1
Analyzing pauses in people with Parkinson’s disease. A linguistic study
about the use and timing of pauses
Patrick Drent
LTS998M20.2019-2020.2 Student number: 3862852 University of Groningen Supervisor: A. Rofes
2
Index
Abstract ... 3
1. Introduction ... 4
1.1 Parkinson’s disease and pausing in speech ... 4
1.2 Word categories and pauses in spontaneous speech ...10
1.3 This study...12
2. Method ...14
2.1 Participants ...14
2.2 Materials and analyses ...16
2.3 Spontaneous speech measurements ...17
2.4 Pauses ...17
2.5 Word categories ...18
2.6 Inclusion and exclusion criteria ...18
2.7 Ethics approval ...18 2.8 Statistical analyses ...18 3. Results ...19 3.1 Pause frequency ...19 3.2 Word categories ...20 3.3 Pause time ...22 4. Discussion ...25
4.1 Introduction and expectations ...25
4.2 Pause frequency ...26 4.3 Word categories ...27 4.4 Pause time ...30 4.5 Side notes ...31 5. Conclusion ...33 6. References...35
3
Abstract
Aim of the study
This study aims to explore the role of pauses in people with Parkinson’s disease. More specifically, we are interested in linguistic patterns specifically for people with Parkinson’s disease in comparison to a healthy group of participants.
Introduction
Parkinson’s disease could impair spontaneous speech, like the use of pauses. Studies about Parkinson’s disease and spontaneous speech are rare, contradictory and focused mostly on nouns and verbs.
Method
Seven older people with Parkinson’s disease (of which three also have dementia) participated in this study. We also have a group of healthy controls out of the literature. The first 300 words of spontaneous speech from our participants were used to study. Pauses were marked and the details about these pauses were noted, like the word categories before they used these pauses.
Results
Pauses were found before eight different word categories. People with Parkinson’s disease tended to have most pauses before conjunctions, adverbs or verbs. On a more individual level, it seems that some people tend to peak at a specific linguistic variable. Compared to the healthy group of people, overall, less nouns and pauses were used in spontaneous speech.
Conclusion & Discussion
Our participants used overall less pauses than the healthy group of people. Most pauses were used before adverbs. We think this might be due to less speech used overall in comparison to the control group, and certain outliers due to the relatively small sample size of seven people with Parkinson’s disease. Further research seems necessary.
4
Introduction
1.1 Parkinson’s disease, general symptons and pausing in speech
Parkinson’s disease is a progressive neurodegenerative disorder characterized by the progressive loss of dopaminergic neurons, and to be more exact, primarily in the Substantia Nigra (Harrison, 2013). The Substantia Nigra is a group of neurons which is part of the basal ganglia and is important for the creation of dopamine in the brain. Dopamine is a
neurotransmitter which is important to move muscles. When someone has Parkinson’s disease, the cells which produce dopamine are slowly dying which results in a shortage of this
neurotransmitter. The Substantia Nigra is located in the middle part of the brain and belongs to the basal nuclei of the central nervous system. The result of the loss of dopamine will eventually result in certain impairments. Some of the most well-known impairments in Parkinson’s disease are impairments in moving the body, at the start barely noticeable like shaking with hands (tremor) and later more noticeable impairments like stiffness or slow movement (bradykinesia). Parkinson’s disease cannot be cured but there are medications which significantly improve symptoms like the medicine Lepodova which is a dopamine precursor and makes everyday life a bit easier.
Parkinson’s disease is, right after Alzheimer’s disease, the most common neurodegenerative disorder, and the prevalence in the Netherlands is about 0,3% for the whole population, and 1% for people which are older as 60 years (Kalf & van Zundert, 2017). The prediction is that in 20 years the population with Parkinson’s disease will double (Kalf & van Zundert, 2017). For the German population, the risk of Parkinson’s disease is around 2,0 % for men and 1,3 for woman (Deutsche Gesellschaft für Parkinson und Bewegungsstörungen, 2018). Around 400,000 people in Germany currently have Parkinson’s disease according to German health insurance data, and around 7000 people die every year because of Parkinson’s disease. Finally, the German Society of Neurology or GSN assumes a crude prevalence (in the category 65+population) of
1800/100,000 persons. Also, cognitive impairment is a frequent complication in PD, at the time of diagnosis, about one third of the people will also have mild cognitive impairments, and over a typical disease duration, up to 80% op people with Parkinson’s disease will develop a kind of dementia (Williams-Gray et al., 2013). One of the symptoms of Parkinson’s disease which also influences daily life is the changes in speech. Like tremor in hands, this symptom might not be immediately noticeable because the earliest symptoms might involve a softer and slower form of speech. Most people can, however, make themselves understandable. Later however, might people with Parkinson’s disease have more problems in making themselves understandable to
5 their direct surroundings like partners. With that in mind, Ramig et al., (2008) describe that only 3-4% of the people with PD receive proper speech treatment. Nonetheless, people with PD can have serious language impairments such as speaking with a soft voice, a monotone voice, breathiness (which hampers intelligibility), a hoarse voice quality and lessened facial expression (Pitcairn, Clemie, Gray, & Pentland, 1990) which makes communicating much harder.
Communication is essential for people in everyday life, and impairments in communication might result in depression and emotional changes, something which is common in a lot of people with Parkinson’s disease (Timmer, Tendolkar, & Esselink, 2014; Mayo Clinic, 2020). This is mostly because people with Parkinson’s disease are having problems which influences quality of life.
Importantly, another type of language issue that is observed in people with Parkinson’s disease are pauses in speech. Pauses are a natural component of speech rate. Indeed, speech rate itself is measured by the amount of speech produced per unit and is typically expressed as syllables per second or words per minute (Tjaden & Wilding, 2011). When speaking about pause time, we refer to the accumulation of silence over the course of a given speech sample. A related concept to this is pause frequency which refers to the number of pauses in a speech sample, so basically the number of silences used in speech. (Tjaden & Wilding, 2011). In a study by Ho et al. (1998), 200 people with Parkinson were examined, and 66% of those people had deviations in their voice production, and 29% had deviations in speech tempo (think of fluency). In the study by Ho et al. (1998), it is argued that the pause patterns can be due to the ‘stuttering’ like speech, which is due to starting problems when trying to speak, which can lead to freezing in speech. Interestingly, people with Parkinson tend to have wordfinding difficulties, due to cognitive decline (Smith, Ash, Xie, & Grossman, 2018). This also can lead to slower speech, and eventually more pauses being used.
There are several impairments which could lead to difficulties in fluency of speech like delayed transmission of information due to loss of connection and impaired cognitive abilities like working memory and inhibitory planning or motor skills. Performance of working memory is generally intact (like for example when doing a digit repetition task), but people with
Parkinson’s disease will perform more poorly as the disease increases in severity. Eventually, people will also have difficulties even when cues (think of verbal help) are provided. Delayed transmission of information through cortico-stiatio-pallido-thalamocortical circuits due to loss of connection resulting from Parkinson’s disease related lesions could interfere in the smooth flow of information between language areas, which on its turn will result in impaired fluency of speech because the language system has to wait for the next sentence elements to become available in the brain (Braak & Braak, 2000). Lewis, Lapointe, Murdoch and Chenery (1998) pointed out that language production might rely on cognitive abilities like working memory, planning, initiation and inhibitory planning which leaves language production very vulnerable in Parkinson’s disease. Visually, people with Parkinson’s disease are generally impaired which also is due to cognitive dysfunction (Beatty & Monson, 1989). The result of impaired cognitive abilities can for example be seen in an early article of Illes, Metter, Hanson and Iritani (1988).
6 They found that the speech of the persons with Parkinson’s disease was disfluent, marked particularly by more pauses and more pauses per word than speech of the healthy control group. They argue that syntactic language changes associated with Parkinson’s disease might reflect an adaptation to increasing dysarthria.
In another study after this one, Iles (1989) examined analyzed responses of open-ended
autobiographical questions from 10 individuals with Parkinson’s disease, as well as people with Huntington’s disease (10), Alzheimer’s disease (10) and another healthy group of older adults (10). Overall, the group with Parkinson’s disease did not differ from the healthy adult group on measures like locations of pauses, number of fluency disruptions et cetera. However, the people with Parkinson’s disease seemed to include more pauses in their responses than the control group, and also with perceptible silent hesitations at the initiation of sentences and between main clauses and optional clauses. They described that in both Parkinson’s disease, as in Huntington’s disease, the people had periods of freezing or rigidity followed by a limited burst of action or in this case, a burst of speech. Because of this, Illes (1989) concluded that the disfluencies within and between clauses utterances in people with Parkinson’s disease are a result of difficulties in activating concepts and formulating sentences, as well as to adaptation to the motor impairments in Parkinson’s disease.
In the earlier quoted article of Lewis et al. (1998), the language production of 20 people with Parkinson’s disease (of which half also had dementia) and 20 healthy age-matched controls was tested. They found that the people with more severe dementia in comparison to people with just Parkinson’s disease and healthy controls had increased impairments in defining words and generating sentences. They attributed that this was a result of cognitive impairments associated with frontal lobe dysfunction and this conclusion seems to be validated given the fact that they based it on the face validity of the tasks used. The language tasks on which they scored more poorly were that those required organization, abstract thought, integration of information and planning. These functions are all related to the frontal lobe. Finally, Jacobs et al. (1995) argue that verbal fluency which is impaired in people with Parkinson’s disease, is a sign of executive dysfunction rather than a primary impairment of language and may be a harbinger of the more severe forms of cognitive dysfunction.
When speaking more specific about pauses, people need breath pauses in their speech and are necessary in everyday spoken language. Speech is a function, which needs balance in producing speech, and metabolic demands of oxygen/carbon dioxide exchange (Hoit, Lansing, & Perona, 2007). All these factors involve deciding when to pause to inhale. Pauses are used to mark the ends of prosodic phrases, changing intonation and syllable duration (Steinhauer, 2003). In the everyday spoken language, listeners use pauses along with other prosodic cues to parse semantic and syntactic units in their running speech. Some studies describe that pausing is more important to understanding syntactical complex sentences than is pitch contour (Shah, Baum, & Dwivedi, 2006).
7 In an article from 2012, it is described that there are not many studies regarding syntax and breath pausing in older adults and in people with speech disorders like Parkinson’s disease (Huber, Darling, Francis, & Zhang, 2012). What we do known is that typical aging and PD could also result in some changes in the cognition, which on their turn can lead to problems in planning breath pauses together with language formulation (Zgaljardic et al., 2006).
Furthermore, in another article from Stine et al., (1995) showed that during silent reading tasks, older adults tend to pause for less time at the end of sentences, and pause more and longer at minor syntactic boundaries when comparing with younger adults. In people with PD, the impact of cognitive factors have not been well studied, however, some of these cognitive impairments in people with PD are similar to typically aging people although the impairments are more severe in Parkinson’s (Huber et al., 2012). This is why Huber et al. (2012) conclude that it is highly likely that cognitive impairments in PD also play a role in language impairments. This also counts for impairments in working memory, reduced inhibitory mechanics and difficulties in set switching could likely play a role in impaired language impairments. Because syntactic
processing seems to conclude in subtle language processing difficulties, people with PD have more difficulties in planning breath pauses with syntax (Grossman, Lee, Morris, Stern & Hurtig, 2002). As described by Weismer (2006), pauses are highly important for the listeners ability to use syntactic predictions for lexical access and lexical segmentation, when occurring at
unexpected or atypical locations.
Huber et al. (2012) studied people with PD and looked more specifically at breath pauses. They hypothesized that people with PD would take a smaller percentage of breath pauses at major boundaries. However, there were no differences found in comparison with healthy controls for breath pauses taken at major or minor boundaries. They did, however, not investigate/rapport in which situation these breath pauses were taken, except if they were taken in the previous spoken major and minor boundaries. This is interesting and should be examined more. What they did found, was that older adults pauses more frequently at minor syntactic boundaries and less frequently at major ones compared to young adults. This result was being viewed as a result of respiratory rather than cognitive changes. Also, although no differences were found for breath pauses taken at major or minor boundaries between people with Parkinson’s disease and healthy older adults, it was found that individuals with Parkinson’s disease produced more pauses at ‘atypical’ boundaries than healthy older adults. However, this result came out of an oral reading task, a task which does not involve spontaneous speech production.
Pauses can be made in all kinds of places. For people with Parkinson’s disease, it seems that literature speaks more about so-called ‘atypical’ places in which pauses take place, these are places like in the phrase itself, or in a word (Lee et al., 2019). It was found that 24-60-year-old and healthy adults produced increased atypical pauses in an extemporaneous language production task in comparison to tasks like oral reading or written passage tasks. In typical aging, both respiratory-psychological and cognitive-linguistic changes occur. Due to reduced recoil pressure, decreased muscle mass, reduced vital capacity and decreased lung volume may cause older adults to pause more frequently. Also, due to a reduction in cognitive processes like
8 inhibitory control and working memory, older adults may experience difficulties in producing and comprehending complex sentences, like in spontaneous speech (Caplan & Waters, 1999). For people with Parkinson’s disease, it is well known that declines in higher-level cognitive skills like working memory and inhibitory control are very common at a relatively early stage of the disease (Troche & Altmann, 2011). At this day, there are not many studies regarding production of pauses in relation to language formulation in people with Parkinson’s disease. Some things are known. In the study of Bunton (2005), the effects of typical aging in Parkinson’s disease were investigated and to be more specific, the lung volume use in spontaneous speech. Results indicated that people with Parkinson’s disease paused less frequently at syntactic boundaries (50-71% of times) as healthy older adults (75-87% of times). In the study of Lee et al. (2019), older adults tend to pause more at atypical linguistic locations than young adults (like in the sentence: Her /pause/ eyes were bright). Lee et al. (2019) considered out of these results that an increased production of pauses at atypical locations might be part of normal aging. Also, it was found interestingly enough, that people with Parkinson’s disease did not differ from age- and education matched healthy older adults in terms of the distribution of silent pauses. They did not produce more atypical breath pauses compared to age-matched healthy adults, actually, a lot of factors like location and frequency of pauses were similar between these two groups. This result is considered to be rather interesting given the fact that in the study of Huber et al. (2012), people with Parkinson’s disease produced more pauses at atypical locations, something which was not entirely in line with the result of Lee et al. (2019).
In general, pauses in spontaneous speech is a highly variable phenomenon for all people. Pausing is a highly fragmented and discontinuous activity. It is more likely and longer before words with less predictability and that pauses can be used to sort out which parts of certain verbal sequences are verbal habits and which are being created at the time of speaking (Kendall, 2013). But how about the pause pattern when speaking about a healthy group of people
without Parkinson’s disease. Studies about these kinds of patterns seems divided and somewhat rare given the vague definition of the term ‘pause’. Some of the earlier studies quote that people with Parkinson’s disease produce more pauses than healthy controls (Hammen, Yorkston & Breukelman, 1989; Torp & Hammen, 2000), while some other studies like the one from
Skodda and Schlegel (2008) stated that people with Parkinson’s disease produced fewer pauses in comparison to healthy controls. Also, one older article from Metter and Hanson (1986) found large individual differences between both people with Parkinson’s disease and healthy controls. As can be seen, the results are inconsistent and differ quite a lot. Experts say that this might be due to small sample sizes and range of disease severity. We do know some general facts about normal aging. Penny et al. (1996) found no differences in speech rate for people between the age of 65 and 90 years of age. However, with regard to pauses, several studies do show that an increase in the frequency and duration of silent pauses. This could be due to cognitive decline which is normal when aging. This also is a logical consequence when looking at word finding difficulties which increase when aging and which are reflected in pausing (Schwartz & Frazier, 2005; Zellner-Keller, 2007). Then again, there are also studies which show the opposite. In an
9 older study of Kynette and Kemper (1986), no significant changes were found in the production of fillers or incomplete sentences between a healthy group of people ages between 50 and 90 years old. It is also known that older healthy adults take more pause time in phrases and sentences in comparison to a younger group of healthy adults (Little, Prentice, Darrow & Wingfield, 2005), so while typically aging, pauses will take more time.
In the article of Igras-Cybulska, Ziółko, Żelasko, and Witkowski (2016) some more specific information can be found. In this study, a group of Polish experienced speakers, and
inexperienced speakers (which were labeled as students) were examined on their speech and to be more exact, their spontaneous speech and forms of presentations and other
non-spontaneous variants. Pauses were divided in silent pauses (in which no voice activity is
recorded when pausing), filled pauses (in which pseudo-words were used as pauses like ‘hmm’ or ‘eee’ which are normal pseudo-words for pausing in Polish) and breath pauses. This study showed that professionals spoke more slowly, used less disfluencies and formulated shorter sentences in comparison to the group of students. Their dynamic breathing rhythms were much more concordant with sentence boundaries. Overall, the use of breath pauses was much higher for experiences speakers than for students. Which was also interesting was the fact that filled pauses were the least used pauses for experienced speakers but used the most by students. Breath pauses and silent pauses were used almost the same for students, while silent pauses were used on average 26% of the times, where breath pauses are used in 50% of the times. Unfortunately, it is not described how many words were used in this study samples.
In another study from Redford (2013), the spontaneous speech of both children and one of their parents (both without problems in language, typical developing children) was examined with the use of a storytelling task. The story is known as the frog story from Mercer Mayer, in which there are only pictures and no real text. The child and the parent needed to tell a story with the pictures provided in the book. Both the children and the parents needed to tell the story twice. Ten parent-child dyads were recruited. The results showed that children produced on average 23.34 pauses per minute in the first storytelling, and 26.18 per minute during the second. For their parent, the results were a little closer to each other with 21.92 pauses per minute in the first, and 21.63 per minute the second time. Pause durations differed too. Children used on average 1416 (first) and 1094 (second) milliseconds while their parent used 993 and 901
milliseconds. Although these results are more interesting, the fact is that these results might not be connected to the term ‘spontaneous’ speech, due to the fact that the participants in this study were pushed to certain words and directions (like the pictures in the frog story), although the participants were free to interpret and tell the story. Again, the study did not mention how many words were used in total, but the pauses per minute makes it a little clearer for
comparison in this study. Another interesting study of Angelopouloua et al. (2018) examined people with aphasia and compared them to 19 healthy adults. Healthy adults were asked to provide a brief narration of the patient’s history, so for example the stroke incident itself and how it affected their lives (the healthy group consisted of the caretakers of the people with aphasia). Unfortunately, it was not noted how many words were used per individual, or how big
10 the speech sample was. Well were some more general noted. In the healthy group, the mean number of silent pauses used for all healthy participants was around 20 pauses. It ranged from a minimal of 11 pauses up till a maximum of 38 pauses. This was done in a mean amount of seven utterances with a mean length of utterances of 15,92. This means that utterances were quite long (around 16 words used per utterance and 7 utterances used). Since the mean length of utterance stands for the number of words or morphemes used in each of the utterances, there are some raw conclusions which can be speculated when interpreting the results. A raw
calculation teaches us that this means that around 112 words per sample were used, in which around 20 pauses were used for the healthy group. So, for every 5,6 words/morphemes used in spontaneous speech, 1 pause takes place, but this is a very raw conclusion and will differ
between participants.
In summary, pausing in speech is highly variable phenomenon for all people. People with Parkinson’s disease tend to have problems in their speech in relation with pausing, due to factors which are related to Parkinson’s disease like disfluency of speech and cognitive/motor factors. Literature about people with Parkinson’s disease and their pause pattern seems divided, with some articles claiming that people with Parkinson’s disease produce more pauses in
comparison to a control group (Hammen et al., 1989; Torp & Hammen, 2000) while others claim the opposite (Skodda and Schlegel, 2008). This might be due to the vague definition of the term ‘pause’.
1.2 Word categories and pauses in spontaneous speech
But when talking about people with Parkinson’s disease and pauses they make, where can these pauses exactly be found? To begin, it is important to understand the nature of lexical
impairments in people with Parkinson’s disease. For what is known, different classes of words can be selectively impaired in people with different types and sites of cerebral damage. To be more exact, this includes words which belong to different grammatical classes like nouns and verbs. It is known from several lesion studies that lesions in the left temporal lobe may induce deficits in the processing of nouns (Martin & Chao, 2001). For verb processing, this is a bit more complex, as problems with verbs can occur due to a much wider range of deficits, including the left frontal and inferior parietal lobe, and also in subcortical structures like the basal ganglia (Daniele, Giustolisi, Silveri, Colosimo, & Gainotti, 1994; Silveri & Ciccarelli, 2007).
People with Parkinson’s disease are known to have problems in producing verbs, and to be more exact, the use of action related words or verbs (referring to movements of parts of the arm/leg/face like the words ‘kick’ or ‘lick’) (Boulenger et al., 2008). Other studies like the ones of Rodriguez et al. (2009) and, Herrera, Rodriguez-Ferreiro and Cuetos (2010) seem to agree with Boulenger et al. (2008). The production of action related verbs has been associated to neural activity in cell assemblies which include areas in the cortical regions involved in motion planning and execution (Hauk, Johnsrude & Pulvermüller, 2004). In a study of Cotelli et al. (2007), the capacity of 32 people with Parkinson’s disease to name 120 drawings (60 objects and 60 drawings of actions) was tested. The results showed significant impairments in the action
11 naming task, and the authors concluded that this research could support the idea that the motor system is involved in lexical-semantic processing of motion related words. Furthermore, in the same study, the results indicated that people with Parkinson’s disease not only showed impairments in naming action related events, but also in object naming, compared to healthy controls. It must be noted that people with Parkinson’s disease were significantly more impaired in naming the action related verbs, than in object naming. These results can indicate the
following. The fact that verb retrieval is severely more impaired in PD could be consequential due to dopamine depletion in the striatum, which disrupts the function of subcortical prefrontal networks. The left prefrontal and parietal areas are important in processing verbs (Daniele et al., 1994). Another study found similar results for verbs (verb-specific responses) in both the left prefrontal and parietal areas when using event-related MRI, while noun-specific activations involved activation in the inferior temporal lobe (Shapiro, Moo & Caramazza, 2006). Some other studies pointed out that people with Parkinson disease have problems with verbs due to
extensive executive resource demands associated with verbs (Auriacombe et al., 1993; Grossman et al., 1994). Parkinson’s disease and Parkinson related disorders are as stated not associated with significant disease burden in the pyramidal system that includes
motor-associated cortices. Furthermore, regression analyses did not relate action verb performance to cortical atrophy (progressive disruption of the cerebral cortex).
Interestingly enough, the article of York et al. (2014) seems not to be in line with the study of Boulenger et al. (2008) and relative studies. In this study, 22 people with Parkinson’s disease took part in a two-alternative forced-choice associativity judgment task. In this task, 120 frequency matched action verbs, cognition verbs, concrete nouns and abstract nouns were probed, this was also done with a group of people with ALS (amyotrophic lateral sclerosis). Their results were being compared to a healthy control group of 13 healthy seniors. Results showed that people with Parkinson’s disease were not significantly disturbed in action verb associativity judgments, while people with ALS were. The authors mention that this was partly due to
degraded representation of action knowledge feature in motor-associated cortical regions. In this study, the group of people with Parkinson’s disease had more problems in the concrete noun category, something which the opposite with the results of Boulenger (2008). York et al. (2014) stated that this was due to a well-known ‘concreteness effect’ (Paivio, 1991). Also, they stated that earlier work (Boulenger et al. 2008; Herrera et al. 2010; Rodriguez et al. 2009) did not evaluate the so-called verbs of cognition to eventually determine whether the action verb deficit is related to the executive resource demands or grammatical demands associated with verb classes. Another interesting theory which they state is that receiving dopaminergic
supplementation might also interfere in results regarding action verbs (Herrera & Cuetos, 2012), although this does not seem to have interfered in the study of York et al. (2014). Again, like the pause frequency, the literature seems to be divided on this matter and more exploration to certain cases seem to be necessary.
While there are some articles regarding the spontaneous speech of people with Parkinson’s disease in relation to nouns and (action-related) verbs, there are not that many articles
12 regarding other word categories which might also shine more information in the speech pattern of people with Parkinson’s disease. Like in the recent study of Smith et al. (2018), the title of the study suggest linguistic markers in people with Parkinson’s disease, unfortunately like in many articles, this study seems to be more related to nouns and verbs specifically, leaving out word categories like pronouns, adverbs (Cann, 1993; Frey, 2003; Frey & Pittner, 1998; Huddleston & Pullum, 2002) and adjectives for a start. Also, results between most study’s seems to be the same, with people with Parkinson’s disease seem to pause more frequently before verbs in comparison to healthy age-matched controls (Smith et al., 2018). This is rather strange given the fact that nouns and verbs are only a few factors in linguistics, but always seem to be the most used word categories in analyzing spontaneous speech in people with Parkinson’s disease. So while we already know that people with Parkinson’s disease probably pause more frequently before the use of verbs, certainly action-related ones in comparison to nouns, the literature does not state too much about the use of pronouns, adverbs, adjectives and other word categories. One of the articles which does investigate some other linguistic variables is the one from Holtgraves, McNamara, Cappaert and Durso (2010). In this study, the spontaneous speech in a group of people with Parkinson’s disease was examined. The occurrence of verbs, function words (which are all other linguistic variables like adverbs and prepositions), and sentence length were investigated. They found some interesting things, like that greater left-side motor severity was associated with the production of significant fewer verbs, function words and shorter sentences. Furthermore, they would produce fewer complex utterances. Unfortunately, it remains unclear which function words in particular were the most affected. This article
specifies everything as function words and keeps it that way in the conclusion and discussion. We have to think about the earlier described studies of Lee et al. (2019) and Huber et al. (2019) to think about where pauses might take place. These studies do agree that both healthy older adults and people with Parkinson’s disease use pauses at atypical places like in the middle of sentences or in words itself. We might expect due to that, that pauses will take place in/at words which connect main- and subordinate clauses like conjunctions which are highly important in connecting phrases or clauses with one another. Due to the relative absence of studies regarding conjunctions and Parkinson’s disease we cannot say with certainty that this will be the case for our study, but we will keep this in mind.
1.3 This study
Researchers argue that the impairments that people with Parkinson’s disease have in sentence comprehension might be related to a specific grammatical or syntactic deficit (Lieberman et al., 1992; Cohen et al., 1994). People with Parkinson’s disease displayed longer response times and more errors in the comprehension of certain syntactic structures, and certainly in the more complex syntactic structures. Furthermore, Altmann and Troche (2011) discussed the
interruptions of fluency which can take many forms. Some of the forms include longer and more pauses between or in the middle of sentences, these problems can be attributed to difficulties
13 in the early stages of language production like formulation and conceptualization. Because of this, we hypotheses that both the production of syntactic longer and more complicated sentences and because of interruptions of fluency will eventually result in more problems for people with Parkinson’s disease, and more uses of pauses throughout their spontaneous speech. We hypothesis that this will also happen in this study, and because longer sentences might be more difficult, we think that people with Parkinson’s disease will have more pauses in the syntactic important variable conjunctions, given the fact that they connect utterances and are important in the buildup for sentences.
Aims and hypothesis of this study
The aim of this study is to examine more about silent gaps (pauses) in people with Parkinson’s disease and to compare these results with a healthy group of people. Studies about these topics are rare and the literature seems to be inconsistent between people with Parkinson’s disease and healthy adults. This study will examine the pause pattern of people with Parkinson’s disease, this will be done in their spontaneous speech. There are not many articles regarding this kind of research, and the ones that exist are inconsistent and sometimes vague. This study tends to be more exact and tries to be as specific as possible. Our research will be centered around nouns and verbs, given the fact that people with Parkinson’s disease certainly have problems with both word categories and most literature builds around people with Parkinson’s disease and nouns and verbs. While using this literature regarding nouns and verbs in analyzing people with Parkinson’s disease, this same literature can be used in analyzing other word categories like pronouns, adverbs, adjectives et cetera. The term pause still seems to be too vague, given the fact that most studies use different definitions, making it difficult to interpret certain results. This study has its own specific definition of a ‘pause’ and will use this in the whole research. Given the fact that literature is so divided when comparing between pauses in people with Parkinson’s disease and healthy controls, it is also difficult to predict how many pauses (pause frequency) in spontaneous speech are normal for both groups. We will use the article of Redford (2013) for that, because this is the most specific one found in the literature. But because of the divided literature, we will not make a specific prediction which group will use more pauses in spontaneous speech.
Furthermore, because we will look specifically at word categories, we predict and expect that some word categories will have more pauses in front of them. We think of the word category conjunctions, because of their importance in connecting and building role in formulating sentences/phrases, something which is more difficult for people with Parkinson’s disease (and also for older people which are healthy) who tend to include longer and more pauses in the middle of sentences (Altmann & Troche, 2011).
14 1. People with Parkinson’s disease will use more pauses in their spontaneous speech as the healthy group of people. Because most of the literature found seems to result in more uses of pauses in speech of people with Parkinson’s disease. This is our main hypothesis over this study;
2. People with Parkinson’s disease will have more and longer pauses before (action related) verbs than nouns in their spontaneous speech. Boulenger (2008) states that people with
Parkinson’s disease have more frequent pauses when producing verbs in comparison to healthy age-matched participants;
3. People with Parkinson’s disease will have more problems in the word category ‘conjunctions’. We expect due to the literature and their placement in sentences and utterances to find a high number of pauses in front of conjunctions.
Method
Participants
For this research, seven older German speaking adults (Gender = two females and five males) with Parkinson’s disease participate. From these seven people, three of them also have dementia (PD and D=3) (labeled as participant 5, 6 & 7). On the time of screening, all adults ranged in age (range = 78-84 years). The data of these participants comes from Dementiabank, which is a shared database of multimedia interactions for the study of communication in dementia, or Parkinson’s disease in this case (MacWhinney, 2007). All participants have
Parkinson’s disease and are German native speakers. They were all screened by F. Jalvingh, who is an academic speech therapist who also works at the University of Groningen. Dementiabank is freely available for everyone to use, but the permission to use the data for this research comes from the University of Groningen. Because of the current situation regarding the
Coronavirus, I as examiner did not have the chance to screen the participants used in this study in person. In Table 1, more of the details of the participants can be found.
15 Table 1
Demographic characteristics of participants
Participant 1 Participant 2 Participant 3 Participant 4
Participant 5 Participant 6 Participant 7 Age (when screened) 81 78 79 84 84 82 81 Diagnosis Parkinson’s disease Parkinson’s disease Parkinson’s disease Parkinson’s disease Parkinson’s disease and dementia Parkinson’s disease and dementia Parkinson’s disease and dementia
Gender Male Female Male Male Female Male Male
Native
language German German German German German German German
Control group
Furthermore, we have two groups of healthy control participants which were used in the
studies of Redford (2013) and Angelopouloua et al. (2018). We used two control groups to see if the results would match both between our group of participants and the controls groups, or if results would differ and why this is possible. Literature seems divided on pause frequency between people with Parkinson’s disease and control groups so two groups might give a better picture if this was true. Both groups of participants were screened for their spontaneous speech and their pause patterns were noted in the results of both studies. The first group from the study of Redford (2013) contains ten healthy controls (n= 10, gender= one male/nine female) which used a spontaneous speech story telling task to their children. All participants were native speakers of American English. The other study contained 19 healthy adults (n=19, range= 45-86 years of age, gender= ten male/nine female) who used spontaneous speech to tell something about people with aphasia and how their life changed due to their strokes. All participants were native Greek speakers and right-handed.
Both groups will be used to compare pause frequency results in the results of this study. We used the electronic database SmartCat to find our healthy participants for this study. The
keywords we used to find these participants were: spontaneous speech, Parkinson’s disease and pausing and we used advanced searching to better match what we wanted to search.
16
Materials and analyses
For this research, we used the audio recordings provided by the University of Groningen. The data we used is password protected and cannot be reached without permission. Video recordings were not available. The length differed quite between the participants (range = 03:26-11:45). See also Table 2. Audio recordings were completed in one try. Conversations topics were all the same for the participants. In the conversations, the people with Parkinson’s disease spoke about their life with Parkinson’s disease. Only participants and the researcher were present during these conversations, except for the conversation with participant six which also included the partner of the participant.
Table 2
Time audio-recordings and general information of participants
Audio sample Total length (minutes) Present during audio
recording Participant 1 03:26 Ra + Pb Participant 2 09:44 R + P Participant 3 05:22 R + P Participant 4 06:06 R + P Participant 5 10:31 R + P Participant 6 11:45 R, P and PAc Participant 7 09:29 R + P
Note. a Researcher. b Participant. c Partner of participant.
The audio recordings were analyzed by listening at least four times at all recordings and noting every (possible) pause. After this, pauses were noted in which cases they were present (before which word category) and time of the pauses, see also paragraph 2.4 and 2.5. Words which were not familiar were translated. For measuring the time of the different pauses, a stopwatch was used twice, in which the mean time of those two measurements was used. Samples were analyzed in Microsoft Word. After noting pause frequency and pause times the results were put into tables. After that all different word categories were labeled and put into tables. Afterwards the amount of different word categories was noted. When all the data was collected, we looked at more specific details. We split verbs in two categories, which will be elaborated in the word category paragraph. We also separated general mean times (which are mean pause times over all pauses used in samples made by a participant) from individual word category times (which are mean pause times used per word category). After all our data was analyzed we matched it with studies which were available to compare it with the rare literature available, like the control groups we used.
17
Spontaneous speech measurements
The participant told as much as possible, while the examiner sometimes has to ask some extra questions to keep them talking, for example when they need help in certain situations like eating or when they dress. The first 300 words that participants use in their sample were used. The literature states that 300 words is the commonly used measure (Bastiaanse & Jonkers 1998; Prins & Bastiaanse 2004). After marking the 300 words, all pauses will be noted. See the next paragraph for the definition of pauses, and what is considered to be a pause in this research. Verbs were split in action and non-action ones (see the word categories paragraph for the definition of both kinds of verbs).
The control group of Redford (2013) investigated spontaneous speech pause frequency by using a storytelling task without text. The book used contained only pictures (the Frog story of Mercer Meyer). With the pictures provided parents needed to tell a story they made up at that
moment, so the subject was mostly related to the story, however, the parents were free to tell whatever they wanted. The story was told twice so none of the stories was exactly the same. This was done to measure the effect of cognitive load and/or familiarity on pause frequency behavior.
The control group of Angelopouloua et al. (2018) used spontaneous speech to tell about their relatives which suffered a stroke which resulted in aphasia (a brief narration).
Pauses
This study uses the following terminology in the search for pauses in spontaneous speech: ‘A pause is a short period like in speech in which something such as a sound is stopped before starting again (Cambridge Dictionary, 2020).’ When looking at pauses, we will look at moments of silent gaps in speech. For this research, a pause time is used between a minimum of 500 milliseconds and longer. Because we did not have programs available to specifically measure time, we used this 500 millisecond boundary. A pause which will be interrupted by the examiner will not be considered a pause. Pauses will be timed twice with a stopwatch and the mean length measured will be used as time for a pause. Furthermore, when analyzing the speech output, we will describe when we think a pause is used and why. The total uses of pauses will be described and analyzed, the time of how long such a pause will take, and in which situations the pause will take place (for example before nouns/verbs or other psycholinguistic variables like adjects/prepositions. See next paragraph).
To compare our pause frequency, we will use the control groups mentioned in paragraph 2.1. Because we want to compare these for 300 words samples, we will the number of pauses used for both groups in such samples. The study of Angelopouloua et al. (2018) teaches us that around 112 words per sample were used, in which around 20 pauses were used for the healthy control group. For every 5,6 words/morphemes used in spontaneous speech, 1 pause takes place. Because we know that 112 words were used per sample with around 20 pauses used, we
18 measure that ((20/112)) x300 = 54) 54 pauses will be used in 300-word samples. So, based on the study of Angelopouloua et al. (2018) we expect that the healthy group of people would use 54 pauses in samples comparable to ours. The study of Redford et al. (2013) is a little more difficult to use given they only mention the amount of pauses per minute which are 21.92 pauses per minute in the first, and 21.63 per minute the second time they do the spontaneous storytelling task. However due to the rare amount of studies available, this study will be used to compare.
Word categories
When working on this research, the following linguistic variables were found and used in this research. Most of the word categories like nouns or articles are pretty basic and shall not be explained, however we will split verbs in action related and non-action related verbs.
There are action-related verbs, in which someone is literally doing something (like ‘kick’ or ‘lick’ or ‘eat’) and non-action related verbs in which someone does not take action (like ‘believe’ or ‘be’ or ‘own’). These two were added together for the general verb section but are taken separately too to see the difference in both and if people with Parkinson’s disease have more problems with verbs over concrete nouns in particular, or more in action related verbs over non-action related verbs (Boulenger et al., 2008).
Inclusion and exclusion criteria
For our participants we did not state inclusion and/or exclusion criteria as our data was already available to us for use.
Inclusion criteria of our control group consisted of adult groups of people with no psychiatric or neurological history and using spontaneous speech. Exclusion criteria were other forms of speech related tasks which did not involve the spontaneous variant of speaking. Although the study of Redford (2012) involved around a storytelling task, it was mentioned as being a spontaneous variant because of the lack of text.
Ethics approval
Approval to use the data was given by the University of Groningen.
Statistical analyses
For this research, the number of pauses will be calculated and will be analyzed. After this, the pause pattern will be analyzed, in which the examiners tend to examine if there is a
psycholinguistic pattern to be found (for example more pauses before verbs, and if yes: what kind of verbs). This will all be described. We used the following statistical tests/analyses to explore differences in our data. All of our data was analyzed in Microsoft Excel.
19 To measure a possible significant difference between action related verbs and non-action
related verbs we used a X² test (Chi Square test) (p=0,05) in Microsoft Excel with the command (=CHIKW.TEST). We measured both total percentages and total numbers, placed these results in tables in excel and did the X² test. After this we measure our X² value and will compare our result with the p-value of 0,05. The 0,05 p-value is a highly used p-value in most literature and for this it will be used in this study.
To measure how far the individual results of our participants were from the mean amount of time used in front of both general pauses and individual word category pauses we used the z-scores (z-test) and looked per category which scores are two or more standard deviations (sd) away from the mean if these were present. We used the formula (z=(x-µ)/ σ) in which z= z-score, x=individual mean z-score, µ= mean score for one category and σ= sd. Mean times were general mean times for all pauses used, or mean times per word category used and can be found in paragraph 3.3. Standard deviations were measured in Microsoft Excel with the command =STDEV.s().
Results
3.1 Pause frequency
The total amount of pauses used by all participants was 142. The total amount for every
individual participant can be found in Figure 1. The mean number of pauses was calculated with the following formula: M=(P/TP) in which M= Mean, P=Total amount of pauses and TP= Total amount of participants. This gives the following formula: M= (142/7). The mean number of pauses used is (M=20,29) per participant. While looking at range, the following formula is used: R=(HV-LV) in which R=range, HV= highest value of pauses and LV=lowest value of pauses. This gives the following formula: R= (35-5) which gives (R=30). The standard deviation or (sd) was measured and resulted in (sd=10,47). See also Table 3. How these results compare to a healthy group of participants will be discussed in the discussion of this study, see paragraph 4.2.
20
Figure 1. Total number of pauses used in 300 words of spontaneous speech.
Table 3
Number of pauses, mean, range and sd of pauses used
Total amount of pauses Meana Rangeb Sd
142 20,29 30 10,47
Note. a Mean number of pauses. b Range of pauses. c Standard deviation between pauses.
3.2 Word categories
Most of the pauses appeared in front most of verbs, adverbs, conjunctions and pronouns with pauses before adverbs as most common (32 times out of the 142 pauses). After adverbs, most pauses were in front of verbs (24) followed by pronouns (23), conjunctions (22), nouns (20), prepositions (12), articles (5) and the least in front of adjectives (4). In Figure 3, the word categories before pauses were made can be found for every participant with Parkinson’s disease. 21 9 19 27 5 35 26 0 10 20 30 40 50 60
Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6 Participant 7
N u m b er o f p au se s
21
Figure 3. Psycholinguistic variables
So, a large amount of the pauses used were in front of verbs and nouns. When looking more specifically at verbs, most pauses overall seem to have been made before action related verbs (16) as before non-action related (8) from our total amount of 24 pauses used in front of verbs. Participant 5 was the only participant which did not use pauses in front of verbs. Furthermore, some of our participants only used 1 out of each category (participants three (one non-action verb and four (seven action verbs)). See Figure 4 for the use of action and non-action related verbs over all our participants.
To see if there was a significant difference between pauses in front of action and non-action related verbs, we did a X² test. We did not find a significant difference between pauses in front of action related verbs and non-action related verbs (X² (6) =0,367748777, p=0,05). We
conclude that we did not find a significant difference between action related verbs and non-action related verbs (p>0,05).
2 2 2 2 7 5 4 4 1 7 5 3 1 1 2 7 2 2 2 2 6 1 3 1 8 9 1 7 4 1 1 2 5 2 3 4 2 1 5 8 1 1 1 1 1 0 5 10 15 20 25 30 35 40
Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6 Participant 7
N u m b er o f p au se s
Psycholinguistic variables
Pauses before nouns Pauses before verbs Pauses before articles
Pauses before conjunctions Pauses before adverbs Pauses before prepositions
22
Figure 4. Action and non-action related verbs and pausing
3.3 Pause time
As for the time used per pause, the results can be found in Table 3. The mean amount of time for a pause used ranged from 1,39 seconds (participant 2) minimum up till 4,84 seconds maximum (participant 7). We did a z-test (Z test, mean= 2,65, sd=1,18) and found that none of the participants had a z-score which was more as two standard deviations away from the mean time. The time which was closest to more as two standard deviations away came from
participant seven who had a z-score of 1,86. All other z-scores can also be found in Table 4.
Table 4:
Mean time used per pause and z-scores
Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6 Participant 7 Total time
for all pauses (seconds) 42,62 12,55 52,43 64,83 8,92 91,65 101,62 Number of pauses 21 9 15 27 5 35 21 Mean time per pause (seconds) 2,03 1,39 3,5 2,4 1,78 2,62 4,84 Z-scores -0,53 -1,07 0,72 -0,21 -0,74 -0,03 1,86
When looking more closely at the results, the z-scores were calculated per word category which gave the following results (see Tables 5 and 6). None of the scores measured were more as two standard deviations away from the mean amount of time used. Participant seven was the only participant which has positive z-scores for all the word categories used and the highest z-score overall for in this case adverbs (Z=1,92).
1 3 7 0 3 2 3 1 1 0 2 1 0% 20% 40% 60% 80% 100% Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6 Participant 7 N u m b er o f p au se s in p erce n ta ge s
Pauses in front of action and non-action verbs
23 Table 5:
Mean pause time and sd per word category
Table 6:
Z-scores per word category
Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6 Participant 7 Nouns -0,87 -1,29 -0,16 0,74 - 0,15 1,41 Verbs -0,42 -0,7 - -0,51 - -0,12 1,74 Articles -0,68 - - -0,46 - - 1,15 Conjunctions -0,42 -1,28 1,69 0,25 -0,95 0,59 0,12 Adverbs -0,37 -0,74 0,74 -0,29 -0,94 -0,31 1,92 Prepositions -0,28 - 0,35 -0,58 - -1,03 1,55 Pronouns -0,44 - -1,09 0,76 0,16 0,46 1,67 Adjectives - - - 0,30 -1,12 - 0,81
Note. Open spaces represent no use of pauses in front of specific word category
Finally, we measured both the action related verbs, as the non-action related verbs. As we already mentioned in paragraph 3.2, the total amount of verbs was 24. From these 24 verbs, of which 16 were action related verbs and 8 were non-action related verbs. Although this amount is rather low, we measured mean amount of times for five of our participants to see if there were interesting conclusions to be found, see Table 7. Participant three (1) and five (0) did not have enough verbs to use for our study. Participant four only used action related verbs. The mean time over all action and non-action related verbs can be found in the last column in table 6.
Mean pause time per category Standard deviation
Nouns 2,97 1,43 Verbs 2,61 1,81 Articles 1,90 0,76 Conjunctions 2,71 0,77 Adverbs 2,95 1,98 Prepositions 2,62 0,76 Pronouns 2,78 0,96 Adjectives 2,64 1,51
24 Table 7.
Mean amount of times action and non-action related verbs Participant 1 Participant 2 Participant 3 Participant 4 Participant 5 Participant 6
Participant 7 Mean time used overall Mean time action related verb 1,81 1,17 - 1,96 - 2,49 6,03 2,69 Mean time non-action related verb 1,77 1,90 - - - 1,98 5,25 2,73
25
Discussion
4.1 Introduction and expectations
This study examined the impact of Parkinson’s disease (of which some people also have dementia) on using pauses in spontaneous speech by looking at multiple spontaneous speech samples while examining word categories and pause time. The central question was if people with Parkinson’s disease would differ in the frequency of pauses in comparison to a healthy group of people, and to look where these pauses would manifest and for how long. Before we started our study, we had some expectations which we elaborated in the introduction. We will first quickly repeat our expectations.
We hypothesized that:
1. People with Parkinson’s disease will use more pauses in their spontaneous speech as the healthy group of people. Because most of the literature found seems to result in more uses of pauses in speech of people with Parkinson’s disease. This is our main hypothesis over this study;
2. People with Parkinson’s disease will have more and longer pauses before (action related) verbs than nouns in their spontaneous speech. The literature already states that people with Parkinson’s disease have more frequent pauses when producing verbs in comparison to healthy age-matched participants;
3. People with Parkinson’s disease will have more problems in the word category ‘conjunctions. We expect due to the literature and their placement in sentences and utterances to find a high number of pauses in front of conjunctions.
As we look back at our study results, it seemed like our participants did not use many pauses overall in their samples. While this was the case will be further explained in paragraph 4.2 and the conclusion area of this study. Overall, more pauses were made before action related verbs, as before non-action related verbs. However, pause times for both word categories were almost identical, see paragraph 4.4. The number of pauses in front of conjunctions (22) was one of the highest we found overall, but the number of adverbs was way higher (32), something we did not predict at first, see paragraph 4.3.
26
4.2 Pause frequency
To begin, all of the participants used several pauses during their spontaneous speech. Some of those pauses were made on logical places, like in the beginning of sentences. However, most of the pauses which were used tended to be at so-called ‘atypical’ places like in the middle of sentences (from the 142 pauses used by the participants, 86 of them were at such places). This was something which should not be surprising, given the studies of both Lee et al. (2019) as the study of Huber et al. (2012) which both state that both older healthy adults, as people with Parkinson’s disease tend to use more pauses at irregular or as they call it ‘atypical’ places as young healthy adults. Even more so, Lee et al. (2019) already considered that normal aging results in an increased production of pauses at more ‘atypical’ locations, and that it not
necessarily has to be the result of the Parkinson’s disease, or dementia with Parkinson’s disease. With that in mind, the results found in the spontaneous speech regarding the places in which the pauses occurred of the seven participants of this study is logical and was to be expected. Given the fact that this is the first study which examines the first 300 words of spontaneous speech from people with Parkinson’s disease or Parkinson’s disease with dementia, it was not clear beforehand what is seemed to be normal when investigating the amount of pauses used in spontaneous speech. This is also due to the fact that the term ‘pause’ seems to be abstract, given the fact that a proper term for pause could not be found. This was why we used a specific term to describe the kind of pauses which were being investigated in this research. More about this term later. When examining the number of pauses used in this research, we already noticed very soon that the amount differed quite much between the participants. Participant 5 was one of the most interesting people because she ‘only’ used five noticeable pauses in her sample of 300 words (1,66% of words were combined with pauses) in this sample. The fact that this was interesting was the fact that most other participant used way more pauses, and seemed to have more problems in producing speech, which might be due to the consequences of the
Parkinson’s disease (or PD with dementia) on their speech. Participant six produced the greatest number of pauses in the 300-word speech sample (35) (11,66% of words were combined with pauses). The mean number of pauses for all seven participants was around 20 pauses used per sample.
But how does this compare to a group of healthy adults. We begin with comparing our pause frequency with the one from Redford (2013). We know that the ten parent-child dyads were examined with the use of the frog storytelling task (see paragraph 1.1) and that this happened twice. The healthy group of adults produced 21.92 pauses per minute the first time, and 21.63 pauses the second time. Without even measuring we can already see that adults will use more pauses in comparison to our group of people, which used a mean amount of 20,29 pauses over their 300-word samples. Samples of our participants took around several minutes to come to 300 words, this differed quite with for example 03,26 minutes (participant 4) to 04,40 minutes (participant 7). So why did our participants used so little pauses in comparison to the healthy adults in the study of Redford? We speculate that this might be the result of some factors. Firstly, because the participants in the study of Redford (2013) had already a certain point to
27 start from, which they could continue all the way till the end of their stories, they always had a form of inspiration in the pages in which they could use words to describe the pictures. Because they could always use the next pictures when they were done with a certain one, they could easily tell more and probably use longer sentences than our people with Parkinson’s disease, who did not use many words per minute during their clear slow speech and needed to be encouraged (by asking questions) certain times due to not knowing what to say or because certain answers were so brief it was not clear what the participant wanted to say. Because of this, it could be the case that the healthy group used more pauses overall due to more speech used per minute. Secondly, it was an average amount of pauses in the study of Redford. The average amount of pauses used for our participants was as we already said, 20 pauses while there were big differences between our individual participants. We thought about measuring our pauses per minute, but this would give a distorted image overall, given that some of our participants used more pauses later on and it differed quite heavily per minute how many pauses were used (which is probably due to how the Parkinson’s disease influenced speech per individual). Finally, we need to consider the fact that our participants have cognitive problems and were older than the parents of the children. This might also have interfered in the total amount of words/utterances used when comparing this to healthy adults (Pitcairn et al., 1990). But with all things considered, it seemed that overall, less pauses were used by our participants when comparing our results with those of Redford (2013) and his healthy group.
This seems to also be the case when comparing our results with those of Angelopouloua et al. (2018). In the introduction, we calculated that a mean amount of around 112 words were used in the spontaneous speech samples of healthy controls, with around 20 pauses in their samples. A total of 300 words were used for our spontaneous speech samples. This means that around a mean amount of ((20/112) X300) 53,58 pauses (silent) were used when 300 words would be used in total over 19 healthy participants. Those are more pauses used as the 20 mean pauses which our seven participants used. The reasons for this might be the same as the ones we spoke about when comparing it with the study of Redford (2013). Out of two studies we can now assume that all our participants individually used fewer pauses in spontaneous speech than the healthy group of people from various studies.
4.3 Word categories
When we started our study, it was already clear that the word categories nouns and verbs would be the main word categories as most literature seems to be centered around these specific word categories. With that in mind we centered expectation two around (action-related) verbs. The literature clearly stated that people with Parkinson’s disease have more trouble with verbs, and mostly for action related ones, as to concrete nouns (Boulenger et al., 2008; Cotelli et al., 2007). With this evidence in mind, we expected that our participants would show the same by using more pauses in front (action-related) verbs instead of nouns, and with longer pauses in front of them in comparison to concrete nouns. While we examined the
28 number of pauses used for both nouns and verbs, and while splitting verbs in action - and non-action related verbs, we concluded the following. Pauses were measured for both word categories in six out of our seven participants. Only one participant (participant 5) did not use pauses in front of both. Why this was the case was unknown, but overall, this participant was the one with the least number of total pauses. So, while one participant did not pause in front of both, the other six participants did make pauses in front of both nouns and verbs. The total amount of nouns and verbs was almost the same over these six participants (20 nouns and 24 verbs) and both categories represented a large amount of the total amount of pauses used (44 out of 142 pauses or 31% of the total pauses). So there were slightly more pauses in front of verbs as to nouns, but there seemed to be some outliers as participant six used seven pauses in front of nouns while most only used two pauses (participants 1-4) so participant six enlarged the mean quite. For verbs, these results were more balanced. When examining which number of verbs consisted of action-related verbs and non-action related verbs we found the following. In total there were 24 verbs, from which sixteen were action related verbs and eight were non-action related verbs. So, like we expected given the literature, more pauses were made in front of action related verbs as to non-action related verbs.
Also, the study of Angelopouloua et al. (2018) also tells us something interesting when
comparing the use of nouns and verbs in a healthy group of people and the people we used in our research. The total amount of nouns and verbs were calculated in all samples, with a mean amount of 18.19 nouns and 20.5 verbs used over around 112 words in total. When calculating this to 300 words, we get 48.72 nouns 54.91 verbs. We already calculated the mean number of nouns and verbs used in our samples, and these were 38,29 nouns and 51.71 verbs used. We already see that the number of verbs is nearly identical with around 3 verbs more for the healthy group. However, the number of nouns used by our participants is way lower with around 10 nouns less per sample. So, while verbs were used around the same number, nouns were used less and this is interesting given the fact that verbs are more problematic for people with Parkinson’s disease but the amount of verbs used is nearly the same as compared to the control group used (Boulenger et al., 2008; Angelopouloua et al. (2018). This might however be due to the relatively small sample size we had with only seven participants.
Also, we did a X² test to examine if there was a significant difference to be found between action related verbs and non-action related verbs over all our participants. We conclude that such a difference could not be found (X²≈ 0,37) with our p value (p=0,05). We conclude that we did not find a significant difference between action related verbs and non-action related verbs as our p-value was higher as the result we found.
Furthermore, another interesting result was the fact that all participants did have a large number of pauses in front of both conjunctions, and adverbs, and that both variables took a large part in total pauses made by our participants. Given the literature, it seems highly logical that conjunctions would be more problematic. Interruptions of fluency, like in spontaneous speech, take many forms. Think of more and longer pauses between or in the middle of
29 sentences (the atypical places). Given the fact that conjunctions are the bridge in speech which connects words, phrases, clauses or sentences, they are placed most of the times in ‘atypical’ places and because of this, the fact that a lot of pauses were found for conjunctions is in line with the literature (Altmann & Troche, 2011). They stated that because of interruptions of fluency, more and longer pauses will be found between or in the middle of sentences, which happened with our participants. With that in mind, the amount of 22 pauses in front of conjunctions over all participants seems rather small in comparison to adverbs (32 pauses). The fact that every participant used pauses in front of adverbs, and most pauses were in front of the word category adverbs might be due to the following. One of the earliest definitions of ‘adverb’ is: Adverbs are words which change, qualify or modify some types of words like
adjectives, verbs or other adverbs (Curme, 1935). Later he added that adverbs modify not only a single word but also prepositional phrases, subordinate clauses or independent statements as a whole, which makes it a category which likely modifies a large amount of words and phrases. More recent literature seems to mostly agree with these definitions (Huddleston & Pullum, 2002). They sometimes cause some confusion in linguistics due to their wider range of semantic concepts and because their syntactic distribution is disparate. Some linguists even regarded the category of ‘adverbs’ as being the ‘linguistic garbage bin’ (Cann, 1993). All this still causes for discussion. Adverbs itself can be divided in more classes, as several sub-classifications has been proposed. This means that the term adverb as it is, could represent a broad amount of words which means that linguistically speaking, many words can be labeled as adverb. As most of the pauses made were in front of adverbs this might be more logical. Like we mentioned earlier the most used word categories which are being investigated nowadays seem to be centered around nouns and verbs while the so-called ‘linguistic garbage can’ might represent more words as we would expect, given that they are optional and rather vague. Frey (2003) and Fred and Pittner (1998) even go so far in distinguishing five separate adverb classes, although their sub-system is not implanted easily in modern linguistic research. With all this information it might be logical that many words represent adverbs or at least, more than we initially thought. Furthermore, because of their modifying nature and the fact that many pauses were made at ‘atypical’ places like in the middle of sentences (the place where people with Parkinson’s disease tend to use more pauses), it might be logical that most pauses were made in front of adverbs.
