northcott et al., 2016). (The impact of aphasia versus cognitive-communication disorder has not been directly compared in a large representative sample.) These negative outcomes include more limited activity and participation, higher incidence of depression, decreased social networks and support, and diminished health-related and overall quality of life (Cruice et al., 2006; Hilari, 2011; Hilari et al., 2012; Hilari & Northcott, 2017; Northcott et al., 2016; Pike et al., 2017; Ross & Wertz, 2003; Simmons-Mackie, 2018; Wallace, 2010). Persons with aphasia (PWAs) are often unable to fulfill pre-stroke life roles and responsibilities, which leads to increased reliance on others, reduced ability to independently navigate society, and overall reduced autonomy. PWAs of working-age are un- or under-employed (Pike et al., 2017; Simmons-Mackie, 2018), and decreased vocational opportunities, coupled with increased medical costs, introduces financial strain (Ellis et al., 2012; Ganzfried, 2018). Because language is used daily, and frequently throughout the day, to survive and/or to thrive, the finding that language deficits, in the form of aphasia, lead to such negative and wide-ranging consequences is therefore not surprising. Aphasia diminishes the ability to share feelings and ideas, make comments, rely on language for learning, ask and answer questions, debate, advocate, protest, worship, and much more. This affects communication opportunities with family members and friends, and broader social and community networks, including those related to navigating the community for survival (e.g., essential services like groceries, healthcare), continued learning and development, vocation, and avocation (Dalemans et al., 2008; Davidson et al., 2008; Gavreau et al., 2019; Viscogliosi et al., 2011).

At the global level, sensitive detection and accurate diagnosis of aphasia is critical for accurate characterization and quantification of the global burden of aphasia so that suitable healthcare opportunities and funding are provided (e.g., Rohde et al., 2018). Using the United States as an example, Simmons-Mackie (2018) estimates that the prevalence of aphasia due to sudden acquired causes (i.e., not progressive) ranges from 2,463,681 to 4,108,469. The wide range is largely due to differences in detection and diagnostic criteria, but even at the lower bound, the numbers are staggering. These numbers are expected to grow, as the year 2030 will mark a demographic turning point in the United States, when the entire baby boomer generation will be over 65 years of age (Vespa et al., 2018). While stroke and other related causes of aphasia do occur throughout the lifespan, they are more common in the elderly; as the population ages, the incidence and prevalence of stroke and related causes will increase. The global burden of stroke is already alarming and has not improved since 1990, whether measuring mortality, disability-adjusted life years, or years lived with disability (Feigin et al., 2015). The presence of aphasia increases that burden, for in addition to the aforementioned listing of poorer outcomes associated with aphasia, aphasia significantly correlates with greater use of inpatient and outpatient services (e.g., inpatient complications, longer stays, more visits) and subsequent higher financial costs, as well as mortality (e.g., Lazar & Boehme, 2017; Wu et al., 2020). These aphasia-related consequences for PWAs and their families, and for society at large, highlight the importance of sensitive detection and effective rehabilitation of aphasia (Brady et al., 2016).

Aphasia treatments are under continuous refinement and development in the quest for improved outcomes. There are now more aphasia treatment options than ever before, and recent advancements even reach the gold standard of generalization to untrained stimuli. Further, though some have expressed concerns that age may play a deleterious role in efficacy (even suggesting that perhaps compensation rather than remediation should be the focus), aphasia treatment has proven to be effective for the elderly (Fabian et al., 2020). There remains much progress to be made for intervention, and there is just as much work to be accomplished for assessment. Undetected and/or misdiagnosed aphasia goes unmanaged and untreated, leading to potentially “avoidable disability” and greater overall burden (Balogh et al., 2015; Power et al., 2015; Sackett, 1992). Critically, the earliest diagnosis, often made by readers of this chapter, has a significant impact on eligibility for later services, and it is difficult, sometimes even impossible, for patients to receive a diagnosis of aphasia if there is historical medical documentation stating that they do not have aphasia. This will be discussed later in the chapter when the topic of very mild or latent aphasia is highlighted.

1.2. Surface differential diagnosis

One of the primary purposes of initial aphasia screening is a surface-level differential diagnosis. The first decision is in response to the question “Does the patient have a communication problem?”. If the answer is yes, the second decision is in response to the question “Is it aphasia and/or some other communication disorder (apraxia of speech, cognitive-communication disorder, dysarthria)?”; the “and/or” is intentionally included here since these problems can and do co-occur. Either way, if it is aphasia and/or other communication disorder, it should trigger a referral to a qualified SLP.

There is no universally agreed upon classification system for aphasia subtypes; in fact, there are over 20 different classification systems (Ardila, 2010). The most popular classification systems are those that arose from what is known as the “Boston Group”, with one of the earliest iterations proposed by Benson and Geschwind in 1971, and the most widely adopted today (Wallace et al., 2019), via use of the Western Aphasia Battery-Revised (Kertesz, 2006), first proposed by Kertesz in 1979. Step one for classification for this system involves determination of fluency, where fluency is defined as the ability to effortlessly and quickly produce speech, although the speech may contain errors and may not be meaningful. Many neurologists (and related professionals) commonly describe aphasia as being “fluent” or “nonfluent” in referring paperwork. However, caution is urged here, as fluency is a complex construct, even for the specialist. Fluency is typically determined, either implicitly or explicitly, by performance on three underlying abilities: grammatical competence (e.g., an individual’s ability to produce grammatically complete sentences), lexical retrieval (e.g., an individual’s ability to use the specific content-rich words necessary to communicate meaning), and motor speech production (e.g., an individual’s ability to clearly and accurately produce speech sounds in words, phrases, and sentences) (Clough & Gordon, 2020).

Persons with nonfluent aphasia often present with halting, limited output that is topically relevant. They are often described as producing “telegraphic” speech given their difficulty producing functors (conjunctions, determiners, prepositions, etc.), which leads to a higher ratio of content words (nouns, verbs, adjectives, adverbs) to functors than is typically found in healthy speakers (also referred to as agrammatism). However, not all individuals with nonfluent aphasia demonstrate agrammatism -- for some, difficulty initiating speech leads to a determination of nonfluent production, while for others, difficulty accurately and easily producing speech sounds may contribute more to the determination. At the other end of the spectrum, persons with fluent aphasia may have relatively intact grammatical competence or may overuse grammatical markers (known as paragrammatism). These individuals may be able to produce words, phrases, and sentences fairly easily, but the words produced may not match well with the topic and may be nonsensical. For persons with fluent aphasia with better auditory comprehension and thus greater awareness of their errors, fluent speech may be interrupted by longer pauses as they identify errors and attempt to repair the production. These behaviors may lead to the perception by some raters that the person is more nonfluent, especially if fluency is determined by overall output, such as the number of words produced over time (e.g., words per minute [wpm]). Persons with fluent aphasia may produce word-level errors during speech, such as “cot” instead of “cat” (referred to as a phonological paraphasia). However, these errors can be difficult to distinguish from motor speech errors that lead to incorrect sound production more common in individuals with nonfluent aphasia. Importantly, fluency determinations are highly subjective decisions which rest upon the three underlying abilities listed above, and there is a great deal of room for disagreement regarding the relative impact of each ability to the overall classification. For this reason, unless fluency is very clear, “aphasia” or “suspected aphasia” is a simpler referral terminology that does not rely upon subjective judgements regarding fluency.

The terms “expressive” and “receptive” are also commonly used to characterize aphasia in an attempt to highlight the more prominent deficits. A person with more obvious difficulty expressing their thoughts through spoken (and/or written) language might be described as having “expressive aphasia” whereas a person with prominent difficulty understanding language presented through auditory and/or visual modalities might be described as having “receptive aphasia”. Here again caution is urged – this dichotomous characterization can lead to erroneous assumptions. For example, if someone is described as having expressive aphasia, it may be interpreted that receptive abilities are intact. However, it is more likely in this example that receptive abilities are a relative strength compared to expressive abilities, but that they are not “intact” or “normal” and are still in need of assessment and/or intervention. As above, the recommendation is that the general referral terminology “aphasia” or “suspected aphasia” is utilized until auditory comprehension can be examined more thoroughly.

2. LANGUAGE checklist for aphasia assessment

The word “language” was leveraged as an acronym to create a useful and memorable checklist to guide navigation of aphasia screening and assessment (See Table 1). The definition of language (L) is first provided, followed by the definition and diagnostic criteria for aphasia (A). Then, language abilities and characteristics to be considered in assessment are presented: naming (N); grammar and syntax (G); unintelligible words, jargon, and paraphasias (U); auditory comprehension and repetition (A); graphemic abilities – reading and writing (G); and everyday communication and discourse (E).

Table 1.

LANGUAGE checklist

L	Language
A	Aphasia
N	Naming
G	Grammar and syntax
U	Unintelligible words, jargon, and paraphasias
A	Auditory comprehension and repetition
G	Graphemic abilities
E	Everyday communication and discourse

2.1. Language

Language is a form of communication that uses a finite set of arbitrary symbols in various combinations to express an infinite number of concepts between community members for the purpose of social cooperation and survival. Symbols can include phonemes (i.e., speech sounds), graphemes (i.e., letters), handshapes (as in American Sign Language), and logograms (i.e., symbols that represent a complete language unit or word). Language consists of five domains: 1) phonology - the sounds in a language (phonemes) and rules that govern sound combinations; 2) morphology - the smallest units in a language which have meaning (morphemes) and their rules for use and combination (these can include words, like “cat”, or grammatical markers like “-ed” to show past tense); 3) syntax - the rules that determine how words are organized into sentences; 4) semantics - the meaning of words and groups of words; and 5) pragmatics - how language is used in specific contexts to communicate and the rules governing its use (Berko Gleason, 2005). These domains can be organized into three overarching categories related to: 1) the form of language (i.e., phonology, morphology, syntax); 2) the content of language (i.e., semantics); and 3) the use of language (i.e., pragmatics and discourse) (Bloom & Lahey, 1978). In most typical adults, language processing occurs primarily in the brain’s left hemisphere, specifically in areas around the lateral sulcus (Sylvian fissure) that separates the frontal, temporal, and parietal lobes and lies over the insula lobe. This area is often referred to as the perisylvian language area, cortex, region, or zone.

2.2. Aphasia

Aphasia results from damage to an established language system following a stroke, traumatic brain injury, and/or other nontraumatic brain injury (e.g., brain infection) that alters the structure and/or function of language centers in the brain and/or of connections between those centers. (Progressive aphasias are not addressed directly in this chapter.) Aphasia is most frequently associated with cerebrovascular damage to the primary supplier of blood to perisylvian language regions, the left middle cerebral artery. Though aphasia is derived from the Greek “a” + “phatos”, meaning “speechless” or “not speakable”, it is not restricted to deficits in the speech, or spoken language, domain. Aphasia involves diminished production and/or comprehension of the underlying language system and the language symbols, meaning that PWAs will demonstrate varied combinatorial profiles of expressive language difficulty (i.e., via speaking, writing, and/or signing/gesturing) and receptive language difficulty (i.e., hearing words, reading words, and/or watching gestures).

There have been many definitions of aphasia over the years, and Hallowell (2016) recently provided a functional definition of aphasia with the following four essential elements: 1) it is an acquired disorder that affects already mastered language abilities; 2) it arises from a neurological cause; 3) all language modalities are impacted to some extent; and 4) it is a disorder of language, not sensory, motor, psychiatric, or intellectual abilities. Aphasia is therefore an acquired language disorder with an underlying neurological cause that affects reception and production of language across all modalities (e.g., auditory comprehension, spoken language, signed language, reading, and writing, etc.) and is not the result of a sensory, motor, psychiatric, or intellectual disorder.

2.3. Naming

Anomia is the “absence of a word or words” in language (Bayles & Tomoeda, 1982) and is the primary diagnostic feature of all aphasia subtypes (Benson, 1988; Goodglass & Geschwind, 1976; Goodglass & Wingfield, 1997). Anomia is most often linked to difficulty with retrieval of noun (pictured or real objects) or verb (pictured or demonstrated actions) names, but it can also arise from difficulty with selection, use, and/or generation (Bayles & Tomoeda, 1982). There are many ways to assess naming, but the focus here is on the two types of naming most commonly utilized in screening.

2.3.1. Confrontation naming

Standardized picture and object confrontation naming assessments for nouns are the most widely used tools for PWAs in clinical and research settings (e.g., Brady et al., 2016). It involves “confronting” the patient with a picture or an object and asking them to provide the name. Confrontation naming assessments are easy to administer, well-validated, and reliable. These assessments have consistently revealed differences between PWAs and healthy controls, as well as differences between different aphasia subtypes (Bastiaanse & Jonkers, 1998; Kohn & Goodglass, 1985). Confrontation naming correlates highly with overall aphasia severity and thus diagnosis (presence or absence) and classification (subtype). However, how well these assessments predict connected spoken language (i.e., discourse) depends on aphasia subtype. For example, Richardson et al. (2018) investigated picture naming (nouns and verbs) and the ability to convey “gist” in a large sample of PWAs and found that correlations were strong for individuals with Broca’s and Wernicke’s aphasia, moderate for those with conduction and anomic aphasia, and not significant for individuals with very mild or latent aphasia (the latter related to obvious ceiling effects of the naming assessments).

Recently, increased attention has been paid to confrontation naming of verbs. There is ample evidence that verb production is impaired in PWAs (see Cho-Reyes & Thompson, 2012), and may be more impaired than noun naming in some PWAs (e.g., Berndt et al., 1997b). Evidence suggests that verb naming impairments may be equally pervasive across all aphasia types (Akinina et al., 2020). In English, verbs are central to sentence construction, which means that performance on verb naming may better predict sentence-level and connected speech abilities than noun naming (Berndt et al., 1997a; Bastiaanse et al., 2003; Cho-Reyes & Thompson, 2012). In fact, preliminary evidence suggests that therapy focused on remediating verb naming abilities may yield improvements in sentence production, a critical generalization for improving functional communication (e.g., Thompson et al., 2013).

2.3.2. Generative naming

Generative naming tasks require patients to “generate”, or name, different words beginning with a certain letter (i.e., letter fluency) or within a prescribed category (i.e., semantic fluency) within a restricted time frame (i.e., 60 seconds). The primary variable is the number of accurate and unique words produced. Secondary variables involve analysis of clustering and switching. In letter fluency, clusters are sequentially generated words that meet the following criteria: 1) share the first two letters; 2) differ only by a vowel sound; 3) rhyme; or 4) are homonyms (i.e., same letters or sounds but differ in meaning). For semantic fluency, clusters are sequentially generated words that belong to the same subcategory, and Troyer (2000) provided sample populated subcategories (e.g., Farm animals: chicken, cow, donkey, ferret, goat, horse, mule, pig, sheep, turkey; p 377). Switches are the number of transitions between clusters (e.g., the total words generated minus the words that were part of clusters). Similar to confrontation naming assessments, generative naming assessments are easy to administer, well-validated with large normative databases (since they are also a staple of neuropsychological research), and reliable. Detailed rules for scoring along with norms for several letters and semantic categories have been provided by Troyer (2000).

Letter fluency was first introduced during a written task in the 1930s (Thurstone, 1938). It was then modified by Benton and colleagues and eventually incorporated into two aphasia assessments, the Multilingual Aphasia Exam (MAE; Benton & Hamsher, 1976) and the Neurosensory Center Comprehensive Examination for Aphasia (NCCEA; Benton, 1967; Spreen & Benton, 1969, 1977) (see Tombaugh, Kozak, & Rees, 1999 for a more detailed history). Letter fluency tasks have also been called phonemic fluency tasks, but this is not entirely accurate. Phonemes correspond to speech sounds, which may or may not correspond to letters. The instructions for letter fluency tasks instruct a patient to “name as many words as [they] can think of that begin with that letter”, not that begin with a certain sound. Letter fluency is also known as the Controlled Oral Word Association Test (COWAT; Benton & Hamsher, 1976) and is included in the popular neuropsychological task referred to as verbal fluency, used in that domain because it probes executive functioning and working memory in addition to language abilities. While letter fluency completed with the letters F-A-S is the most well-known (Tombaugh et al., 2000; Loonstra et al., 2010), there are well-established norms for two alternate lists using letters C-F-L and P-R-W (Ross et al., 2006; Ruff et al., 1996) that clinicians will find useful but that have yet to be fully adopted for aphasia assessment. Analysis of clusters and switches for letter fluency may also provide more insight into patient deficits.

Semantic, or category, fluency was introduced as a part of aphasia assessments, first the WAB (Kertesz, 1982) followed soon by the BDAE (Goodglass & Kaplan, 1982) (see Tombaugh, Kozak, & Rees, 1999 for a more detailed history). For semantic fluency, patients are asked to generate names of nouns within a category, and the category of animals is the most commonly utilized and researched (Tombaugh, Kozak, & Rees, 1999; Loonstra, Tarlow, & Sellers, 2010). Other notable categories include fruits, vegetables, colors, boys’ names, and cities/towns (Swanson, 2005; Tombaugh et al., 1999). Troyer (2000) also introduced rules and norms for several categories.

Generative naming may be more sensitive to aphasia than confrontation naming, though it is confounded by the increased reliance on other cognitive abilities. For example, in a study of Cognitive-Linguistic Quick Test (CLQT; Helm-Estabrooks, 2001) performance in PWAs, patients performed fairly well on nonlinguistic tasks as expected, with 31% to 69% performing within the normal range depending on the task. However, only 15% performed within the normal range for confrontation naming, and 0% for generative naming (Helm-Estabrooks, 2002).

2.4. Grammar and syntax

Grammar refers to the language rules utilized by speakers and listeners to convey and understand meaning. These rules involve complex interconnections between semantics, the lexicon (word meanings paired with combinations of sounds, letters, and/or gestures), phonology, morphology, and finally, syntax (Jacobs, 1995). Too often, this highly complex rule system is reduced to a discussion of agrammatism versus paragrammatism, and agrammatism further reduced to decisions regarding whether or not a PWA is telegraphic/non-fluent (see previous fluency discussion).

An example of a quick probe of expressive grammar and syntax is on the Quick Aphasia Battery (QAB; Wilson et al., 2018). Patients are presented with a picture of a semantically reversible event (e.g., a picture of a boy kissing a girl) and are asked “What is happening here?”. The clinician is also to engage the patient in conversation on a list of suggested topics. The clinician then rates the connected speech, and two of the 10 connected speech feature ratings are 1) “Agrammatism: omission of words and morphemes, especially closed class; ‘telegraphic speech’”, and 2) “Paragrammatism: substitution of closed class items; inappropriate juxtaposition of words and phrases; garbled syntax.” Even this simple description and straightforward rating requires some specialized knowledge, contributing to the absence of such tasks or ratings on non-specialist measures. For example, the word “morpheme” is not a common term, and it refers to the smallest meaningful unit of language that cannot be further subdivided. Morphemes can be free, meaning they can stand alone (e.g., “test”, “scope”), or bound, meaning they cannot stand alone (e.g., “pre-“, “-ing”). Open class words refer to the larger subset of vocabulary that is constantly growing and evolving - nouns, verbs, adjectives, and adverbs. Closed class words, also called function words, refer to the smaller subset of vocabulary that is primarily fixed - prepositions, conjunctions, pronouns, and determiners (Shipley & McAfee, 2015). The assessor is therefore listening for a number of morpheme-related variables, for example: 1) Is plural and/or possessive “-s” used correctly? 2) Is the speaker using closed-class morphemes or omitting them? 3) Are verb tenses marked correctly (e.g., “-ed”, “-ing”, “-s/z”)? 4) Are rules for combining free and bound morphemes followed (e.g., “-ly” can be appended to the adjective “quick” but not to the noun “house”)? The assessor is also listening for related syntax violations, such as: 1) Is a subject or object in the wrong place in a sentence? 2) Is a plural used for a single item, or vice versa? 3) Do complex and compound sentences have the appropriate connectors (e.g., coordinating clauses)?

Sampling connected spoken language, or discourse, is critical for assessment of grammar and syntax. However, this type of assessment is not as common as other areas (e.g., lexical retrieval) and may be missing from screening or diagnostic instruments entirely. This leaves a great deal of rich language information unexamined, and even dismissed as uninformative, if the instrument or checklist does not call it to the assessor’s attention. Clinicians should monitor connected spoken language for these variables, even if informally until formal instruments are developed, for even if a patient is mostly accurate and just a little “off” occasionally (e.g., the wrong verb tense, not marking plurals, etc.), it likely warrants a referral as “suspected aphasia” to an SLP.

2.5. Unintelligible words, jargon, and paraphasias

Word-level errors are an obvious aphasia symptom and can assist with identifying what is often coined a “frank” aphasia. However, the absence of these errors, or the presence of these errors only in connected speech, should not be used to rule out aphasia. Some assessments may have formal inventories or rating scales related to word errors, but if not, it is still an item on the LANGUAGE checklist needing attention for adequate screening for aphasia.

Unintelligible words may arise due to differences in the motor speech domain or the language domain. If caused by motor speech difficulties, errors can occur at the speech sound level - substitutions, distortions, omissions, etc. - and make the word unrecognizable. Motor speech disorders commonly co-occur with nonfluent aphasias and should trigger an SLP referral, whether it occurs in isolation or in combination with aphasia. In the language domain, unintelligible words are often linked to neologisms, or “new words” - nonwords that are not recognizable as being related to any potential target, even with context. For example, a patient may be trying to name a picture of a “tomato” (/təmeɪɾoʊ/) and they say “kruge” (or /krudʒ/).

While this nonword does follow the phonological rules of English, it is not related to the target word - it does not share any phonemes (speech sounds), it differs in syllabic structure, etc. Sometimes neologisms may share phonemes with the target word, but there are various thresholds of overlap applied to determine if a word is a neologism (i.e., a non-word that bears little similarity [e.g., <50% overlapping phonemes] to the target) versus a nonword phonological paraphasia (i.e., a non-word that bears some similarity [e.g., >50% overlapping phonemes] to the target). For our “tomato” example, if the patient said “tuge” (or /tudʒ/), it would still be considered a neologism, since it only shares one phoneme (/t/). However, if the patient said “mitumo” (or /Mətumoʊ/), it would be considered a nonword phonological paraphasia because it shares more phonemes (/m,ə,t,oʊ/). When the connected spoken language of a patient is full of neologisms, this is called neologistic jargon.

Another type of jargon is semantic jargon, which is when a patient’s connected spoken language involves real words, but in combinations that have little meaning. For example, when telling the Cinderella story, a person with conduction aphasia said the following: “And so now the lady the guy who’s owned is gone so his dead and so nobody can help the two girls. And the old one has been lift alone the other one who’s lifted by himself and been telling a love from everybody and can’t be making the piss for himself. She’s got the up and we chipped at the little health that keep them to been above. Just looked around and everything is kept for.” Semantic jargon may occur alongside neologistic jargon and/or with other paraphasias.

Semantic paraphasias are real words that are semantically related to the target word. In our “tomato” example, a semantic paraphasia could be “onion” or “pepper”. Formal phonological paraphasias are real words that are phonologically related to the target word. In our “tomato” example, whereas “kruge” and “tuge” and “mitumo” are nonwords, “motor” is a real word that is phonologically (i.e., shares /m,oʊ,ɾ/), but not semantically, related to the target. Mixed paraphasias are real words that share both semantic and phonological similarity to the target. For example, a mixed paraphasia for “tomato” could be “potato”, as it is semantically similar (i.e., they are both edible, they belong to the category of fruits and vegetables) and phonologically similar (i.e., shares /ə,t,eɪ,ɾ,oʊ/). Patients may produce other errors, such as perseveration, which is the inappropriate recurrence or continuation of a previous word or phrase, and verbal stereotypies, which are repeated words or nonwords, or repeated phrases, that may be the only thing the patient can say.

2.6. Auditory comprehension and repetition

Before conducting an assessment of auditory comprehension and/or repetition, it is critical to determine whether or not the patient’s auditory perceptual abilities are intact and if not, to ensure their hearing is aided during assessment. Assessment of auditory comprehension and repetition is commonplace in aphasia assessment, as they are both considered critical tasks for making decisions regarding aphasia subtypes. Auditory comprehension tasks often include point-to tasks (by object name, action name, descriptor, etc.), following simple directions, and answering Y/N questions. Repetition tasks often include word and sentence repetition tasks.

Despite the usual inclusion of probes for auditory comprehension, the information gained is under-utilized for diagnostic purposes. Subsequently, auditory comprehension deficits are underemphasized in clinical practice as well as the broader aphasia literature (e.g., Ellis et al., 2020). For example, a recent analysis of the contributory weights of subtests on the WAB-R found that auditory comprehension contributes ~20% or less to the overall aphasia quotient (AQ), so that expressive language tasks contributed the remaining ~80%. The WAB-R is thus essentially a test of expressive language. Historically, a 5-point change in the AQ has been thought to be “clinically significant” (Katz & Wertz, 1997; Gilmore et al., 2019), despite the flaw in this judgement that ignores the test’s reliability (Bothe & Richardson, 2011; Jacobson & Truax, 1991). This means that the “clinically significant change” pursued for many years has been change in expressive language, which has implicitly shaped the focus and development of treatment approaches. Another common practice that devalues auditory comprehension assessment information relates to our previous discussion of “expressive” and “receptive” aphasia, for once a patient is determined to have an expressive profile, it is often erroneuously assumed that their receptive abilities are intact.

Repetition tasks require auditory perception, retention of the auditory signal, and then conversion of the auditory signal to an articulatory trajectory so the patient can repeat the target. Repetition tasks are most useful for differential diagnosis of aphasia subtype (e.g., Kertesz, 2006) and for characterizing the integrity and function of language areas in the brain (white and gray matter) (Fridriksson et al., 2018). While repetition tasks can be used as “gateway” tasks to facilitate other relevant language abilities, improving repetition should not be targeted as the end-goal for PWAs.

2.7. Graphemic abilities - reading and writing

Assessment of reading and writing is often overlooked even though these abilities are key defining elements for both language and aphasia. PWAs typically experience some degree of alexia, or acquired reading impairment (Beeson & Insalaco, 1998; Cherney, 2004). While rare, alexia can also occur in isolation, without other language deficits of speaking, understanding, or writing. The research for alexia treatment lags behind that for spoken language deficits, although several treatments have been developed (Beeson et al., 2010; Cherney et al., 1986; Coelho, 2005; Moyer, 1979; Woodhead et al., 2013, 2018). As a result, clinical assessment (and treatment) of alexia is less frequent (Simmons-Mackie et al., 2005), despite the fact that only a brief reading screening is usually needed to determine the presence or absence of alexia (Cherney, 2004). Reading assessment tasks may include the following (e.g., Cherney, 2004, Rapcsak et al., 2007): reading comprehension (from letter recognition to paragraph comprehension), oral reading of words and nonwords, oral reading of sentences and paragraphs, and spelling. If reading deficits are detected, more in-depth assessment to determine alexia subtype (e.g., pure alexia, surface alexia, semantic alexia, etc.) can be conducted following a referral to an SLP.

It is also common for PWAs to experience agraphia, or acquired writing impairment, along with aphasia and/or alexia, though it can occur in isolation as well (Purdy, 2016; Roeltgen, 1993). Agraphia research and clinical practice is similarly underdeveloped. The presence of agraphia can be detected with a brief writing screening, to be followed by a thorough assessment by an SLP to determine agraphia subtype (e.g., phonological agraphia, semantic agraphia, deep agraphia, etc.). Writing assessment tasks may include the following (e.g., Beeson et al., 2010; Kertesz, 2006): letter conversion (e.g., lower to upper case), copying (shapes, letters, words, sentences), writing-to-dictation, and written picture description.

Assessment of reading and writing is needed to ensure that the communication supports utilized by the healthcare team are appropriate and effective. For example, when someone has difficulty understanding spoken language because of their aphasia, it is common to provide supplemental written materials, which will have limited utility if they have alexia. Similarly, when someone has difficulty expressing their thoughts with spoken language because of their aphasia, it is common to prompt them to try to write it down, which will provide little help if they have agraphia. When assessing reading and writing, it is important to collect relevant history related to their pre-injury level of abilities. It is also important to screen visuoperceptual abilities to assist with differential diagnosis or to confirm that the patient has both visuoperceptual deficits and reading and/or writing deficits (Cherney, 2004).

2.8. Everyday communication and discourse

Because of the increasing awareness and acknowledgement of the devastating impact of aphasia, providers of PWAs are encouraged to adopt a life and social participation approach when assessing, treating, educating, and counseling. This brings to the forefront activity and participation domains in the World Health Organization International Classification of Functioning, Disability, and Health (WHO-ICF) framework (2007) and the complementary Life Participation Approach to Aphasia (LPAA; Chapey et al., 2001; Elman, 2016; Simmons-Mackie & Kagan, 2007). This is supported by stakeholder input into outcomes, in which PWAs, caregivers, and providers consistently prioritize activity and participation goals (e.g., increased autonomy, participation in life roles, participation in leisure, improved psychosocial well-being, return to work, etc.) alongside language and/or communication goals (Brown et al., 2012; Wallace et al., 2016; Worrall et al., 2011). A recent and important change was the redefining of primary outcomes in an esteemed Cochrane Library systematic review (Brady et al., 2016): “The primary outcome chosen to indicate the effectiveness of an intervention that aims to improve communicative ability must reflect communication activity in real world settings, that is functional communication.” (p. 6). Impairment-based measures were consigned to secondary outcomes, most appropriate for use as a surrogate or delegate for functional communication when psychometrically sound and/or widely adopted measures of functional communication are not available. Functional communication measures should address the ability to convey a meaningful message to a listener in everyday communication exchanges, and measurement of communication ability above the sentence level, or discourse, is an endorsed measure of functional communication (Brady et al., 2016).

2.8.1. Naming and discourse

As aphasia outcomes are re-prioritized and the focus turns to functional communication, the utility of naming assessments is called into question. As previously mentioned, naming measures have long and frequently been utilized for their ease of use and interpretability. This widespread adoption has thus shaped clinical and research practice, including billing processes and research directions. As such, there is abundant evidence of treatment-induced change in naming measures as well as other measures that rely heavily on naming (e.g., WAB-R aphasia quotient [AQ], an index of overall aphasia severity). Naming assessments may be useful for estimating the potential impact of aphasia in connected spoken language, but so far, research focused on describing relationships between lexical retrieval during picture naming and lexical retrieval during narrative and conversational tasks has returned equivocal results. Studies have shown a lack of significant correlation between naming and real-world communication abilities (e.g., telling a story, having a conversation) and suggest naming may not be a strong predictor of discourse abilities (e.g., Mayer & Murray, 2003; Fergadiotis & Wright, 2016). Correlations between naming and discourse abilities may be present, but only for some subtypes - for example, Richardson et al. (2018) found strong correlations between naming (object and verb) and narrative “gist”, but only for persons with Broca’s and Wernicke’s aphasia. Other measures are needed to predict functional communication across aphasia types and severities.

2.8.2. Discourse and functional communication

Discourse is “language that is beyond the boundaries of isolated sentences” (Ulatowska & Olness, 2004, p. 300) and requires successful combination and stringing of linguistic elements into a coherent and meaningful message (Wright, 2011). Discourse pervades activities and participation opportunities - socializing and relationship building, talking with family members and coworkers, reminiscing about old stories, expressing needs in the community, seeking or sharing information for the purposes of cooperation, etc. There is some evidence that discourse abilities measured in structured clinical settings predict how well PWAs can participate in everyday real-world conversations (e.g., Doyle et al., 1995; Mayer & Murray, 2003). Common discourse elicitation techniques include picture description, story retelling, procedural descriptions, and unstructured topic-related or open-ended conversations (Bryant et al, 2016).

Calls have been made for development of a core outcomes set for discourse measures (Dietz & Boyle, 2018) similar to core outcomes already developed for more commonly used aphasia tests (Wallace et al., 2019). A discourse core outcome set is possible thanks to repositories such as the AphasiaBank database (Fromm et al., 2020), which hosts hundreds of discourse samples from PWAs and healthy, non-brain-injured controls, contributed by researchers around the world. These samples are collected with a standardized elicitation protocol that includes conversational speech, story retell, picture description, sequenced picture description, and a procedure (or “how-to”) task. AphasiaBank also includes a web page for publications and discourse measures developed using the database (https://aphasia.talkbank.org/discourse/), giving clinicians access to standardized elicitation protocols, standardized discourse measures with detailed scoring instructions, and normative data that may aid in interpretation of client scores. To ensure that these calls for a core discourse outcome set become a reality, the FOQUSAphasia (FOstering QUality of Spoken discourse in Aphasia; https://foqusaphasia.com/) group was established, bringing together clinicians and researchers with the intent to increase the use of spoken discourse measures in clinical and research settings (Stark et al., 2020).

2.8.3. Streamlined discourse assessment

There are over 500 discourse measures (Bryant et al., 2016); thorough review is beyond the scope of this chapter and is not needed for the level of screening discussed in this chapter. Presented below are measures that can be administered without lengthy transcript preparation or specialized phonetic transcription as long as the clinician has access to the procedures and manuals for each approach. With these measures or similar, clinicians can be sure to assess a critical component of the LANGUAGE checklist, everyday communication and discourse, which is perhaps the most critical component for those with very mild deficits. In addition to the measures discussed below, some assessments may include useful rating scales with discourse sampling. For example, the rating scale for the Quick Aphasia Battery (QAB; Wilson et al., 2018) addresses speech rate, errors, grammar, utterance length and complexity, and more.

2.8.3.1. Content units (CUs) were first formally introduced by Yorkston and Beukelman (1980) for use with the Cookie Theft picture. A CU is defined as “a grouping of information that was always expressed as a unit by normal speakers” (Yorkston & Beukelman, 1980, p. 30). The example they give is “The ¹ little ²boy is ³on the stool and ⁴reaching up for a ⁵cookie and he’s going to ⁶fall over”, which includes 6 content units. When scoring, one point is assigned to each CU on the list, then all points are tallied for a total CU score. Words that are similar to the units listed are allowed (e.g., lady for woman). Analysis of CUs does not require phonetic transcription and can be scored online. Recent work has further expanded CUs by adding a left-to-right ratio (L:R CU ratio), which is “the ratio of the number of CU produced from the left side of the picture to the number of CU produced from the right side of the picture” (Keator et al., 2020). This ratio can provide additional information about brain dysfunction. Early work with CU analysis revealed differences between PWAs and controls for content and efficiency (Yorkston & Beukelman, 1980), which is supported by the more recent study for PWAs in acute to chronic phases of recovery (Keator et al., 2020). Further, Keator et al. (2020) showed that CUs in the acute phase post-stroke strongly predicted CUs in the chronic phase, suggesting that this efficient measure has high prognostic value.

2.8.3.2. Core lexicon (CoreLex) analysis is used to investigate the typicality of words in discourse (Dalton & Richardson, 2015; Fromm et al., 2013; Kim et al., 2019; MacWhinney et al., 2010). For example, during the Cinderella story retell task, a healthy control might use lexical items such as “slipper” and “prince”, whereas an individual with aphasia might use lexical items such as “shoe” and “guy”. CoreLex is scored by assigning one point for each checklist item produced by a participant and then all points are tallied for a total CoreLex score. Synonyms of CoreLex items are not given points, but inflected CoreLex items are. For example, the Cinderella CoreLex checklist includes the verb “run”; productions such as “jog”, “sprint”, and “race” would not receive credit, but productions of “ran” and “running” would. The gist of the message might still be communicated with these less typical items (“shoe”, “guy”, “jog”, etc.), but the quality of the communicative interaction may decrease (e.g., reduced listener perception of success and/or comfort, increased cognitive effort). CoreLex correlates with both word-level and utterance-level discourse performance (Dalton & Richardson, 2015; Kim et al., 2019). CoreLex does not require phonetic transcription, and is amenable to online scoring, as clinicians could simply check off each item that a client produces from a checklist. Recent work with CoreLex has involved the publication of a tutorial on CoreLex (Kim & Wright, 2020), a compendium of CoreLex checklists for various stimuli (Dalton, Kim, et al., 2020), and norms (Dalton, Hubbard, & Richardson, 2020).

2.8.3.3. Main concept analysis (MCA) is an approach that facilitates characterization of how effectively an individual communicates the gist or essential elements of a discourse task (Nicholas & Brookshire, 1993; Nicholas & Brookshire, 1995). According to Nicholas and Brookshire (1993), a main concept (MC) consists of a single main verb and any constituent arguments (e.g., subject, object, direct object) and associated clauses. The verb, arguments and clauses are considered “essential elements” that must be produced in order to communicate the gist of the concept. For example, MC1 from the Cinderella story retell (“The ¹father ²remarried a ³woman”) contains three essential elements: 1) father; 2) remarries; 3) woman. In order to conduct MCA, utterances that align with MCs are evaluated for the accuracy and completeness of essential elements. The following codes are assigned to utterances matching an MC: Accurate/Complete (AC) - all essential elements are present and all elements are correct (following the example above, “the ¹dad ²married a ³lady with two daughters”); Accurate/Incomplete (AI) - some essential elements are missing, but those produced are accurate (for example, “the ¹dad ²remarried”); Inaccurate/Complete (IC) - all essential elements are present, but some are not accurate (for example, “the ¹father ²dated a new ³woman”); Inaccurate/Incomplete (II) - some essential elements are missing, and of those produced, some are incorrect (for example, “the ¹uncle ²married”). Missing MCs are scored as Absent (AB).

MCA is sensitive to differences between clinical and control populations (Kong, 2009; 2011; Kong et al., 2016; Fromm et al., 2017; Hameister & Nickels, 2018; Dalton & Richardson, 2019) and correlates with listeners’ perceptions (Ross & Wertz, 1999). Recent work in MCA has involved the development of checklists and preliminary normative data in English (Kong et al., 2016; Richardson & Dalton 2016; 2019; Hameister & Nickels, 2018), Irish-English (Kong et al., 2012), Spanish (Rivera et al., 2018), Cantonese (Kong 2009, 2011), and Taiwanese (Kong & Yeh, 2015). While the initial development of MCA may not have been conducted with non-transcription-based analysis in mind, these recent advances may allow clinicians to reliably use it in this manner.

2.8.3.4. MCA has recently been enveloped and expanded into Main Concept, Sequencing, and Story Grammar, or MSSG, analyses (Greenslade et al., 2020). In a single analysis of the Cinderella story retell, MSSG measures both gist as well as organizational structure and sequencing. To begin, the 34 MCs established by Richardson and Dalton (2016) were pre-assigned one of six story grammar component codes (e.g., setting, initiating event, direct consequence, etc.) and were also pre-assigned an episode number (1-5, as there are 5 episodes in the Cinderella story). Utterances corresponding to MCs were coded for accuracy and completeness according to the rules and guidelines described above (Nicholas & Brookshire, 1995; Richardson & Dalton, 2016) and then assigned a numeric value representing the continuum of accuracy and completeness (Kong, 2009; Richardson & Dalton, 2016). Using our example above, a speaker would receive 3 points for the accurate and complete production of “The ¹father ²remarried a ³woman”. A total of those values across all concepts yields the main concept (MC) composite score. Each MC then receives a sequencing score representing whether it was produced in the correct order, and scores are totaled across MCs. Returning again to our example, if the speaker produced “The ¹father ²remarried a ³woman” in the appropriate place (i.e., the opening of the story), they would receive 3 points; if this utterance was produced out of place, they would receive fewer points. These two scores are then summed for the main concept + sequencing (MC+Sequencing) score. Next, the presence or absence of required episodic story grammar components (i.e., initiating event, attempt, or direct consequence) is marked and tallied for the total episodic components score. The MC “The ¹father ²remarried a ³woman” would not be included in this tally, as it is a “setting” component, whereas MC17 “¹Cinderella ²went to the ³ball” would be included because it is a required episodic component (i.e., direct consequence). Finally, complex episodes (episodes with at least two out of three required story grammar components) ae marked and tallied for the episodic complexity score. Recent investigations of MSSG indicate it is sensitive to differences between healthy, non-brain-injured controls and PWAs as a group, as well as between healthy controls and aphasia subtypes, including individuals with latent aphasia (Richardson et al., 2021).

2.8.4. Discourse and detection of latent aphasia

There are PWAs who score at or above ceiling on commonly used standardized tests (e.g., Western Aphasia Battery-Revised [WAB-R]) and therefore usually do not meet diagnostic criteria for aphasia. These individuals may receive labels such as “clinically undetectable” aphasia (Olsen et al., 1986), latent aphasia (Boller & Vignolo, 1966; DeDe & Salis, 2020; Pichot, 1955; Silkes et al., 2020; but see Heilbrun, 1958), not aphasic by WAB (NABW; Dalton & Richardson, 2015; Fromm et al., 2017), or very mild aphasia (Cavanaugh & Haley, 2020). (For the purposes of this chapter, we will use the term and abbreviation “not aphasic by WAB” and “NABW”, since the majority of research conducted with this group has used the WAB-R aphasia cutoff to define them.) This group has even been used as “controls” for test development (Kertesz, 2006) and also likely composes a substantial proportion of those described in the literature base as “recovered”. The lack of mainstream assessment measures sensitive to their deficits leads to an inability to qualify for clinical services or meet research inclusion criteria and a scarcity of treatment programming options for addressing higher-level deficits, even though they experience limited activity and participation and negative psychosocial impacts (Armes et al., 2020; Cavanaugh & Haley, 2020).

PWAs who are not aphasic by WAB differ significantly from controls for CoreLex and MCA. This group differs from controls also for number of utterances, lexical diversity (as measured via moving average type-token ratio), and lexical entropy (as measured via word information measure ), during discourse (Cunningham & Haley, 2020; Fromm et al., 2017). Temporal discourse measures are also sensitive to latent aphasia. For example, this group performs significantly differently for the derived efficiency measure AC/min (or accurate and complete main concepts produced per minute) but not for CoreLex/min (Dalton et al., 2020). Other differences have been reported for words per minute, silent pause durations, and percentage of formulation time (DeDe and Salis, 2020; Fromm et al., 2017).

PWAs who are not aphasic by WAB seem to differ from their nearest clinical neighbor, persons with anomic aphasia, for word- and utterance-level measures (e.g., words per minute, mean length of utterance, moving average type-token ratio) (Cunningham & Haley, 2020; Fromm et al., 2017) as well as temporal measures (e.g., articulation rate) (DeDe and Salis, 2020). However, for suprasentential measures such as those related to MCA and story grammar, these two groups – anomic and not aphasic by WAB - look quite similar, with very few significant differences between the two groups observed (Dalton et al., 2015; Dalton & Richardson, 2019; Dalton et al., 2020; Fromm et al., 2017; Richardson et al., 2021). While the focus of this discussion has been on spoken language, there is evidence that persons with left hemisphere damage determined to be non-aphasic demonstrated subclinical auditory comprehension deficits on the Token Test, called latent sensory aphasia (Boller & Vignolo, 1966), that are likely to be mirrored by reading comprehension deficits (McNeil et al., 2015). There is not currently a brief clinical instrument for frontline detection of such mild deficits, but the long history of developing shorter and/or and computerized versions of the Token Test (e.g., De Renzi & Faglioni, 1978; McNeil et al., 2015; Park, McNeil, & Tompkins, 2000) is encouraging.

It is clear that PWAs who are not aphasic by WAB constitute a distinct clinical group positioned between, and overlapping with, both healthy controls and persons with anomic aphasia and are in need of additional research and clinical focus. To illustrate this point, Figure 1 shows the similarities and differences between controls and two PWAs who are not aphasic by WAB on a range of traditional aphasia tests as well as discourse measures. The Western Aphasia Battery – Revised aphasia quotient (WAB-R AQ; Kertesz, 2006) determines the presence or absence of aphasia. Scores range from 0-100, and the cut-off for aphasia diagnosis is a score below 93.8. The Boston Naming Test (BNT; Kaplan et al., 2001) and Verb Naming Test (VNT; Cho-Reyes & Thompson, 2012) are commonly used to evaluate the severity of noun (BNT) and verb naming impairments. The Sentence Comprehension Test (SCT; Martin et al., 2008) evaluates comprehension of complex sentence structures such as passive constructions. Finally, a shortened version of the Discourse Comprehension Test (DCT; Brookshire & Nicholas, 1993) evaluates auditory comprehension of short stories. For these traditional measures, we do not display boxplots for distribution of control performance because controls perform at or near ceiling on these tests. The normative data and boxplots of the distribution of control performance for the discourse measures - Core Lexicon, Main Concept Composite Score, Sequencing Score, and Total Episodic Components - are taken from previously published data (Dalton & Richardson, 2019; Richardson et al., 2021). As the boxplots illustrate, there is a wide range of “normal” discourse performance (and no ceiling). For each measure we have plotted the scores for three individuals - one healthy, non-brain-injured control (blue diamond), and two persons not aphasic by WAB (orange pentagon, purple cross) matched for sex, age, and level of education. The healthy control performs at or near ceiling for all traditional measures and within the range of healthy control performance on the discourse measures. The two PWAs who are not aphasic by WAB perform similarly on traditional measures whereas discourse measures clearly demonstrate impaired language production in connected speech. Specific scores are as follows, in order of control (blue) and persons not aphasic by WAB (orange, purple): Western Aphasia Battery – Revised aphasia quotient (100) – 100, 97.2, 95; Boston Naming Test (60) – 60, 56, 52; Verb Naming Test (22) – 22, 18, 17; Sentence Comprehension Test (20) – 20, 18, 18; Discourse Comprehension Test (24) – 24, 20, 17; Core Lexicon (94) – 88, 34, 42; Main Concept Composite (102) – 86, 34, 35; Sequencing (102) – 88, 36, 42; and Total Episodic Components (15) – 15, 7, 9.

Figure 1.

Scores for commonly administered aphasia tests and discourse measures. Blue boxplots and points indicate healthy, non-brain-injured control performance. Orange and purple points indicate performance of two individuals not aphasic by WAB (NABW) on each test and measure. Maximum scores for each test and measure are displayed at the top of the figure. (WAB-R AQ - Western Aphasia Battery - Revised Aphasia Quotient; BNT - Boston Naming Test: VNT - Verb Naming Test: SCT - Sentence Comprehension Test; DCT - Discourse Comprehension Test; MC Comp - Main Concept Composite score; Seq - MSSG Sequencing Score; Episodic Comp - MSSG Total Episodic Components Score)

Figure 1 highlights the lack of sensitivity in common aphasia tests, and also shows the value, indeed the necessity, of discourse assessment to identify “mild” but functionally devastating language impairments. Importantly, for the two PWAs who are not aphasic by WAB whose data are shown in this figure, while both were working age, neither was able to maintain employment at their pre-stroke level or even underemployment, primarily due to their undiagnosed residual language impairment.

3. Screening and tracking aphasia

3.1. Assessment fidelity

There has been a recent focus on assessment fidelity for clinical trials to ensure study validity (Richardson et al., 2016; Spell et al., 2020). The recommendations made for clinical trialists can also inform and improve clinical assessment, as they draw attention to the importance of qualifications, training, assessment psychometrics, and adherence to assessment procedures. The recommendations also emphasize the vital role of the assessor as part of the assessment instrument - setting the pace, time limit monitoring, providing instructions/cues/feedback, and scoring - meaning that the assessor must be taken into consideration when evaluating the reliability and validity of assessment instruments and the resultant inferences.

If assessment results are to be interpreted within the context of the normative references accompanying the instrument, then the assessor must adhere to the instrument procedures (e.g., Magasi et al., 2017). When procedural deviations occur (e.g., to accommodate patient abilities or fatigue, procedural drift or contamination, etc.), as is bound to happen in clinical settings, then the assessor must interpret results with caution, as the sensitivity and specificity of the instrument will also be changed. For example, naming subtests on the Quick Aphasia Battery (QAB; Wilson et al., 2018) incorporate timing of response into the scoring, so that a higher score is obtained if the person responds within 3 seconds, and a score of 0 is given if the person responds after 6 seconds. In contrast, the Mississippi Aphasia Screening Test (MAST; Nakase-Thompson, 2004) has no instructions regarding how much time is allowed for response, and no differential scoring according to time of response. Contaminating one of these instruments with the administration procedures of the other would render the normative references meaningless. For the aforementioned letter fluency example, people may perform differently if they are incorrectly instructed to “name as many words as [they] can think of that begin with the sound /f/” (as they may include words that begin with letters “ph”, e.g., phone, photo, pharmacy, etc.) versus the correct instructions to “name as many words as [they] can think of that begin with the letter ‘f’”, impacting the interpretability of results.

This discussion of course also pertains to elicitation of discourse. Examples of different instructions for picture scene description that may influence the quality, quantity, and organizational structure of discourse are as follows: for the picture of the river scene from the Frenchay Aphasia Screening Test (FAST; Enderby et al., 1986), speakers are instructed to “Tell me as much about the picture as you can”; for the study of the Cookie Theft picture by Keator et al. (2020), speakers were instructed to “Tell me everything you see going on in the picture”; and for the updated Modern Cookie Theft picture by Berube et al. (2019), speakers were instructed to “Tell me everything that is happening in the picture, as though you were describing it for the blind.” As another example, for the Cinderella storytelling, speakers are supposed to first look through a wordless Cinderella picture book, and then tell the Cinderella story after that book has been removed, according to the standardized AphasiaBank protocol. However, some clinicians and researchers allow the speaker to tell the story while looking through the book, and one can easily discern that this will impact performance.

3.2. Selected screening and tracking instruments

Table 2 lists several brief measures reviewed for their appropriateness and psychometric strengths by Salter et al. (2006), El Hachioui et al. (2017), and/or Rohde et al. (2018), or that were discovered via related literature searches. As several recent reviews point out (Salter et al., 2006; Hachioui et al., 2017; Rohde et al., 2018), the fields of neurology and aphasiology are lacking measures proven for differential diagnosis of aphasia in stroke (or other brain-injured) populations. The most strongly recommended measures are listed, including some newly developed instruments with strong psychometrics. A standalone instrument is not recommended at this time, and there is insufficient data to provide rankings of instruments. It is recommended that several instruments be combined to be sure the LANGUAGE checklist is efficiently but thoroughly addressed.

Table 2.

Brief Aphasia Assessment Measures

Brief Measures – Screening
FAST	Frenchay Aphasia Screen Test	Enderby et al. (1986); Enderby and Crow (1996)
LAST	LAnguage Screening Test	Flamand-Roze et al. (2011); Flowers et al. (2015)
MAST	Mississippi Aphasia Screening Test	Nakase-Thompson (2004); Nakase-Thompson et al. (2005)
QAB	Quick Aphasia Battery	Wilson et al. (2018)
Brief Measures – Discourse
CoreLex	Core Lexicon	Dalton et al. (2020); Kim and Wright (2020); Dalton, Hubbard, and Richardson (2020)
CUs- Cookie Theft	Content Units – Cookie Theft	Keator et al. (2020); Yorkston and Beukelman (1980)
CUs – Modern Cookie Theft	Content Units – Modern Cookie Theft	Berube et al. (2019)
MCA	Main Concept Analysis (with checklists)	Richardson and Dalton (2016, 2020); Dalton and Richardson (2019); Dalton, Hubbard, and Richardson (2020); Nicholas and Brookshire (1995)
MSSG	Main Concept, Sequencing, and Story Grammar	Greenslade et al. (2020); Richardson et al. (2021)
Brief Measures – Screening with language subtests of cognitive assessments
ACE-R	Addenbrooke’s Cognitive Examination – Revised	Gaber, Parsons, and Gautam (2011) (language component)
OCS	Oxford Cognitive Screen	Demeyere et al. (2015) (language component)
Tracking Measure (after diagnosis of aphasia)
ART	Aphasia Rapid Test	Azuar et al. (2013)
Brief Measures – Impact (after diagnosis of aphasia)
BOMPA	Basic Outcome Measure Protocol for Aphasia	Kagan et al. (2020)
CETI	Communicative EffecTiveness Index	Lomas et al. (1989)
CCRSA	Communication Confidence Rating Scale for Aphasia	Babbitt et al. (2011)
SIS	Stroke Impact Scale	Duncan et al. (1999, 2001)

Glossary

Agrammatism

Difficulty producing grammatical markers. Agrammatic speech is characterized by a higher ratio of content to function words, missing agreement markers (tense, person), and inaccurate pronoun use.

Agraphia

Acquired writing impairment that commonly co-occurs with other acquired language impairments.

• Deep – Writing impairment which combines difficulty converting sounds in a word to the corresponding letters with a semantic impairment.
• Global – Writing impairment caused by incomplete or absent activation of word representations in the mental lexicon.
• Graphemic Buffer – Difficulty holding words in memory while selecting the appropriate letters.
• Pure – Writing impairment without co-occurring language deficits in speaking, understanding, or reading.
• Surface – Writing impairment caused by a breakdown between the mental representation of a word and its corresponding graphemes.

Alexia

Acquired reading impairment that commonly co-occurs with other acquired language impairments.

• Deep – Reading impairment that has characteristics of phonological alexia along with semantic impairment.
• Neglect – Reading impairment caused by visual neglect.
• Phonological – Reading impairment caused by breakdown between the visual word form and the phonological system such that letters cannot be converted to the corresponding sounds.
• Pure – Reading impairment without co-occurring language deficits in speaking, understanding, or writing.
• Semantic – Reading impairment caused by difficulty connecting the visual word form and sounds in the word to their meaning.
• Surface – Reading impairment caused by the incomplete or absent activation of word representations in the mental lexicon after processing the visual word form.

Anomia

Difficulty producing a word or words in language. This is the primary diagnostic feature for all aphasia subtypes.

• Confrontation Naming – Patient is presented with an object or image and asked to provide the name.
• Generative Naming – Patient is presented with a letter or semantic category and asked to state items that belong to that group.

Aphasia

An acquired language disorder with an underlying neurological cause that affects reception and production of language across all modalities (e.g., auditory comprehension, spoken language, signed language, reading, and writing, etc.) and is not the result of a sensory, motor, psychiatric, or intellectual disorder.

Apraxia of Speech

Motor speech disorder caused by a disruption in the planning and programming of motor plans for producing speech sounds.

Assessment Fidelity

Procedures and trainings used to ensure that all assessments are completed in a reliable and valid manner so that results can be compared to normative data and/or results can be replicated across time and assessors.

Cognitive-Communication Disorder

Communication impairment arising from impairment in cognitive domains such as attention, memory, and executive function.

Discourse

Language that is beyond the boundaries of isolated sentences, which requires successful combination and stringing of linguistic elements into a coherent and meaningful message.

Dysarthria

Sensorimotor speech disorder caused by changes in the strength and/or control of the musculature used to produce speech sounds.

Expressive Language

the ability to communicate one’s thoughts with others. This can be via any modality, such as speaking, writing, or using signed language.

Fluency

The ability to effortlessly and quickly produce speech, although the speech may not be meaningful and may contain errors. Fluency is determined by grammatical competence, lexical retrieval, and motor speech production.

Grammar

The language rules utilized by speakers and listeners to convey and understand meaning. These rules involve complex interconnections between semantics, the lexicon, phonology, morphology, and syntax.

Grapheme

Written letters of a language.

Jargon

Pattern of speech where connected spoken language is fluently produced but is not meaningful.

• Neologistic – Connected spoken language is full of non-word productions that lack meaning (e.g., neologisms).
• Semantic – Connected spoken language is full of real word productions that have meaning individually, but do not combine to create a meaningful utterance.

Language

A form of communication that uses a finite set of arbitrary symbols in various combinations to express an infinite number of concepts between community members for the purpose of social cooperation and survival.

Lexicon

The words in a language, or in an individual’s repertoire for a language.

Logogram

Symbols that represent a complete language unit or word. These could be sounds, letters, handshapes, etc.

Morphemes

The smallest units in a language which have meaning. These can include free morphemes which can stand alone (e.g., cat) and bound morphemes which must be produced with a free morpheme (e.g., past tense -ed).

Morphology

The smallest units in a language which have meaning and their rules for use and combination.

Paragrammatism

Overuse of grammatical markers in language production.

Paraphasia

Word production errors that are a result of a language impairment. Paraphasias are classified into different types to better understand where language difficulties may be occurring.

• Formal phonological – Errored word production that results in a real word that is phonologically related and semantically unrelated to the target word.
• Mixed – Errored word production that meets the criteria for both phonological and semantic types.
• Neologism – Errored word production that results in a non-word; literally, “new word”.
• Nonword phonological – Errored word production that results in a nonword that is phonologically similar to the target word.
• Semantic Errored word production that results in a real word that is semantically related to the target word.

Perseveration

The inappropriate recurrence or continued production of a previously produced word or phrase.

Phonemes

The speech sounds of a language.

Phonology

The sounds in a language and rules that govern sound combinations in a language.

Pragmatics

Language used in context to communicate and the rules governing its use.

Primary Progressive Aphasia

Acquired language disorder resulting from progressive neurological disease that presents initially as an isolated language impairment. Often the result of an underlying Alzheimer’s or Fronto-temporal dementia type pathology.

Receptive Language

The ability to understand the communication of others. This can include any modality, including auditory comprehension, text or braille reading, or viewing signed language.

Repetition

The ability to say back, or repeat, a word, phrase or sentence.

Semantics

The meaning of words and groups of words.

Syntax

Rules that determine how words are organized into sentences.

Verbal Stereotypies

Repeated frozen forms of words, nonwords, or phrases that are produced across a variety of contexts, regardless of the appropriateness of their use. These may be the only utterances the individual is able to produce.