Comparing two picture naming tasks in Primary Progressive Aphasia: insights from behavioural and neural results

Picture naming is a multi-componential process encompassing dissociable yet interacting stages, from mapping the visual input to the semantic representation, to linking it to the word form for spoken output (Gleichgerrcht et al., 2015). It entails a widespread brain network, including the occipital regions, running along the fusiform and temporal gyri and the temporal pole, and further extending to the inferior frontal cortices (Catricalà et al., 2020; Migliaccio et al., 2016). Picture naming tests are widely used to assess language impairments in neurological patients, both after cerebrovascular accidents (McKinnon et al., 2018) and in different neurodegenerative diseases (Catricalà et al., 2020), including Primary Progressive Aphasia (PPA, Gorno-Tempini et al., 2011).

Different kinds of errors, e.g. visual, semantic, phonological, and anomia, may arise during the naming process. Although there is no one-to-one correspondence between the error type and the impaired processing stage, qualitative scoring provides relevant information about the neuro-cognitive components that are likely to be damaged (Balthazar et al., 2010; Schwartz et al., 2009). In our previous study on naming across neurodegenerative patients, we found: a) visual errors to be correlated with hypometabolism in the occipital lobes, involved in visual processing; b) semantic errors to be correlated with brain metabolism in the left fusiform and inferior and middle temporal gyri, extending to the temporal pole, regions which are known to be associated with conceptual knowledge; c) phonological errors to be correlated with hypometabolism in the left superior and middle temporal gyri, brain regions involved in the coding for the phonological forms (Catricalà et al., 2020).

Impaired naming is among the core diagnostic features of both the semantic variant of PPA (sv-PPA) and the logopenic variant PPA (lv-PPA) (Gorno-Tempini et al., 2011). Impaired naming qualitatively results in lexical-semantic errors and anomia. These kinds of errors are present since the early stages of the disease both in sv-PPA and in lv-PPA, but they may be attributed to different causes. In sv-PPA, naming errors have been ascribed to the progressive loss of semantic knowledge, emerging from the progressive involvement of the anterior temporal lobe (Jefferies and Lambon Ralph, 2006; Catricalà et al., 2021; Mion et al., 2010). Coordinate semantic errors reflect the initial blurring between the boundaries of concepts belonging to the same semantic category, resulting from the degradation of relevant distinctive features (Catricalà et al., 2015). As the disease progresses, additional features are lost and superordinate errors and anomia tend to predominate (Hodges et al., 1995). Impairments at the lexical level have been also reported in this syndrome (Mesulam et al., 2013). In lv-PPA, lexico-semantic errors have been mainly attributed to a dysfunction in the retrieval of phonological word form, associated with cortical atrophy in the left middle and superior temporal gyri (Leyton et al., 2017; 2019; Win et al., 2017; Leyton et al., 2012; Migliaccio et al., 2016) and resulting in anomia and phonological errors. Accordingly, lv-PPA patients may show preserved knowledge of the item, for example by producing accurate circumlocutions, and their performance is improved by cueing (Leyton et al., 2013). A semantic origin of naming errors is however possible also in these patients (Roher et al., 2010; Teichmann et al., 2013). Phonological errors are associated with cortical thinning in the left posterior superior temporal gyrus (Leyton et al., 2014).

Naming is also affected in patients with the non-fluent variant of PPA (nfv-PPA), even though the impairment is generally less severe as compared to the other variants (Henry et al., 2015). In nfv-PPA, in addition to articulatory distortions, errors are mainly phonological, including omission, substitution, deletion, or insertion of phonemes (Migliaccio et al., 2016).

Picture naming tests are the most frequently used for the assessment of word retrieval abilities. The currently available picture naming tests differ in many factors potentially affecting the performance, such as the format of the stimuli (e.g., coloured pictures as in Savage et al., 2013 vs black and white line drawings as in Kaplan et al., 1983; Mckenna and Warrington, 1983), the number of stimuli (e.g., from 6 as in Lindeboom, 2002 to 80 as in Deloche and Hannequin, 1997), the presentation time (e.g., ranging from no limits, as in Mckenna and Warrington, 1983, Savage et al., 2013, to 20 seconds as in Kaplan et al., 1983), as well as the psycholinguistic properties, i.e. differences in frequency, familiarity and typicality. Previous studies in aphasia reported, although not univocally, several insights on how different variables may influence not only performance accuracy but also differences in error types (Nickels and Howard, 1995; Cuetos et al., 2002; Kremin et al., 2003; Kittredge et al., 2008). Studies on neurodegenerative diseases have largely focused on the semantic variant of PPA (Barbarotto et al., 1995; Lambon Ralph et al., 1998, Woollams et al., 2008) and suggested that naming accuracy is predicted by typicality, frequency, familiarity, and age of acquisition. Anomic errors are elicited by items with low typicality, characterized by unusual features failing to activate any name, whereas semantic errors are more frequent in the case of high typicality items, whose features overlap among semantically related exemplars (Woollams et al., 2008).

Of note, different naming tests are used interchangeably, sometimes according to the usual practice of the clinical and research centres, without any clear agreement on which test is preferable in evaluating a specific performance in specific types of patients.

A recent systematic review reported in fact how not all naming tests are equally effective in discriminating, for example, lv-PPA from other PPA variants (Conca et al., 2022). Indeed, in the majority of cases, the Graded Naming Test (Mckenna and Warrington, 1983; Warrington, 1997) revealed no difference in naming accuracy between sv-PPA and lv-PPA, while other instruments, such as the DO80 naming test (Deloche and Hannequin, 1997), the Boston Naming Tests (Kaplan et al., 1983; Mack et al., 1992), and the Northwestern Naming test (Thompson and Weintraub, 2014), showed a worse accuracy of sv-PPA patients. All these tests included black and white line drawings, but the Graded Naming Test has no time limit for the response, while the other two tests have time limits of 6 and 20 seconds, respectively. Different results have been also reported in comparing lv-PPA and nfv-PPA, with no difference in the Boston Naming Test (Kaplan et al., 1983), and worse performance of lv-PPA than nfv-PPA in SYDBAT (Savage et al., 2013) and DO80 naming test (Deloche and Hannequin, 1997). SYDBAT, differently from the Boston Naming and DO80 tests, includes coloured photographs and has no time limit for the response. Crucially, the lack of information about the characteristics of the stimuli and the diagnostic metrics prevents a straightforward interpretation of the results. In fact, the characteristics of the naming tests may result in different weighting on components of the neural networks linked to specific aspects of language, such as visual perception, lexico-semantics and phonology, leading to specific performance profiles in PPA variants.

The present study aims to investigate the behavioural characteristics, both in terms of proportion of correct responses and error type, of two Italian naming tests, i.e. CaGi naming subtest (CaGi, Catricalà et al., 2013) and Screening for Aphasia in NeuroDegeneration naming subtest (SAND, Catricalà et al., 2017), administered to a sample of PPA patients. We investigated the capacity of the two tests in discriminating PPA versus controls and among PPA variants, taking also into account the psycholinguistic variables affecting performances. We also explored the correlation between naming performance in both tests and regional brain metabolism as assessed by FDG-PET in PPA cases. Besides the research purpose, identifying the psycholinguistic characteristics of different naming tests, which best discriminate across PPA variants and best unveil the nature of the naming impairment, may improve the management of PPA patients in both diagnostic and rehabilitation settings.

留言 (0)

沒有登入
gif