Primary Progressive Aphasia: A study of picture naming, interference and lexical access

(1)

Primary Progressive Aphasia:

A study of picture naming, interference

and lexical access

Laura ten Dijke

University of Amsterdam

MA Thesis General Linguistics

2019

(2)

1. Introduction

The speech production in day to day life is thought to be an effortless and automatic mechanism. The actuality, however, reflects a much more complex picture. It involves a multitude of processes including conceptualization, semantic, syntactic, phonological processing and articulation (see Harley, 2014b for a more detailed discussion). Schiller and Verdonschot (2015) highlight lexical access as its crucial component. Lexical access refers to ‘‘the process of retrieving a word from the mental lexicon and encoding it for speech’’ (Wilson et al., 2017, p. 90). While the underlying workings and models of lexical access are still debated (e.g. Garrett, 1975; Dell, 1986; Levelt, 2001), most of these models assume that the process of retrieving words is a competitive one (e.g. Starreveld and La Heij, 1996). This indicates that there is a competition between candidates for lexical selection (Schiller & Verdonshot, 2015). A testing paradigm that has been used to empirically to examine this claim is the (visual) picture-word interference (PWI) paradigm, which is a modification of the Stroop task (e.g. Glaser & Düngelhoff, 1984). The paradigm elicits interference effects and relies on executive (interference) control processes needed for successful word selection (Pai, Roelofs & Schriefers 2012). Piai, Roelofs and Schriefers (2012) claim that these mechanisms ‘‘allow the participants to respond to the target picture, rather than to the distractor word’’ (p. 615). This implies that selective attention toward the target is needed. Selective inhibition plays an additional role. Shao, Meyer and Roelofs (2013) define selective inhibition as the ability to ‘‘suppress specific competing responses’’ (p. 1200). This is relevant when a distractor needs to be ignored in order to name the target1_.

The experimental design of PWI and its implications will be further discussed in chapter three.

The present study examines naming abilities2_{of individuals with Primary Progressive Aphasia (IwPPA)}

with the use of a picture-word interference task (with related and neutral distractors), in which response time and accuracy are measured as experimental output parameters. In addition, the cognitive and linguistic performance is taken into consideration. The present investigation aims to explore the breakdown in naming in relation to the interference control processes mentioned above. Moreover, this thesis explores the factors that play a role in the difficulties and errors that arise. Word finding difficulties or anomia3_in

naming are especially present among individuals with aphasia, a language disorder due to brain damage. This is also the case with Primary Progressive Aphasia (PPA), a disorder characterized by the progressive deterioration of language functions with relatively spared cognitive domains (Mesulam, 1982, 2003; Grossman & Ash, 2004). PPA has been categorized in three main variants of a semantic, nonfluent and logopenic kind (Gorno-Tempini et al., 2011).

1_{Diamond (2013) states that interference control consists of selective attention and cognitive inhibition). Moreover,}

Thomas, Rao & Devi (2016) reported that the processes of attention and inhibition are related, showing a significant association between the two. They will therefore not be analysed separately, and instead I will use the term

interference control.

2_{In this thesis, naming is understood as the process comprised of three stages including the identification of the}

object, access to its semantic representation so that it can be recognized and the activation of its phonological representation so that its lexicalized and used for speech (Spezzano & Radanovic, 2010).

(4)

Thus far only few studies have included PWI to investigate aphasia and PPA (e.g. Hashimoto & Thompson, 2010; Thompson et al., 2012; Vandenberghe et al., 2005 Thompson et al., 2012). There is much room to apply PWI in order to obtain new insights into different variants of PPA. This study may provide a stepping-stone for future more in-depth research with aim to come to a better understanding of lexical access in PPA and of the stratification of PPA patients into subgroups.

In the following sections I elaborate on PPA and the characteristics of the recognized subtypes. In addition, I discuss the various theories concerning the nature of naming deficits and how they are exhibited by the variants. Along with the aims, the experimental study is introduced. These results are presented and subsequently discussed. I emphasize the need for including more participants of different subtypes and interference stimuli to obtain more conclusive results. I reflect on how my results compare with those obtained in previous studies and conclude with a future perspective. I stress the importance of using a broader approach by performing additional measures, testing executive functions as well as other modalities like neuroimaging in future studies.

(5)

2. Primary Progressive Aphasia

Primary Progressive Aphasia (PPA) is a relatively rare neurodegenerative disorder characterized by a progressive deterioration of specific language functions. Marked by an insidious onset, there is a gradual degeneration of neural networks governing production, comprehension, word finding, and object naming, in particular. Considering that a decline in cognitive processing domains4_{is relatively spared in the early}

stages of the disease, the language impairment should be viewed as an isolated problem. Moreover, whereas cognitive processing decline and behavioural changes5_{increase during the (later) progression of the illness,}

the language deficit remains the most prominent feature (Mesulam, 1982, 2001; Grossman & Ash, 2004; Gorno-Tempini et al., 2011). PPA is typically related to brain atrophy of the frontal and temporal regions of the left hemisphere and absence of strokes, head traumas, or other specific causes that might account for the language deficits (Mesulam, 2003; Grossman & Ash, 2004).

It should be noted that while these criteria are straightforward, the diagnosis of PPA is not (see Marshall et al., 2018 for more in-depth discussion). Like in many cases of aphasia, the linguistic impairment may have effects on the outcome since the diagnostic materials often rely on ‘‘linguistically mediated instructions’’ (Grossman & Ash, 2004, p. 4). Any results of performed neuropsychological testing in PPA should therefore be carefully interpreted (Amici et al., 2006). Moreover, this caution is often necessary to avoid possible misdiagnosis in view of other progressive disorders displaying certain language impairment symptoms as well, such as corticobasal degeneration or progressive supranuclear palsy (Battista et al., 2017).

Aside from this root diagnosis of PPA, a consensus has been established in further classifying the disorder into a semantic (svPPA), a nonfluent (nfvPPA) and a logopenic variant (lvPPA). Following the proposed clinical guidelines by Gorno-Tempini et al. (2011), the categorization of these subtypes is mainly based on clinical diagnosis, the underlying pathology6_{and brain atrophy that is assessed through}

neuroimaging. In trying to classify the variants based on clinical diagnosis, a series of main language domains should be considered. Gorno-Tempini et al. (2011) propose the evaluation of ‘‘speech production features, repetition, single-word and syntax, comprehension, confrontation naming, semantic knowledge and reading/spelling’’ to be able to correctly identify the language symptoms of a subtype (p. 1008). Notable is that various language batteries (e.g. Boston Naming test, CAT-NL, location learning test) can be used as long as they address these areas.

In addition to the speech and language abilities, specific pathology presents itself indicative of the type of PPA, namely its variants are often linked to degeneration of the frontotemporal lobar (FLTD) or pathologies within the large spectrum of Alzheimer’s disease (Gorno-Tempini et al., 2011; Marshall et al., 2018). However, numerous studies have pointed out that these are rather heterogeneous (e.g.

4_{As stated by Mesulam (2003, 2013) this entails faculties like episodic memory, executive functions, visuospatial skills,}

reasoning and comportment.

5_{For example, individuals with svPPA can experience changes including absent or misplaced sympathy, social}

disinhibition, exaggerated reactions to pain among others (see Marshall et al, 2018 for in-depth reading).

6_{This refers to the disease correlations and underlying mechanisms with which the disease, i.e. the PPA variant, is}

(6)

Tempini et al., 2011; Marshall et al., 2018; Battista et al., 2017), and that observations related to them only reflect ‘‘groupwide probabilities’’. This suggest only relative relation between the specific subtypes and pathology (Gorno-Tempini et al., 2011, p. 1007).

While the study by Gorno-Tempini et al. (2011) provides a compelling case for PPA categorization, there are cases where deficit patterns do not align with the proposed criteria and represent an atypical or mixed subtype. In a study reported by Harris et al. (2013), for example, some patients could not be classified, exposing problems with the categorization criteria. Similar reports emerge in an investigation by Vandenberghe (2016). In a meta-analysis study, 15-30% of the tested individuals could not be assigned to any of the three variants. He clarified that this is due to a lack of specific features that are necessary for a categorization, or because an individual has features of multiple subtypes, suggesting a mixed variant. Possible ways to improve this matter include the modification on imaging as well as more in-depth neuropathological testing. In the same manner, Battista et al. (2017) recommend the use of suitable neuropsychological and linguistically fine-grained tests with methods more suitable in order to evaluate a patient’s linguistic abilities. Moreover, PPA research advocate for the implementation of specific ad-hoc clinical tests for PPA to improve current categorization (Battista et al. 2017; Bisenius, Neumann & Schroeter, 2016).

While this certainly calls for deeper investigation, it exceeds the scope of the present thesis. This paper only takes the main three variants into consideration, whose characteristics are further specified in the sections below.

2.2 The semantic variant

The semantic variant is categorized by a progressive loss of semantic knowledge (Gorno-Tempini et al., 2011; Kertesz & Harciarck, 2014). According to Gorno-Tempini et al. (2011), individuals diagnosed with svPPA typically experience word finding difficulties and single word comprehension deficits. Even though word finding difficulties are common across all subtypes, the anomia in svPPA is severe. Faced with a loss of meaning of words, verbal communication becomes gradually empty and verbose (Marshall et al., 2018). Patients employ simplifications, circumlocutions, and increasingly use close-cased words (Mesulam, 2001; Wilson et al., 2010). Single word comprehension is gravely impaired as well, where comprehension of less familiar and low-frequency items is lost first. Patients are likely to use generalizations, omissions, circumlocutions or substitutions for atypical items (Vandenberghe, 2016). Additionally, semantic paraphasias are consistent as well, especially in object naming (Kertesz & Harciarck, 2014). Other features present in the subtype are the manifestations of surface dyslexia, an impairment in the ability to read irregular words, and dysgraphia, a similar disorder concerning writing. Irregular words are regularized in their pronunciation and spelling. This is due to the fact that patients rely of phonology rather than meaning (Gorno-Tempini et al., 2011; Kertesz & Harciarck, 2014; Harley, 2014). Patients retain relative speech production (grammar and motor speech), as well as repetition (Gorno-Tempini et al., 2011; Kertesz & Harciarck, 2014). In terms of neuroimaging, the subtype is associated with cerebral atrophy in the anterior

(7)

temporal lobes (Kertesz & Harciarck, 2014; Leyton & Hodges, 2014). Additionally, the semantic variant is linked to ubiquitin-positive, TDP43-positive pathology (Gorno-Tempini et al., 2011).

2.2 The nonfluent variant

In contrast to the semantic variant that is characterized by fluent speech, the nonfluent subtype is defined by a strenuous, effortful language production and evident signs of agrammatism (Gorno-Tempini et al., 2011; Leyton & Hodges, 2014). Patients demonstrate telegraphic speech by omitting grammatical elements like verbs, function words and inflectional morphemes (Hillis, Tussiash & Caramazza, 2002; Gorno-Tempini et al., 2011). Sentence comprehension is another affected component, where patients are unable to understand sentences due to their grammatical complexity. Phrases containing negation, passives and object relative clauses are examples of this (Leyton & Hodges, 2014; Peelle et al., 2007). In addition to deficits in syntax, nfvPPA is categorized by an impaired motor planning or apraxia of speech. Individuals exhibit motor problems resulting in slurring, as well as phonological errors including deletions or transpositions of sounds. Moreover, prosody is affected with individuals displaying abnormalities in loudness, pitch and intonational stress (Vandenberge, 2016; Gorno-Tempini et al., 2011; Marshall et al., 2018). Per contra, single word comprehension and object knowledge is relatively preserved. Oher than the existent word finding difficulties, no profound semantic deficits are present (Gorno-Tempini et al., 2011; Kertesz & Harciarck, 2014). The majority of patients with nfvPPA show atrophy in the interior frontal gyrus and the insula cortex of the dominant hemisphere (Marshall et al., 2018). Neuropathologically there are greater individual differences, although frontotemporal lobar degeneration with tau-positive inclusions pathology is often expressed (Gorno-Tempini et al., 2011).

2.3 The logopenic variant

The logopenic variant is dominated by word retrieval and sentence repetition deficits (Gorno-Tempini et al., 2011). Similar to nfvPPA, the fluency of speech is compromised. In lvPPA, speech output is slowed, consisting of word-finding pauses and hesitations (Gorno-Tempini et al., 2011; Vandenberghe, 2016). Considering these retrieval difficulties, patients show no obvious syntax deficits and have relative intact semantics and single word comprehension (Kertesz & Harciarck, 2014; Marshall et al., 2018). Impairments in confrontation naming7_{are less severe than in the semantic variant. Sentence comprehension is distorted}

however. A manifestation that is caused by the length rather than grammatical complexity of the phrase (Leyton & Hodges., 2014). With patients displaying short term phonological memory deficit, the repetition of longer phrases is also impaired (Vandenberghe, 2016). Individuals with lvPPA often present phonological paraphasias, including syllable mis-selections, either in spontaneous speech or in naming

7_{Rayner (2017) defines this as ‘‘a type of task used in assessment when problems with anomia or word retrieval are}

of concern. Confrontation naming involves the selection of a specific label corresponding to a viewed picture of an object or action’’ (p. 1).

(8)

(Gorno-Tempini et al., 2011; Marshall et al., 2018). Reading aloud is affected is a similar manner as well (Marshall et al., 2018). In regards to neuroanatomy, most cases of the logopenic variant reveal atrophy involving the temporo-parietal junction zone, as well as regions including the left parietal and temporal cortices, (Leyton & Hodges, 2014). The variant is moreover linked to an underlying pathology of Alzheimer’s Disease (Gorno-Temipini et al. 2011).

(9)

3. The present study

As previously mentioned in the introduction, the present study examines the naming ability of IwPPA and the role of executive interference control processes including attention and selective inhibition. Whereas the importance of executive control in naming abilities was shown in stroke patients (e.g. Piai & Knight, 2018), very little is known in this respect regarding PPA. In this section, I will present the various papers covering the naming deficits that are exhibited in PPA. Furthermore, I will discuss the use of the PWI as an experimental design in investigating these notions and how the task taps into the involved control processes. The main aims of this thesis are to (1) investigate the performance of picture naming of IwPPA and control participants using a PWI task (through response times and error analysis), (2) investigate the response times and errors between the three PPA variants, and (3) to measure how the performance on the PWI task compares to other cognitive and language tests. At the end of this section I offer my predictions on the results that I will obtain.

3.1 Naming abilities and executive functions in PPA

A common denominator in IwPPA is the difficulty in word finding/retrieval and naming (Race et al., 2013; Vandenberghe et al., 2005; Kirschner, 2014). As noted in chapter two, individuals with svPPA demonstrate this as a consequence of deterioration of conceptual knowledge. Van Scherpenberg et al. (2019) affirm that ‘‘semantic features constituting semantic representations of objects are progressively lost in people with svPPA and are therefore consistently unavailable during naming’’ (p. 13). Though Wilson et al. (2017) have shown that some patients also reflect difficulties in the post semantic stage of word retrieval8_{. With respect}

to the other non-semantic variants, Henry et al. (2013) suggest that in lvPPA lexical retrieval is impaired at a post-semantic or phonological stage. In this case the access to phonological representations is distorted. Though pointed out in an earlier study, in nfvPPA and lvPPA the deficit extends into the prior levels of lexical semantics as well (Rogalski et al., 2008).

Another point to raise in the (possible) effect of impaired executive functions on PPA. In the diagnosis of PPA it is made explicit and assumed that cognitive abilities i.e. executive functions, should not be deteriorated in the beginning stages of the disease. However, Macoir et al. (2017) indicate that a decline in these functions is present in all three variants of PPA with a magnitude that exceeds what is expected from the natural aging process and PPA diagnosis criteria.

3.1 The use of the PWI paradigm

The PWI paradigm is an on-line naming task where participants are presented with a set of pictures, each

8_{This would mean that the semantic stage is intact but subsequent stages reflect difficulties. These difficulties could}

(10)

accompanied by a written (or spoken9_{) distractor. They are usually instructed to name the picture whilst}

ignoring the distractor. The relation to the distractor to the picture can be modified in several ways. The distractor can be presented as semantically related or unrelated to the picture, promoting a semantic interference effect. Lexical interference can also be measured when distractors are presented as unrelated or neutral. Phonological facilitation is another possibility when distractors are phonologically related or unrelated to the picture. Additionally, presentation time of the distractor can be altered as well. They can be presented before (-300 ms), simultaneously (0 ms) or after the picture (+500 ms), also known as the stimulus onset asynchrony (SOA) (Collina, Tabossi & De Simone, 2013; Shao & Meyer, 2017). Critical determinants that are measured during this task, is the accuracy and the time it takes from the presentation of the picture to its naming (the naming latency). As presented by Harley (2014a), how the participants are instructed may influence the response. Telling the participants to react quickly makes them respond faster but at expense of making more errors. Conversely, stimulating the participants to be accurate makes them respond slower. Therefore, both error rates and response times are measured in PWI studies. The errors found in these tasks may say something about the naming abilities of participants. Impairment can be shown by a heightened pattern of hesitations or no responses, as well as phonological and semantic paraphasias (Hashimoto & Thompson, 2010).

The use of the PWI paradigm is an appropriate experimental method for investigating naming abilities of aphasic individuals. A study by Hashimoto and Thompson (2010) corroborates this claim. They analysed the use of the PWI in aphasia by studying the course of semantic and phonological activation needed for naming in aphasic individuals (n=11) and control participants (n=20). In their PWI task they utilized semantic, phonological and unrelated distractors presented at SOAs ranging from -300 ms, 0 ms and +300 ms. Results showed that significant semantic interference as well as phonological facilitation was present at 0 ms. In the case of semantic interference, the individuals with aphasia were significantly slower than the control group at SOAs of -300 and 0 ms, with the largest effect measured at 0 ms. The subsequent error analysis showed that more are errors present in the semantic related than in the neutral condition. The error of ‘no responses’ was the most frequent for the aphasic group. They indicate that the observed effects were caused by breakdowns in the phonological processing stage.

The claim that PWI can be useful for exanimating lexical selection and interference control is provided by studies reported by Piai, Riès and Wick (2016) and Piai and Knight (2018). The possibility that interference control has an influence on the naming difficulties in PPA as measured by PWI. For example, an investigation reported by Vandenberghe et al., (2005) analysed word finding difficulty and word selection in the semantic (n=6) and the nonfluent variant (n=5) of PPA. A primed picture naming paradigm was applied with the use of a semantically related, unrelated and a neutral nonword condition. Results demonstrated more errors and slower word retrieval for IwPPA when the prime was semantically related.

9_{A spoken distractor usually refers to an auditory PWI paradigm. The paradigm that is taken into consideration in}

(11)

Moreover, they proposed that during lexical retrieval, the characteristics10_{of PPA interfere with the ability}

to select and to discriminate semantically related words. Moreover, a noteworthy study by Thompson et al., (2012) used a PWI task to study the time course of object naming between individuals with nfvPPA (n=8) and lvPPA (n=13) in comparison to controls (n=17). They utilized related and unrelated stimuli in order to test overall semantic interference. Results showed slower reaction times for all groups in the related condition in comparison to the unrelated condition over SOAs of -500, -100 and 0ms. However, only the PPA groups presented effects at a SOA of -1000 ms, implying a greater magnitude of semantic interference. They suggested that this magnitude is due to heightened activation of semantic networks as well as higher vulnerability to the interference from distractors. They further deduced that this is caused by semantic mapping deficits marked by difficulty in rapidly dissociating the distractor from the target. The study highlights that besides phonological encoding deficits, defective semantic processing is accounted for also in the naming deficits in these subtypes. Important to note is that the presence of these deficits due to poor cognitive functions is ruled out. Patients performed well on the cognitive tests and results showed no significant correlations.

3.3 Predictions

In accordance to the previous studies on PPA using PWI (Vandenberghe et al., 2005; Thompson et al., 2012), I expect to find that the IwPPA exhibit slower response times and more frequent naming errors in comparison to the control participants. I predict that more difficulty will arise in the related compared to the neutral condition. With respect to the error analysis, based on the general difficulty in word-finding, I expect hesitations to be present in all PPA variants (Kirschner, 2014). In addition, I expect that the semantic variant demonstrates more semantic paraphasias and the logopenic variant displays a heightened frequency of phonological paraphasias. Moreover, as certain variants only have few to only one patient, and the high diversity within each subtype, I expect that no sufficient statistical power will be reached to categorize the different PPA variants.

(12)

4. Methods and materials

4.1 Participants

Twelve individuals with PPA (6 svPPA, 5 lvPPA and 1 nfvPPA) and thirteen healthy controls, matched on age (p = 0.60), sex (p = 1) and education (p = 0.495), participated in the study11_{. They were recruited from}

various hospitals in the Netherlands through geriatricians, neurologist and/or neurophysiologists. Patients were clinically diagnosed with PPA. The diagnosis, based on a multi-disciplinary evaluation, was established according to official clinical criteria (Gorno-Tempini, 2011). Patients did not have any neurological diseases other that PPA and, like the control participants, did not present any physical or sensory deficits that could interfere with their performance in the study. The controls were acquired through the personal contacts of my supervisors at the Radboud University Nijmegen. All participants were native speakers of Dutch. The study has been approved by the ethical committee and written consent was obtained from all the participants. Only controls received monetary compensation. Specific demographic data on the participants is illustrated in Table 1 below.

Table 1. Demographic information of patients and controls

Participant

Age at testing

(in years) Sex (Verhage, 1964) Education level classification PPA symptoms Years of

P1 73 M 4 lvPPA 2.5 P2 64 M 4 lvPPA 1.0 P3 55 M 6 lvPPA 2.0 P4 75 M 4 svPPA 2.5 P5 68 F 6 svPPA NA P6 63 F 6 svPPA 4.0 P7 74 F 6 svPPA NA P8 70 F 5 svPPA 2.0 P9 71 M 4 lvPPA 2.0 P10 80 M 6 lvPPA NA P11 67 M 4 nfvPPA NA P12 66 F 4 svPPA 5.0 PPA mean 68.8 ± 6.3 7M/5F Control mean 68 ± 3.2 8M/5F

4.2 PWI task

4.2.1 Stimuli

Stimuli included twelve colour pictures of living objects that the participants were asked to name. These items were selected from two different semantic categories (either an animal or a fruit) with six objects per

11_{All participants are part of the Language and Progressive Aphasia (LAPA) study (N. Janssen, personal}

(13)

category. The pictures were considered for inclusion because of their easy and frequent use in daily life. This was done in order to minimize the response times caused by word finding problems, since the focus was on the degree of inhibition rather than the naming speed (N. Janssen, personal communication, 7 January, 2019). However, I cannot discard the possibility that frequency or word length effected the outcomes of the results. Two different distractor types were used for each picture measuring an overall interference effect12_{(Starreveld & La Heij, 2017). The distractor was either a word from the same semantic}

category, the related condition, or a row of X’s (XXX), the neutral condition. All the distractor words belonged to the response set, meaning that for the image of an apple there was also a distractor word “apple”. The task was divided into six blocks, with each block containing twenty items, 120 items in total. In the test session the participants saw the experimental item each with a neutral and related distractor (five times per condition). Every trial was randomized per participant.

4.2.2 Procedure

The PWI task was conducted in a sound-proof cubicle where participants were seated in front of a 24-inch BenQ (XL2420Z) monitor and a Stage Line DM-5000LM microphone. Before the start of the experiment, participants were shown a sheet of paper containing the images that would appear in the task. After reviewing each one and highlighting the target names, which were provided below each image, participants underwent a practice session. The familiarization phase comprised of twelve practice items in total. Instructions appeared on screen and were offered orally as well. Participants were asked to name the images presented in the centre of a white screen while ignoring a distractor that was superimposed on every image (SOA= 0 ms). Each participant was encouraged to be as fast and accurate as possible. As mentioned previously, the task was divided into six blocks. After each block, the participant had the option to take a small break if needed. To begin each block, the phrase ''ATTENTION! The experiment is beginning'' was presented on the screen for 1500 ms. After this, the experimental item with a distractor appeared for 4000 ms in the middle of the screen. A white screen was shown for 1000 ms before the start of the next trial.

It should be noted that the control group was tested by me and that the IwPPA were tested by one of my internship supervisors. The tests were conducted in a similar environment. A comparable monitor, microphone and similar settings for the audio recording were used. All patients underwent an MRI13_scan

first, and after a small break proceeded to partake in the PWI task and other neuropsychological tests. The MRI task was administered first, followed by the PWI task and neuropsychological research. Likewise, the control participants underwent an MRI scan. In 3 control cases the MRI and the PWI task were performed on the same day (as was done for PPA patients), the scan was carried out first followed after a short break by the PWI task. In the other 10 control cases the MRI was done on a different day as the PWI task.

12_{Note that the interference effect is neither of semantic or lexical nature. This is discussed in further detail in the}

Discussion.

(14)

4.2.3 Response time analysis

The naming response times (RT) were all manually determined using Praat (Boersma & Weenink, 2018). A spectrogram generated by Praat was used to mark the first sound of each word made by the participant. Only the beginning sound, and not the duration or end, was taken into consideration. This was done prior to error coding and separating trials by condition. Responses containing errors or speech dysfluency were coded as incorrect and their corresponding trials were excluded from the RT analysis14_{(Piai, Riès & Swick,}

2016; Piai & Knight, 2017; Thompson et al., 2012). The following responses were coded as errors: (1) an existing word—but not the target—was named. (2) the distractor word was named. (3) response contained hesitations (e.g. these entailed responses that began with stutters and filled pauses). (4) phonological paraphasias. (5) a semantically related response (e.g. instead of responding with dog, responding with Labrador). (6) RT is unreliable (e.g. the previous trial could be heard, participant was talking to the examiner or coughed). (7) response contained a phrase or a determiner in addition to the target (e.g. participants would start with ‘that is a [mouse]’). (8) no response is given. Trials containing slight variations, such as consistent mispronunciation of a word, the use of the diminutive, modification after the target (e.g., cow … calf), and a late response in the next trial, provided no distractor word was named, were coded as correct. In accordance to Piai and Knight (2017), the RTs were log-transformed first to reduce skewedness. Next, they were analysed with a linear mixed-effects model in R (R Core Team, 2019). There were fixed effects for group (control vs. IwPPA) and condition (neutral vs. related), as well as their interaction. Random intercepts were included for both participants and item. Condition was added as a random slope for participant and group for item.

4.2.4 Error analysis

The coding of errors was identical to the one applied in the RT analysis. In this case, all correct responses were excluded, with the addition of responses coded with (6) RT is unreliable and (7) response contained a phrase or a determiner in addition to the target. The responses coded as 6 and 7 were excluded from the RT analysis due to their influence on the response time but did not contain any errors.

The errors were analysed with a mixed-effects logistic regression in R (R Core Team, 2019). Similar to the formula utilized for the RT analysis, there were fixed effects for group (control vs. IwPPA) and condition (neutral vs. related), as well as their interaction. Random intercepts were included for both participants and item. Only condition was added as a random slope for participant, on account of more complex models failing to converge.

4.3 Additional cognitive and language tests

Additional tests were administered in order to examine the cognitive and language abilities of the participants and to check how these correlate (and act as predictors) to their performance of the PWI task.

(15)

The tests included: (1) The Sydney Language Battery (Sydbat-NL, Savage et al., 2013, Dutch version modified by Eikelboom et al., 2017), designed to help clinicians differentiate the PPA subtypes. It consists of a naming, repetition, comprehension and semantic association task. (2) the Montreal Cognitive Assessment (MoCa), designed to be a screening tool for mild cognitive impairments. There is an assessment made on domains including ‘‘memory, language, executive functions, visuospatial skills, calculation, abstraction, attention, concentration, and orientation’’ (Julayanont et al., 2013, p. 111). (3) the Digit Span (of Wechsler Adult Intelligence Scale-IV), consisting of a forward, backward and sequencing15_component

is aimed at measuring auditory recall, short term memory and working memory (Coalson et al., 2010). (4) the Trail Making Test (TMT), consisting of part A and B, is used to assess domains of speed, visual scanning and executive function (attention and switching) (Llinàs-Reglà et al., 2017).

4.3.1 Procedure

With respect to the healthy controls, the administration of the additional tests occurred after the PWI task in the same audio booth. The investigator was seated in front of the participants and proceeded to administer the tests in a continuing manner. This was done following the specific instructions provided by the test. No specific order of tests was adhered, only if a certain test required sections to follow a specific order16_{. Due to the nature of PPA, the patients were burdened as little as possible. This meant that relevant}

test results from the hospital were used as much as possible. If certain tests were not available or if the neuropsychological research from the hospital had been conducted more than a half year ago, the tests were taken again on the same testing day as the PWI task17_.

15_{Although the Digit Span consist of these three subtests, the results obtained reflect the total score. It would have}

been more insightful if each section was analysed.

16_{For the SYDBAT-NL the order of the test administration is important. The subtest ‘Naming’ must always be}

administered first. The subtest ‘Repetition’ may then be administered 2nd_{, 3}rd_{, or as last. However, the subtest}

'Comprehension' must always be administered before the subtest 'Semantic association’.

(16)

Control PPA

5. Results

The PWI task results are first reported. Figure 1 and Table 2 show individual-average as well as group-average response times and naming accuracy. Notably, the response times for the IwPPA are longer in comparison to the control participants, with relatively more variation present as well. It is evident that that that data points of the control participants are grouped tightly together, but that the data points of IwPPA demonstrate a high degree of heterogeneity, as seen in Figure 1. As expected, the naming accuracy of the healthy controls is much higher (nearing ceiling level) than the IwPPA. Again, there is extensive variation present.

Figure 1. Individual averages of response times and naming accuracy for the two groups

across conditions in PWI task

Table 2. Group averages of response times in ms and naming accuracy (and their standard deviation).

Condition RT in ms Naming accuracy

Control Related 896 (116) 97.4 (2.4) Neutral 759 (94) 99.6 (0.7) PPA Related 1367 (457) 81.5 (18.2) Neutral 1141 (384) 92.1 (7.4)

5.1 Response times

With regard to the response times seen in Table 3, the IwPPA were significantly slower in the naming of pictures in comparison to the control participants by an average estimated difference of 153 ms (p=<0.001). A significant overall interference effect was observed as well. Response times were higher in the related condition in comparison to the neutral condition by an average estimated difference of 77 ms (p=<0.001). Conversely, no inference can be drawn as to whether the IwPPA are slower than the healthy controls over

0 500 1000 1500 2000 2500 3000 Related Neutral Re sp on se Time s (ms ) 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Related Neutral Na ming a cc ur ac y (% )

(17)

the two conditions. There was no significant effect found between these two notions (p=0.227). The same holds true for this comparison between the different PPA subtypes and controls. Only a main effect of group and condition were thus found.

Table 3. Inferential statistics for response times (top) and naming accuracy (below).

Response times b SE t (df) p

Controls vs. PPA 0.153 0.036 4.217 (25) <0.001

Related vs. neutral 0.077 0.006 12.066 (24) <0.001

Related vs. neutral : controls Related vs. neutral : PPA

Related vs. neutral : controls vs. PPA 0.016 0.013 1.240 (23) 0.227 Related vs. neutral : controls vs. svPPA 0.017 0.018 0.977 (18) 0.341 Related vs. neutral : controls vs. lvPPA 0.014 0.012 1.206 (18) 0.243 Related vs. neutral : controls vs. nfvPPA 0.023 0.025 0.909 (28) 0.371

Error rate B SE z p

Controls vs. PPA 3.089 0.724 4.265 <0.001

Related vs. neutral 1.867 0.671 2.782 0.005

Related vs. neutral : controls Related vs. neutral : PPA

Related vs. neutral : controls vs. PPA -0.892 0.688 -1.296 0.195

Related vs. neutral : controls vs. svPPA -1.117 0.751 -1.487 0.137

Related vs. neutral : controls vs. lvPPA -0.592 0.722 -0.820 0.412

Related vs. neutral : controls vs. nfvPPA -1.496 0.915 -1.635 0.102

5.2 Naming accuracy

The error rate results in Table 3 show that the control participants, in contrast to the IwPPA, score significantly better on the task by an estimated log difference of 3.1 (p=<0.001). According to this, the calculated odds ratio is 22.0, meaning that the control participants scored 22.0 times better than the IwPPA. A significant interference effect was again observed. Participants were significantly less accurate in the related condition than the neutral condition by an estimated log difference of 1.9 (p=0.005). The calculated odds ratio is 6.5, meaning that in the related condition 6.5 times more errors were produced than in the neutral condition. Once more, no significant effect in the interaction between group and condition was found (p=0.195), nor between the control participants and PPA subtypes. Only a main effect of group and condition were present.

The error distribution is shown in Figure 2. Based on the graphs of the two groups, a substantial difference in the frequency of errors is noticeable. The IwPPA produce more errors both in the related (n=133) and the neutral condition (n=58). The difference is observed in the number of distractors named, hesitations and no responses given. Considering the small number of PPA participants, a more detailed account of the produced errors per variant and participant can be provided. In Figure 3 the error

(18)

PPA Control

distributions of the logopenic variant are shown. It can be observed that most errors stem from P3 (n=41) and P10 (n=46) in their own variant and the PPA group in total. It is noteworthy that phonological paraphasias are not on the forefront of most made errors. The semantic variant, shown in Figure 4, displays a more varying image. Apart from a higher concentration of no responses from P4, no clear outliers are present. Figure 6 presents the errors made by P11, the only participant in the nonfluent variant, with hesitations as the most frequent made error.

Figure 2. Error distribution for the two groups across the related and the neutral condition

Figure 3. Error distribution for the logopenic variant, shown by participant. nt=not target, dis=distractor, hes=hesitation,

phon=phonological paraphasia, sem=semantically related response, nr=no response 0 5 10 15 20 25 30 35 40 45 50 F re qu en cy o f er ro rs Related Neutral 0 5 10 15 20 25 30 35 40 45 50 Related Neutral 0 4 8 12 16 20 nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr P1 P2 P3 P9 P10 lvPPA Related Neutral

(19)

Figure 4. Error distribution for the semantic variant, shown by participant.

Figure 5. Error distribution for the semantic variant, shown by participant.

5.3 Additional cognitive and language tests

Table 4 presents the group-averaged scores and p-values (Mann-Whitney U test) on the various cognitive and linguistic tests. The results show a significant difference between the two groups on almost all tests. Only the repetition and comprehension components of the SydBat-NL rendered non-significant results. Though it should be clarified that there were missing test values from the IwPPA on some parts of the tests, which has an influence of the overall significance.

Figure 6 reveals the correlations between the cognitive tests (MoCa, Digit-span, TMT A and TMT B) and the RT and error rate. The plot shows various coloured circles with the Pearson’s correlation coefficient presented in the middle. For the sake of this analysis only the first two rows are of interest. It can be seen that error rate is negatively correlated with the TMT B, MoCa, Naming and Repetition of the SydBat-NL, and the Digit Span. RT is negatively correlated with all tests, but especially with the MoCa. This implies that the RT and the numbers of errors increases when the score on the test decreases.

0 4 8 12 nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr nt _dis _hes ph on _sem nr P4 P5 P6 P7 P8 P12 svPPA Related Neutral 0 4 8

nt dis hes phon sem nr

P11

nfvPPA

(20)

Figure 6. Correlations between the error rates, response times, and the additional tests

Table 4. Group averages scores on the cog. And linguistic tests (and their standard deviation).

Cog. and

linguistic tests Controls Scores Scores PPA p

MoCa 27 (2.3) 22 (3.2) <0.001 TMT A 51 (5.9) 39 (13.3) 0.008 TMT B 51 (7.5) 41 10.9) 0.042 Digit Span 11 (2.2) 6 (3.6) 0.003 SB-NL Nam 27 (1.4) 18 (4.6) <0.001 SB-NL Rep 29 (0.5) 28 (2.2) 0.093 SB-NL Com 29 (0.9) 28 (1.6) 0.125 SB-NL Asso 28 (2.5) 24 (4.8) 0.003

(21)

6. Discussion

The objectives of the thesis where to investigate the performance in picture naming of IwPPA and controls participants using a PWI task and to investigate thethree PPA variants more closely. In my study the IwPPA performed significantly slower and showed less accuracy in naming compared to the controls. In addition, a significant interference effect was observed. The related condition proved to be more problematic than the neutral condition in both the RTs and accuracy analysis. The results related to group difference are in agreement with the findings obtained from the studies of Vandenberghe et al. (2005) and Thompson et al. (2012). The significant interference effect in my study and the aforementioned studies, however, cannot be fully compared. Whereas my study employed a combined interference effect of semanticity and lexicality, the other studies investigated the role of semantic interference. The combined effect was so designed to maximize the chance of finding a difference in effect between the IwPPA and healthy controls (N. Janssen, personal communication, 24 June 2019). In terms of interpretation it cannot be said whether the effect is lexical or semantic. The breakdown in naming and support to the claims on naming deficits as shown in chapter three, is therefore challenging to uncover. No time course was measured by virtue of various SOAs (see chapter three). This makes it complicated to uncover at which stage (e. g. semantic representation, post semantic stage or phonological representations) difficulty occurs.

The additional Sydbat-NL test showed significant differences in the naming and semantic association subtests between IwPPA and controls whereas comprehension and repetition did not. Relative intact comprehension, but slower RTs, could imply deficits in lexical access. IwPPA can understand the word, implying relative intact semantic knowledge, but not access it for naming. But again, it is difficult to determine at what stage this may occur. The vital analysis of the three variants and their characteristic features are also missing because of a low number of data points. The comparison between them did not converge in the statistical measures taken in R due to low participant number. My results do not support the results obtained by others (e.g. van Scherpenberg et al., 2019; Wilson et al., 2017; Henry et al., 2013) regarding claims of naming deficits.

Still, the notion of interference control of attention and inhibition was measured through the distractor in the PWI. Though in the same manner, to discover the precise role in the slower RTs of IwPPA is complicated. The limitation of only measuring at 0 ms makes it difficult to infer if a heightened activation and difficulty in lexical selection is present. This was described by Thompson et al., 2012 where activity was present at a SOA of -1000 ms. In any case, the fact that faulty executive functions (Macoir et al., 2017) may have an influence on lexical selection should be further investigated. There were significant differences on the cognitive tests between the IwPPA and controls. Moreover, these tests proved to be predictors in the performance of the PWI task.

A clear distinction of the two groups can be made with respect to the error distribution shown in Figure 2. As predicted, there were high rates of hesitation errors in PPA patients. While not predicted, frequent distractor errors and no response errors were obtained in IwPPA. In the subsequent figures, variance among the patients can be observed. Participants P3 and P10 (lvPPA) and P4 (svPPA) contributed most. Due to

(22)

the small number of participants per PPA type, it cannot be directly said that these individuals can be treated as outliers. Moreover, this may also simply be indicative of the heterogeneity of PPA and also the heterogeneity between the subtypes. The diagnosis and stratification of subtypes based on the errors produced by the PWI was not possible. Characteristic errors of the variants such as semantic paraphasias in svPPA and phonological paraphasias in lvPPA did not appear frequently. In part this can be explained by that the specific test was not sensitive enough to elicit those type of errors.

6.1 Limitations

There are a number of limitations to this study. The most vital, mentioned throughout the paper, is the small number of participants present in the investigation. This is especially holds true for the nonfluent PPA variant, where there was only one individual tested. As mentioned before, this undermines the statistical power of my study, which increases the chance of type І error. Moreover, the lack of SOAs and other interference effects (semantic or lexical) in this investigation limit its scope. It would have been interesting to explore the process of various stimuli onsets (see Thompson et al., 2012) in addition to the SOA of 0 ms used in this study. Likewise, addition of stimuli testing semantic interference or phonological facilitation would allow for a more detailed account of lexical access breakdown.

The assumption that the cognitive tests utilized here truly measure the executive functions necessary for lexical selection is not guaranteed. The low scores obtained could also reflect the language impairment in IwPPA, which was already discussed by Grossman & Ash (2004, see introduction).

The RT analysis is another possible drawback of this study. The manual calculation of the RT data leaves room for human error and inconsistencies. Though I tried to be as consistent as possible, it would have been beneficial if there were specific indicators for word sound onset or to have a second coder check for inter-rater reliability. The error distribution analysis was conducted by using descriptive graphs, speculating about patterns found rather than performing statistical analysis to analyse the differences. Further analysis was not possible due to a limited number of datapoints.

Furthermore, as previously reported, the testing of the IwPPA was performed by my supervisor, whereas I tested the controls, which could lead to interobserver variability. Also, some control participants underwent different testing schedules of MRI and PWI task. This may also lead to different results, e.g. if participants make more errors if they have been subjected just before with MRI testing. To minimize the error rate the PWI task was mainly recorded through computerized operating.

Lastly, the fact that a control participant had to be re-tested due to an error in the recording is also far from desirable. Though it is important to note that the time between test and re-testing consisted of a period of 46 days. Nevertheless, the PWI effect assumes the same underlying processes as the Stroop effect, which adheres to extremely robust findings (Maanen, van Rijn & Borst, 2009; Starreveld & La Heij, 2017).

(23)

6.2 Future perspective

The PWI method has proven useful in distinguishing between IwPPA and control participants, but the classification of patients into the three PPA variants on the basis of PWI was not possible. This is because PPA is a very heterogeneous disease and progression path from initial diagnosis during aging is likely to be different for each patient. The time between onset of the disease and initial disease diagnosis varies for each patient. Moreover, loss of executive function will be confounding factor. PWI testing needs to be integrated with other tests, including imaging and cognitive tests, explicitly testing executive functions like selective attention and inhibition. This will allow for early PPA diagnosis, make accurate predictions on how the disease will progress and allow for the development of new therapeutic modalities as the intervention can be better assessed. Altogether, this will result in better care and optimal treatment for each individual PPA patient.

(24)

7. Conclusion

This thesis investigated the naming abilities of IwPPA with the use of a PWI task. The study tried to explore the breakdown in naming in relation to interference control processes evoked through nature of the task. While the results showed that IwPPA are significantly slower and less accurate than healthy control participants, profound conclusions on the breakdown of naming cannot be made. This applies to the role of interference control as well, which needs to be investigated more deeply. The differences in the PPA variants could not be established, due to the low number of participants per each subtype. Nevertheless, the present study should be seen as a pilot study as many pitfalls in the experimental design became apparent. Important is also that the PWI and linguistic tasks should be integrated with other tests on cognitive functions and neuroimaging to allow for better diagnosis and results that can be used to guide for better treatment of IwPPA.

(25)

References

Battista, P., Miozzo, A., Piccininni M., Catricalà, E., Capozzo, R., Tortelli, R., Padovani, A., Cappa, S. F., & Logroscino, G. (2017) Primary progressive aphasia: a review of neuropsychological tests for the

assessment of speech and language disorders, Aphasiology, 31(12),

1359-1378, http://doi.org/10.1080/02687038.2017.1378799

Bisenius, S., Neumann, J., & Schroeter, M. L. (2016). Validating new diagnostic imaging criteria for primary progressive aphasia via anatomical likelihood estimation meta-Analyses. European Journal of Neurology,

23(4), 704–712. http://doi.org/10.1111/ene.12902

Boersma, P., & Weenink, D. (2019). Praat: Doing phonetics by computer (Version 6.0.420. Retrieved from http://www.praat.org/.

Coalson, D. L., Raiford, S. E., Saklofske, D. H., & Weiss, L. G. (2010). Advances in the assessment of intelligence. In Weiss, L. G., Saklofske, D. H., Coalson, D. L., Raiford, S. E. (Eds.), WAIS-IV clinical

use and interpretation: Scientist–practitioner perspectives (pp. 3-23). San Diego, CA: Academic Press.

Collina, S., Tabossi, P., & De Simone, F. (2013). Word Production and the Picture-Word Interference Paradigm: The Role of Learning. Journal of Psycholinguistic Research, 42(5), 461-473. https://doi.org/10.1007/s10936-012-9229-z

Dell, G. S. (1986). A spreading-activation theory of retrieval in sentence production. Psychological Review,

93(3), 283-321. http://doi.org/10.1037/0033-295X.93.3.283

Diamond A. (2013). Executive functions. Annual review of psychology, 64, 135–168. http://doi.org/10.1146/annurev-psych-113011-143750

Eikelboom, W.S., Janssen, N., van den Berg, e., Roelofs, A., & Kessels, R. P. C. (2017). Differentiatie van primair-progressieve afasie aarianten: de nederlandse bewerking van de sydney language battery (SYDBAT-NL). Tijdschrift voor Neuropsychologie, 12(3), 189–202.

Garrett, M. F. (1975). The analysis of sentence production. In G. H. Bower (ed.), The Psychology of Learning

and Motivation (vol. 9., pp. 133-177). San Diego, Calif.: Academic Press.

Glaser, W. R., & Düngelhoff, F.-J. (1984). The time course of picture word interference. Journal of

Experimental Psychology: Human Perception and Performance, 10(5), 640–654.

Gorno-Tempini, M. L., Hillis, A. E., Weintraub, S., Kertesz, A., Mendez, M., Cappa, S. F., et al. (2011). Classification of primary progressive aphasia and its variants. Neurology, 76(11), 1006–1014. http://doi.org/10.1212/WNL.0b013e31821103e6

Grossman, M., & Ash, S. (2004). Primary progressive aphasia: a review. Neurocase, 10(1), 3-18, http://doi.org/10.1080/13554790490960440

Harris, M., Gall, C., Thompson, JC., Richardson, A. M., Neary, D., du Plessis, D., Pal, P., Mann, D. M., Snowden, J. S., & Jones, M., (2013). Classification and pathology of primary progressive aphasia. Neurology, 81(21), 1832–1839. http://doi.org/10.1212/01.wnl.0000436070.28137.7b’

Harley, T. A. (2014a). Recognizing visual words. In The psychology of language: from data to theory (4th. ed., pp. 167-208) Hove: Psychology Press.

(26)

Harley, T. A. (2014b). Language production. In The psychology of language: from data to theory (4th. ed., pp. 395-448) Hove: Psychology Press.

Hashimoto, N., & Thompson, C. K. (2010). The use of the picture–word interference paradigm to examine

naming abilities in aphasic individuals. Aphasiology, 24(5),

580-611. http://doi.org/10.1080/02687030902777567

Henry, M. L., Rising, K., DeMarco, A. T., Miller, B. L., Gorno-Tempini, M. L., & Beeson, P. M. (2013). Examining the value of lexical retrieval treatment in primary progressive aphasia: two positive cases. Brain and language, 127(2), 145–156. http://doi.org/10.1016/j.bandl.2013.05.018

Hillis, A. E., Tuffiash, E., & Caramazza, A. (2002). Modality-specific deterioration in naming verbs in nonfluent primary progressive aphasia. Journal of Cognitive Neuroscience, 14(7), 1099–1108. https://doi,org/10.1162/089892902320474544

Julayanont P., Phillips N., Chertkow H., Nasreddine Z.S. (2013). Montreal Cognitive Assessment (MoCA): Concept and Clinical Review. In: Larner A. (Ed.), Cognitive Screening Instruments: a practical approach (pp. 111-152). Springer, London.

Kertesz, A., & Harciarek, M. (2014). Primary progressive aphasia. Scandinavian Journal of Psychology, 55(3), 191–201. https://doi.org/10.1111/sjop.12105

Kirshner H. S. (2014). Frontotemporal dementia and primary progressive aphasia, a review. Neuropsychiatric

disease and treatment, 10, 1045-1055. https://doi.org/10.2147/NDT.S38821

Levelt, W. J. M. (2001). Spoken word production: a theory of lexical access. Proceedings of the National Academy

of Sciences, 98(23), 13464-13471. https://doi.org/10.1073/pnas.231459498

Leyton, C., & Hodges, J. R. (2014). Differential diagnosis of primary progressive aphasia variants using the international criteria. Aphasiology, 28(8-9). http://doi.org/10.1080/02687038.2013.869306.

Llinàs-Reglà, J., Vilalta-Franch, J., López-Pousa, S., Calvó-Perxas, L., Torrents Rodas, D., & Garre-Olmo, J. (2017). The Trail Making Test: association with other neuropsychological measures and normative values for adults aged 55 years and older from a spanish-speaking population-Based Sample. Assessment, 24(2), 183–196. https://doi.org/10.1177/1073191115602552

Macoir, J., Lavoie, M., Laforce, R., Brambati, S., & Wilson, M. (2017). Dysexecutive symptoms in primary progressive aphasia: beyond diagnostic criteria. Journal of Geriatric Psychiatry and Neurology, 30(2) https://doi.org/10.1177/0891988717700507.

Marshall, C.R., Hardy, C.J.D., Volkmer, A. et al. (2018). Primary progressive aphasia: a clinical approach. J

Neurol, 265(6), 1474–1490. https://doi.org/10.1007/s00415-018-8762-6

Mesulam, M. M. (2001). Primary progressive aphasia. Annals of Neurology, 49(4), 425–432. https://doi.org/10.1002/ana.91

Mesulam, M. M. (1982). Slowly progressive aphasia without generalized dementia. Annals of Neurology,

11(6), 592-593. https://doi.org/10.1002/ana.410110607

Mesulam, M. M. (2003). Primary Progressive Aphasia — A Language-Based Dementia. New England

(27)

Mesulam M. (2013). Primary progressive aphasia: A dementia of the language network. Dementia &

neuropsychologia, 7(1), 2–9. https://doi.org/10.1590/S1980-57642013DN70100002

Peelle, J. E., Cooke, A., Moore, P., Vesely, L., & Grossman, M. (2007). Syntactic and thematic components of sentence processing in progressive nonfluent aphasia and nonaphasic frontotemporal dementia.

Journal of Neurolinguistics, 20(6), 482–494. https://doi.org/10.1016/j.jneuroling.2007.04.002

Piai, V., Roelofs, A., & Schriefers, H. (2012). Distractor strength and selective attention in picture-naming performance. Memory & Cognition, 40(4), 614–627. https://doi.org/10.3758/s13421-011-0171-3 Piai, V., & Knight, R. T. (2018). Lexical selection with competing distractors: Evidence from left temporal

lobe lesions. Psychonomic Bulletin & Review, 25(2), 710-717. http://doi.org/10.3758/s13423-017-1301-0 Piai, V., Riès, S. K., & Swick, D. (2016). Lesions to lateral prefrontal cortex impair interference control in word production. Frontiers in Human Neuroscience, 9, 721. http://doi.org/10.3389/fnhum.2015.00721 Race, D., Tsapkini, K., Crinion, J., Newhart, M., Davis, C., Gomez, Y., Hillis, A. E., & Faria, A. (2013). An

area essential for linking word meanings to word forms: Evidence from primary progressive aphasia. Brain and Language, 127(2), 167–176. https://doi.org/10.1016/j.bandl.2013.09.004

Raymer A. (2017). Confrontation Naming. In: Kreutzer J., DeLuca J., Caplan B. (Eds.) Encyclopedia of Clinical

Neuropsychology. Springer, Cham. https://doi.org/10.1007/978-3-319-56782-2_875-4

Rogalski, E., Rademaker, A., Mesulam, M., & Weintraub, S. (2008). Covert Processing of Words and Pictures in Nonsemantic Variants of Primary Progressive Aphasia. Alzheimer Disease & Associated

Disorders, 22(4), 343–351. https://doi.org/10.1097/WAD.0b013e31816c92f7

R Core Team. (2019). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.

Savage, S., Hsieh, S., Leslie, F., Foxe, D., Piguet, O., & Hodges, J. R. (2013). Distinguishing subtypes in primary progressive aphasia: application of the sydney language battery. Dement Geriatr Cogn Disord,

35(3-4), 208-218. http://doi.org/10.1159/000346389

Schiller, N., & Verdonschot, R. (2015). Accessing Words from the Mental Lexicon. In John R. Taylor (Ed.),

The Oxford Handbook of the Word (pp. 481-492). Oxford University Press.

http://doi.org/10.1093/oxfordhb/9780199641604.001.0001

Shao, Z., Meyer, A. S., & Roelofs, A. (2013). Selective and nonselective inhibition of competitors in picture naming. Memory & Cognition, 41(8), 1200-1211. http://doi.org/10.3758/s13421-013-0332-7

Shao, Z., & Meyer, A. S. (2018). Word priming and interference paradigms. In A. M. B. de Groot & P. Hagoort (Eds.), Guides to research methods in language and linguistics. Research methods in psycholinguistics and the

neurobiology of language: A practical guide (pp. 111-129). Wiley-Blackwell.

Spezzano, L. C., & Radanovic, M. (2010). Naming abilities: differentiation between objects and verbs in aphasia. Dementia & neuropsychologia, 4(4), 287–292. http://doi.org/10.1590/S1980-57642010DN40400006

(28)

Starreveld, P. A., & La Heij, W. (2017). Picture-word interference is a Stroop effect: A theoretical analysis and new empirical findings. Psychonomic bulletin & review, 24(3), 721–733. http://doi.org/10.3758/s13423-016-1167-6

Thomas, S & Rao, S.L. & Devi, B. I. (2016). Relationship between attention and inhibition: are they two sides of the same coin? Journal of the Indian Academy of Applied Psychology, 42(2), 337-343.

Thompson, C. K., Cho, S., Price, C., Wieneke, C., Bonakdarpour, B., Rogalski, E., ... Mesulam, M. M. (2012). Semantic interference during object naming in agrammatic and logopenic primary progressive aphasia (PPA). Brain and Language, 120(3), 237–250. https://doi.org/10.1016/j.bandl.2011.11.003 Vandenberghe, R., Vandenbulcke, M., Weintraub, S., Johnson, N., Porke, K., Thompson, C., & Mesulam,

M. (2005). Paradoxical features of word finding difficulty in primary progressive aphasia. Annals of

Neurology, 57(2), 204-209. https://doi.org/10.1002/ana.20362

Vandenberghe R. (2016). Classification of the primary progressive aphasias: principles and review of progress since 2011. Alzheimer's research & therapy, 8(16), 1-9. https://doi.org/10.1186/s13195-016-0185-y

van Maanen, L., van Rijn, H., & Borst, J. P. (2009). Stroop and picture-word interference are two sides of the same coin. Psychonomic Bulletin & Review, 16(6), 987-999. https://doi.org/10.3758/PBR.16.6.987 van Scherpenberg, C., Fieder, N., Savage, S., Nickels, L. (2018). The relationship between response

consistency in picture naming and storage impairment in people with semantic variant primary progressive aphasia. Neuropsychology, 33(1), https://doi.org/13-34. 10.1037/neu0000485.

Verhage, F. (1964). Intelligentie en leeftijd: onderzoek bij nederlanders van twaalf tot zevenenzeventig jaar. Assen: Van Gorcum.

Wilson, S., Dehollain, C., Ferrieux, S., Christensen, L., & Teichmann, M. (2017). Lexical access in semantic variant PPA: Evidence for a post-semantic contribution to naming deficits. Neuropsychologia, 106, 90– 99. https://doi.org/10.1016/j.neuropsychologia.2017.08.032

Wilson, S. M., Henry, M. L., Besbris, M., Ogar, J. M., Dronkers, N. F., Jarrold, W., … Gorno-Tempini, M. L. (2010). Connected speech production in three variants of primary progressive aphasia. Brain: a journal

(29)

ACKNOWLEDGEMENTS

The process of finishing the thesis and finally completing my master program has been quite the challenge. Still, above all, it has been a great learning experience for which I owe many people my sincerest gratitude. First, I wish to thank my thesis supervisor dr. Nada Vasic. I am very grateful for the feedback and insights you provided.

I would also like to thank my supervisors at the Donders Institute, dr. Vitória Piai and Nikki Janssen for giving me the opportunity to learn and be a part of valuable research. Thank you so much for guiding me throughout the internship period and making the experience immensely worthwhile.

Special thanks to the all of those who were willing to participate and bared with my initial testing sessions. Although I did not have direct contact with the patients in my study, I extend this gratitude and appreciation.

Likewise, I would like thank the patients and personnel of the speech therapy department at the Radboud UMC, who were so helpful and allowed me to observe and obtain great understanding into the workings of speech or language impairment assessment.

Finally, I express my deepest gratefulness to my family and friends who have supported me throughout my university years, and this thesis especially. Thank you for your constant encouragement18_.

Primary Progressive Aphasia: A study of picture naming, interference and lexical access