Introduction

Primary progressive aphasia (PPA) is a dementia syndrome marked by a gradual loss in communication functioning due to the effects of degeneration on brain networks involved in speech and language processing [1]. This dementia syndrome, can have different causes such as frontotemporal lobar degeneration (FTLD) or Alzheimer´s Disease (AD) has several classical variants, each of which has specific language features.

The different variants of PPA present some typical and predominant linguistic characteristics but there is a wide range of symptom possibilities in addition to those mentioned. Identifying these variants early can improve patients’ quality of life. For instance, the semantic PPA (sv PPA) can be characterized by anomia, evident in discourse and on visual confrontation naming tasks, and by difficulty understanding individual words, deficits associated with gradual loss of semantic knowledge. The non-fluent/agrammatic variant (nfav PPA) is characterized by agrammatism, effortful hesitant speech, as well as phonemic and phonological deficits consistent with apraxia of speech. The logopenic variant (lv PPA) is marked by naming difficulties due to lexical access failures, with the occurrence of anomia in discourse and failures in phrase repetition stemming from deficits in the phonological loop of working memory [1–3]. In addition to these variants, there is also the mixed/unclassifiable form of the disease. These patients do not fit the profiles of any of variants because they either do not have any of the core features of any specific variant or they have more symptoms of the classical variants and in this case the individuals may manifest classic characteristics of more than one subtype [4] precluding a clear diagnosis [5,6].

The language disorder might depend either on the intrinsic deterioration of the linguistic system or on a more general cognitive decline and seem to be the result of an interplay between both. In this sense, the mean of language assessment in overall dementias becomes crucial.

PPA is a rare disease that is complex to diagnose due to the heterogeneous presentation of cases. Despite the importance of language with regard to making differential diagnoses of neurodegenerative diseases, there is still a lack of language assessment tools for all types of dementia, including PPA. Instruments originally developed for assessing post-stroke aphasia are still widely used for diagnosing dementias [7]. Although tests specifically developed for assessing aphasia are not best suited for evaluating language deficits associated with neurodegenerative diseases [8,9], some aphasia symptoms secondary to acute lesions are similar to those seen in PPA [10]. More recently, a number of tests have been proposed [11,12], and some retrospective studies, with large PPA patients have already demonstrated that some combined language tasks are useful to identify PPA and distinguishing patients’ subtype syndrome [13], even though in very early stages from the disease [4] that can help to guide linguistic-cognitive interventions.

In this sense it is also worth bearing in mind that tests should undergo full psychometric analysis, including tests of validity, internal consistency, and reliability [14]. The Montreal Toulouse Language-Brasil Battery (MTL-BR) [15] is the only instrument for language assessment available which normative data in Brazil. It was mainly designed to assess aphasia, then it is important to understand if its tasks (including its stimulus) are useful and suitable for assessing language disorders due to degenerative cases. Moreover, in this context, brief batteries are also needed, since the use of extended batteries are often impractical to implement [13] and ascertaining which language assessment tasks can best identify PPA can be valuable

In view of how some instruments or language tasks can best discriminate PPA the present study sought to identify, based on a comprehensive language assessment - MTL-BR -which language tasks can aid identifying the most and/or early impaired language processing in this population.

Materials and methods

The present study was conducted at the Department of Language, Speech and Hearing Sciences of the Universidade Federal de São Paulo (UNIFESP) – Escola Paulista de Medicina. The study was approved by the local Research Ethics Committee under permit no. 0540/2022. Participants were enrolled onto the study on a voluntary basis after signing the Free and Informed Consent Form, devised in compliance with the recommendations of the National Board of Health (CNS) pursuant to Resolution Ruling 466/12 of the CNS/MS (Ministry of Health).

Results

Sample characteristics

The characteristics of the study sample regarding age and years of education are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive values and comparative analysis of PPA and Control groups according to age and education.

https://doi.org/10.1371/journal.pone.0318155.t001

The results given in Table 1 show a statistically significant difference between the CG and PPAG for education. Although the controls were matched by convenience, with the selection of subjects who had a similar education to patients, a small difference between PPA and Control groups was found (15.97 versus 14.66 years of formal education), but it must be highlighted that it does not seem that this difference could interfere on the results obtained, first because the group with more years of education was the PPA and mainly because there is no difference in performing MTL-BR tasks in subjects with more than 8 years of formal education [14].

There was no statistically significant group difference for age.

Of the 29 patients in the PPAG, 13 exhibited the semantic variant, 11 the logopenic variant, 4 non-fluent and/or agrammatic variant, and 1 had the mixed variant. Regarding performance on the MMSE, the PPA group had a mean score of 23.33 (SD ± 3.70), while the CG had a mean score of 29.17 (SD ± 3.43). Considering the MMSE scores 12 PPA patients were at moderate stage and 17 at mild stage of the disease. Mean time since PPA diagnosis was 30 (SD ± 14.02) months.

The results for the comparison of performance by the PPAG and CG on the MTL-BR Battery subtests are presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Descriptive values and comparative analysis of PPA and Control groups according to performance on subtests of MTL-BR Battery.

https://doi.org/10.1371/journal.pone.0318155.t002

The results revealed a statistically significant difference in performance between the PPAG and CG on the following tasks of the MTL-BR Battery: structured interview, oral comprehension of phrases, oral narrative discourse, written comprehension of phrases, written dictation, sentence repetition, semantic verbal fluency, oral naming of nouns and verbs, object manipulation, phonological verbal fluency, body part recognition and left-right orientation, written naming of nouns, oral text comprehension, number dictation, written narrative discourse, written text comprehension and numerical calculations (mental and written). No statistically significant differences between the PPAG and CG were evident for the other tasks of the MTL-BR Battery.

Discussion

The results of the present study, as expected, showed that performance of the PPAG was worse than the CG on all the language tasks assessed by the MTL-Brasil Battery.

The tasks from MTL- BR Battery allow that many linguistic processing can be investigated and compared. It is very important because language processing is the expression of a complex cognition function and each linguistic task can interact with working and long-term memory, attentional skills, and executive control in different ways [25]. In this sense, understanding the deterioration of the linguistic system can help to establish compensatory mechanisms in an attempt to preserve the communication.

There was a statistically significant difference in performance on the structured interview, oral comprehension of phrases and text, oral narrative discourse, object manipulation, body part recognition and left-right orientation, sentence repetition, oral naming of nouns and verbs, phonological and semantic verbal fluency, written comprehension of phrases and text, written naming of nouns, written dictation, number dictation, and numerical calculations mentally and on paper (Table 2). Our results suggest that maybe these tasks could be considerate for language screening tasks in this population. These findings will be further discussed below. It was also found that the PPAG produced a discourse containing a greater number of words compared to the CG. A higher number of words is not necessarily linked to a better quality of an oral narrative since a higher wordiness or verbosity [26] can be associated with linguistic decline.

Regarding to oral comprehension, the impaired performance on phrases can be expected in PPA because, particularly for the agrammatic/non-fluent variant, patients have difficulty performing syntactic processing [27,28], especially for the non-canonic structures incorporated in the sentences from the MTL-Brasil Battery. In the logopenic variant, this same difficulty may be associated with deterioration of phonological working memory, which influences comprehension of long sentences made up of different syntactic structures [28–30] and although the typical difficulty with svPPA is understanding words, these patients commonly have impaired sentence comprehension as the disease progresses. Similarly, on the object manipulation task by verbal command, the higher the number of command items, the greater the overload of working memory. Likewise, this effect also occurs on the body-part recognition and left-right orientation task, whereby the individual has to process two items of information at the same time (body part plus side). However, when identifying body parts and notions of left and right, difficulties involving semantic recognition may also occur. In fact, a previous study found impaired body-part and body-side identification in AD patients, where the spatial orientation involved in this comprehension task can be difficult for some patients with dementia [9].

Oral comprehension difficulties were also found on the structured interview task. The structured interview task from the MTL-Brasil Battery consists of both simple and complex questions, distributed hierarchically. The simpler questions require comprehension of elements of everyday life and are therefore easier to understand. However, the more complex questions involve analysis of situations about which the examinee is asked their opinion and, therefore, demands more cooperation between language and other cognitive functions, particularly executive function. Similarly, comprehension of text is a complex task that requires interaction of several cognitive domains besides language, especially episodic memory, working memory and attention, together with the demands of oral language and executive functions for retelling the story [27,31]. It is worth mentioning that no statistically significant difference was found between groups considering comprehension of words. Although thirteen patients exhibited the semantic variant, most of them were at mild stage and it seems that MTL stimuli from this task was very familiar and easy for them.

Regarding oral production tasks, difficulties were observed on phrase repetition, possibly due to phonological or grammatical deficits, or associated with apraxia speech. Phrase repetition is one of the main tasks used for identifying logopenic PPA due to failure in short-term phonological memory [32]. In non-fluent/agrammatic PPA, this deficit stems mainly from impaired motor planning of speech which leads to apraxia of speech [33]. However, repetition problems can also be caused by a deficit in syntactic processing [34].

With regard to oral naming of words, this task is especially important in assessing both the semantic [35] and logopenic variants of PPA, due to their associated lexical-semantic or lexical-phonologic deficits [36,37]. In semantic PPA, anomia manifests due to gradual deterioration of semantic knowledge. For logopenic PPA, there are more failures in lexical access, while semantic knowledge tends to remain preserved until moderate stages of the disease [38]. In non-fluent/agrammatic PPA, naming of verbs can be more impacted than nouns [39,40]. Although the stimuli from MTL-BR battery are considerate very easy, in the present study they were able to discriminate groups.

Regarding oral narrative discourse, analysis of language in connected speech provides different information, which is particularly relevant in dementia due to the broad cognitive nature of the disorder. To produce discourse, a speaker must integrate linguistic and non-linguistic skills. For this reason, discourse analysis is seen as a sensitive tool for detecting language impairment [41]. In the MTL-BR battery, discourse production takes place based on the input of a figure. Thus, the task requires adequate visual perception and visual inference, where these elements provide the basis for producing a narrative. The results in Table 2 show that the PPAG produced a discourse containing a greater number of words compared to the CG, but with fewer information units. This finding can be explained by the presence of circumlocutions which occur primarily in semantic PPA and secondarily in logopenic PPA. In other words, despite the presence of anomia, semantic and verbal paraphasias, besides substitution of words with unspecific generic terms, participants with PPA seek to maintain a narrative discourse by repeating ideas and talking around the topic [2,3,42,43]. Moreover, deficits such as agrammatism, morphemic and phonemic paraphasias commonly found in nfav PPA [27] also affect the ability to produce discourse.

Impaired performance on the semantic verbal fluency and phonological tasks were expected given that lexical access issues are common to all dementia syndromes [2,44,45]. Semantic verbal fluency measures the integrity of semantic memory, whereas phonologic verbal fluency is a more sensitive measure of executive dysfunctions and impaired phonologic processing [46]. In this respect, the present study findings corroborate the results of previous studies showing the important role of these tasks in language assessment. Thus, evidence suggests that semantic verbal fluency may be more impaired than phonologic fluency in individuals with svPPA. Conversely, phonological fluency may be more impaired than semantic verbal fluency in patients with nfavPPA. In lvPPA, the evidence is conflicting, with some studies reporting similar performance on both tasks [46,47], while others show greater impairment in phonologic fluency [48].

Considering written comprehension tasks, differences were found between groups on comprehension of phrases and of written text. In the case of comprehension of phrases, dissociation between the lexical and morphosyntactic domains may hamper decoding of grammatical structure [49]. For written comprehension of text, besides the language difficulties outlined, performing this task involves greater cognitive demand, such as information retention, short-term memory, working memory and attention. Moreover, although language deficits predominate in PPAs, patients already exhibited impairments to several cognitive domains, with consequent influence on performing highly demanding tasks.

Regarding written production tasks, compared to CG subjects, patients from the PPAG had lower performance for word dictation, written naming and narrative discourse production.

Writing difficulties are common in individuals with PPA [1], such as central dysgraphia, which can impair word spelling, while difficulties performing grapheme-phoneme conversion and peripheral dysgraphia, such as grapheme buffer failure [50] can occur, with a greater incidence of peripheral errors than central errors in some cases [51].

The word dictation task of the MTL-BR battery consists of non-words together with both regular and irregular words. Hence, this task involves a number of writing subsystems of the lexical and phonological routes, which may be impaired in different variants of PPA [50,52–54]. In addition, the association of central dysgraphia with peripheral components is not uncommon in svPPA [50,55].

On the written naming task, specific deficits caused by the lexical and phonological routes are associated with failures in lexical-semantic access.

In written narrative discourse, unlike in oral narrative discourse, there was a statistically significant difference between the two groups only for the elements making up the scene (robbery at the bakery). Written text production may require more planning time, and there is more opportunity for formulating and reformulating ideas [56], factors which likely contributed to better performance on this task relative to the oral discourse production. Moreover, difficulties beyond the specific writing disorders already mentioned were expected because it is known that producing discourse is a complex task that involves retrieving information from memory and deciding on which elements should be included or excluded, entailing major interaction among cognitive functions.

No group difference was detected on the non-verbal praxis tasks. Although the literature shows that deficits on this test can be expected in patients with agrammatic PPA [33,57], the presence of motor difficulties is relative [58]. In the present sample, only 4 patients had the agrammatic variant, which can develop with apraxia. Furthermore, the non-verbal praxis task, in the manner executed in the battery, without repetitive gestures or time control, appears to have low sensitivity for detecting more subtle motor impairments.

Regarding to numerical knowledge and calculation, the PPAG had poorer performance than the CG on tasks such as number dictation and mental calculations. In fact, the transcoding tasks, since they involve language processing, may be especially hard for people with language deficits [59], while dyscalculia may also be present, seen in both acute lesions and PPA [59,60].

Lastly, although the MTL-BR battery seems to be an excellent instrument for language assessment, as all instruments, it has limitations. As a battery that was originally design for persons with aphasia some stimuli from some tasks, as pointed out before, are easy. At this point, some tasks cannot be so good to differentiate deficits in a very early beginning of the disease and/or for highly educated individuals.

Conclusions

This study identified some differences in language processing in PPA patients. Language changes should be monitored and could make us better understand how compensatory strategies aiding communication could be implemented. In this sense, an overview of language impairment can be helpful.

Therefore, the tasks that PPA patients had poorer performance at MTL-BR such as structured interview, oral comprehension of phrases, oral narrative discourse, written comprehension of phrases, written dictation, repetition of phrases, semantic verbal fluency, naming nouns and verbs, object manipulation by verbal command, phonological verbal fluency, body part recognition and left-right orientation, written naming of nouns, oral text comprehension, number dictation, written text comprehension and numerical calculation (mental and written) should be prioritized in assessments of patients with PPA and incorporated into brief tools for use in routine neurologic exams and for diagnosis of PPA.

Supporting information