Article Text

Download PDFPDF

Communication changes in Parkinson’s disease
  1. Nick Miller
  1. Speech and Language Sciences, Newcastle University Institute for Ageing, Newcastle upon Tyne, UK
  1. Correspondence to Professor Nick Miller, Speech and Language Sciences, Newcastle University Institute for Ageing, George VI Building, Newcastle Upon Tyne NE1 7RU, UK; Nicholas.Miller{at}


This review looks at the nature and impact of communication changes in Parkinson’s disease, approaches to assessment and directions for intervention. This is especially important since medical and surgical interventions that help limb movement are largely ineffective, or even detrimental, for speech. Most people with Parkinson’s disease notice changes to their communication. Voice alters early on—even in the prodromal stage. Later, articulation may impair intelligibility further. These changes impact on mood and social participation. However, a full characterisation of communication changes in Parkinson’s must acknowledge that changes are far more pervasive and varied than a quiet voice. Communication is affected by marked dysprosody, cognitive-linguistic impairment, alterations to social interaction and pragmatics. Changes entail not just expressive elements but also receptive. A comprehensive evaluation of potential communication challenges faced by people with Parkinson’s disease must cover all these aspects. Similarly, interventions that ignore the breadth and depth of changes will always remain incomplete.

  • parkinson's speech language assessment management
View Full Text

Statistics from


Standard neurological textbooks typically summarise that Parkinson’s disease leads to a quieter voice (hypophonia), possible imprecise articulation (hypokinetic dysarthria) and altered speech rate. This is true enough, but it represents only a partial picture. The disruption to communication in Parkinson’s disease extends much further, not just in terms of speech changes, but in respect of many other factors impinging on communication. Importantly, difficulties involve not just spoken output but also receptive aspects of understanding and processing. Changes in cognitive–language function (semantics and syntax/grammar) and social interaction have far-reaching effects, even in the absence of dementia. Crucially, there can be communication impairment even when speech sounds otherwise intelligible.

This article explores how communication alters and how this affects daily living. It introduces strategies for assessment and mentions directions in management—including why medical therapy and deep brain stimulation (DBS) have little influence on speaking, and may even be detrimental. The focus is on spoken communication, rather than reading and writing.

Why consider communication?

Around 90% of people with Parkinson’s disease report changes to their communication.1 These changes impact on daily living and represent a major influence not just on ability to communicate effectively and to participate fully in work and social activities (though few studies on work access actually give any direct consideration to communication), they also directly affect mood and feelings about oneself.2 There are two-way interactions between communication changes and cognitive, affective and social variables.3 The changes can contribute to carer burden. Attention to communication is therefore central to achieving and maintaining an optimum psychosocial quality of life. Monitoring speech changes is also important since rapidly deteriorating speech can indicate atypical parkinsonism or another intercurrent disorder.

What changes?

Voice (phonation)

Hypophonia is a prominent symptom. It may even prompt the initial suspicion in family and friends that something is amiss. Indeed, phonation has proved a promising variable to monitor in attempts to detect the prodromal changes of Parkinson’s disease and in charting change from the earliest stages.4

Hypophonia is usually attributed to rigidity of thoracic, laryngeal and pharyngeal musculature, which arguably leads to reduced breath support, poorer vocal cord approximation and less effective voice resonance. That provides only a partial explanation. The underlying peripheral voice mechanism is not impaired—unlike in some causes of neurological dysphonia, where changes to muscle tone and power affect voice quality. People with Parkinson’s disease can produce a louder voice, certainly early in their disease course.

The crucial disruption to voice intensity arises more centrally, from underscaling of vocal parameters and inability to monitor voice intensity adequately.5 6 Thus, despite underlying ability to produce greater volume in response to environmental cues (the Lombard effect) or listener requests, patients do not reach full intensity or even if they do, they cannot hold it. This is exacerbated by the impaired self-monitoring and reduced awareness of voice intensity, with people with Parkinson’s disease typically not appreciating they have a quieter voice.

There are further disruptions to voice. People with Parkinson’s disease experience problems initiating phonation, similar to limb motor initiation difficulties. At first, this manifests itself in occasional blocks or hesitant, repetitive sounding speech. Later, pauses become abnormally long (>200 ms; see figure 1), making speech dysfluent, with possible freezing of voice-speech analogous to freezing of gait.7

Figure 1

Oscillogram (top) showing intensity variation and accompanying sound spectrogram (bottom). A 74-year-old male with Parkinson’s disease (UPDRS III Speech rating ‘2’, ‘slurred but understandable’). Reading ‘He slowly takes a short walk [pause] in the open air [pause] each day’. Of note: tendency to monoloudness illustrated by relatively even oscillogram and level peaks on intensity trace on spectrogram (horizontal arrow left centre); appropriately placed but abnormally long pauses (vertical arrows).

Voice tremor may be present (reportedly in 15%–55%).8 9 Notably, this may emanate not from intrinsic laryngeal musculature tremor (thyroarytenoids; cricoarytenoids) but more likely from tremor in respiratory muscles, pharyngeal walls, soft palate or tongue. Tremulous voice may be distressing for some, but is not a common cause of significant reduced intelligibility.

Prosody changes

A core disruption to spoken output is marked dysprosody—that is, alterations to the stress and intonation patterns of speech. Utterances are characterised by a tendency towards flattened intonation (speech all on one note, monopitch), and loss of contrast between stressed and unstressed syllables (monoloudness; see figure 1). This represents one reason for the misperception that people with Parkinson’s disease are depressed, uninterested or tired—when they are not—and for listeners’ negative evaluation of people with Parkinson’s disease—a perception reinforced by hypomimia and reduced arm movements that accompany speech.

As with the voice changes, dysprosody in Parkinson’s disease does not rest solely on impairment of the mechanics of speech from rigidity and bradykinesia. Underscaling of movements contributes too. A key component in dysprosody in Parkinson’s is a more generalised higher-level impairment in understanding and producing appropriate prosody. People with Parkinson’s disease have difficulty clearly differentiating between for example, happy–sad, angry–disappointed, joking–serious tone of voice in their own speech, but also demonstrate a parallel problem in appreciating the contrasting prosodic tones in others.10 11 This represents a likely component in carers’ reports that the person with Parkinson’s disease has lost their sense of humour (eg, they fail to convey humour in their voice or misinterpret another’s joking remark as serious); habitually gets the wrong end of the stick (eg, interprets an ironic remark as literal—What time do you call this!); and they never really know how the person with Parkinson’s disease is feeling. It appears likely that the problems with (mis)perception of affect extend to visual aspects of processing, for example, detecting sad versus happy faces or picking up on implications of body posture in communication. This topic is pursued under pragmatic disorders below.

Speech (pronunciation and articulation) changes

Hypokinetic movements and difficulties with rapid alternation of movements of the tongue and lips lead to weaker articulatory contacts and, in turn, to imprecise sounds and indistinct words.6 12 Once more, the issue concerns underscaled movements rather than simply neuromuscular restrictions on the ability to reach target articulatory positions or velocities. Attempting to maintain a normal rate of speech in the face of an underfunctioning system may also contribute to imprecise articulation. Sounds requiring firm contacts between articulators become especially vulnerable—bee, riding, watching, corn sound like vee, rising, washing, horn. Vowels produced high or far forwards or back in the mouth tend to be pronounced more centrally: together with the consonant imprecision heat, boot, queen sound like hiss, foot, win.

To listeners, the speech rate may sound accelerated. This is an auditory illusion. The imprecise articulatory contacts, monoloudness and monopitch and lack of perceptible boundaries between words found in Parkinson’s speech also characterise fast speech in people unaffected by Parkinson’s (count as rapidly as you can to 20 to illustrate this, or listen to the horse race commentator when it’s neck and neck in the final furlong). The objectively measured rate is the same or even slower than matched unaffected speakers.6 People with Parkinson’s disease show a greater tendency to accelerate speech over an utterance/passage compared with unaffected speakers. Some show short rushes of accelerated speech. These have been linked to attempts to produce longer utterances on one (reduced) breath, realignment to the natural rhythm of speech after difficulty initiating phonation and/or speech festination akin to that found in gait.13

Altered speech movements are detectable from early on, but typically do not seriously affect intelligibility until later.1 However, despite impressions on the part of the listener that speech is unaffected because intelligibility remains viable, this should not imply the person with Parkinson’s disease is not already adversely affected by the changes.2 14 They may maintain intelligibility only at considerable cost in terms of attention to effort and monitoring of speech. Intelligibility may be fine one to one in a quiet clinic room, but ineffective in noisy environments, over the phone, or when the speaker is involved in activities that distract their attention from optimising speech output.15

Language changes

People with Parkinson’s disease frequently report difficulty finding words or expressing their thoughts clearly. What underlies this impression has not received so much attention as voice-speech deterioration and is certainly not as recognised in clinical assessment and management. Nevertheless, there is ample evidence that even those without dementia show changes to language processing from early on.16–18 People with Parkinson’s disease produce less complex grammar than unaffected people. They show slower language processing speed, are more susceptible to misunderstanding metaphor and inferred meaning (he’s a real fighter; are you still putting your shoes on!) and experience difficulty in resolving lexical and contextual ambiguities (he picked up the spade—is this about gardening or playing cards?). A particular difficulty centres on retrieving action verbs/semantics,19 which in turn can dissociate from impairment of syntax and object semantics.

Semantic (tell me as many animals as you can think of) and phonemic/letter fluency tasks (tell me as many words as you can starting with F) are widely employed in assessing language in neurology. When controlled for motor speed people with Parkinson’s disease can perform similarly to unaffected individuals on semantic fluency tasks. Some reports claim poorer performance on letter fluency, but even in unaffected individuals this can be susceptible to literacy influences, so the issue remains open. The reasons for reduced output may evolve over time. Impairment may be dominated by motor slowness in early Parkinson’s disease, but later bradyphrenia and other cognitive changes underlie poorer performance.20 People with Parkinson’s disease struggle especially when they need to alternate between categories (tell me the name of an animal and then something you can eat; tell me a word beginning with F and then with A; keep switching between the two).21

Debate continues regarding how far such changes relate to language specific processing deficits and how far they are linguistic reflections of dysfunction in other cognitive domains.16 18 20 The answer is probably both. Particular aspects of language are impaired, but language comprehension and production are not possible without the support of executive functioning, attentional focus and switching, balanced excitation and inhibition in retrieval of words and grammatical structures, and auditory verbal short term, working memory.

Discourse management and pragmatics

Discourse and pragmatics refer here to the practical and social skills necessary to manage conversations. Successful conversations involve knowing how to gain and start a turn in a conversation. One must be able to signal that one does or does not want another speaker to take the floor. Conversations depend of being able to structure an explanation or request clearly and keep track of themes and changes in topic. One must know how and when to introduce new material, know what knowledge can be taken as given versus when to be more explicit, and so forth. Parkinson’s disease affects these abilities.11 18 22 It also impacts on the ability to recognise that misunderstandings have taken place and the ability to repair them.14

Such changes interact too with the other verbal and non-verbal impairments. Alterations to facial, arm and body posture signals that indicate one wants a turn or is not yet finished hamper entry into and staying in a conversation. Listeners misperceive a delay in initiating voice or abnormally long pauses as signals that one does not want a turn or is finished. Bradyphrenia creates a constant struggle to understand fully and to keep up with talk. Difficulties retrieving words, switching mental sets and distinguishing the implicit (tone of voice, irony, and seriousness) content of exchanges because of receptive and productive prosody problems present added challenges.

Part of the problem here may centre on an impaired theory of mind. Theory of mind enables us to anticipate others’ intentions, desires, emotions, (re)actions and beliefs within the social context as well as to reconcile their (re)actions with our own. It has been argued to be affected in Parkinson’s and posited to contribute to breakdown of the discourse and pragmatic skills outlined above.11 18

Psychosocial impact of communication changes

Communication is central not just to functioning successfully on a day-to-day basis, it is intimately tied up with feelings about ourselves and how others react to us. It is unsurprising, therefore, that changes to communication profoundly affect people with Parkinson’s disease and those with whom they communicate. Speech and language changes, even when intelligibility remains apparently intact, can cause the speaker to lose confidence in speaking, give a sense of inadequacy and/or frustration and a feeling that one is being negatively judged.2 Speech that lacks varied stress and intonation patterns, is quiet and imprecise, or where the person is hesitant in their formulations or replies invites negative judgements by listeners,11 add a further barrier to social participation and positive self-perception.


The clinical assessment examines whether, and to what extent, any of the above changes pertain, and their impact on the ability of the person with Parkinson’s disease to make themselves understood, reliably to understand others, and to participate effectively in daily living. Since the central aim of intervention is always to improve intelligibility and participation, these constitute the key focus of assessment.

To ascertain speech intelligibility levels and what variables should form targets in therapy to improve it, one uses diagnostic intelligibility testing.23 Intelligibility rating scales differentiate broad levels of severity, but can suffer the drawback of poor intra-rater and inter-rater reliability and tell one nothing about which targets to tackle to improve the situation. Once more, performance varies markedly between clinic and real-life situations. Ideally, therefore, intelligibility is also assessed under dual-task conditions and in naturalistic settings.15 24

There are questionnaires to support the evaluation of psychosocial impact. They cover (speaker, carer and clinician) perceptions of what aspects of communication have altered; how this affects participation in social roles and situations; and how these may alter affective self-perceptions and interactions.2 14 25 26

Turning to more impairment-based outcomes, voice assessment establishes the person’s ability to produce a sustained vowel (say ‘ah’ as long as you can), to say this with increasing intensity (louder) and varying and wider pitch range. Easily usable instrumental methods provide objective acoustic measurements to quantify performance,27 bearing in mind that the key articulatory–acoustic variables to measure may differ between languages.

Prolonged vowels are good for certain aspects of voice evaluation, but measures based on a standard reading passage (to permit comparisons across time and between individuals), set monologue (eg, tell me how you would make a cup of tea/coffee) or picture description allow insights closer to real-life situation performance. From these one can employ clinical rating scales or objective acoustic analyses (see figures 1 and 2) to gauge appropriateness and consistency of loudness/intensity level, pitch range and variability, fluency (pauses, repetitions and prolongations) and rate of speech.

Figure 2

Oscillogram (top) and accompanying sound spectrogram. A 70-year-old female with Parkinson’s disease (UPDRS III Speech rating ‘1’, ‘slight speech changes’). Fifteen repetitions of ‘pea’. Of note: loss of intensity (around 75 → 55 dB) across 4.45 s, shown on decreasing oscillogram excursions and spectrogram intensity wave (arrow lower right); loss of sound distinction in mid-high frequencies (arrow centre right), associated with loss of energy/intensity of sound signal and weaker lip contacts. Net perceived effect: change from easily to poorly audible voice, heard at start as ‘pea’, in mid-repetitions more as ‘fee’, at end going to ‘he’.

Prosody production can be evaluated based on reading and monologue samples—using acoustic and/or rating scale instruments. Specifically designed sentences that tap control of stress and intonation patterns can deliver a more focused assessment of prosody production and perception.11 For example, asking the person with Parkinson’s disease to say the same sentence (eg, the dress is yellow; the sausage fell into the trifle) in a neutral versus angry versus jocular tone. Listeners judge which tone they think they have heard. One can ask the person with Parkinson’s disease to signal the difference between SHE drinks coffee and she drinks COffee; disambiguate sentences like, they’re hunting DOGS versus they’re HUNting dogs; produce the same sentence as a statement versus a question, You press this button here! versus You press this button here? For receptive prosody testing, the person with Parkinson’s disease indicates from pictured or written choices which tone of voice or meaning they hear.

Rate measurement can be taken from performance on standard reading tasks, and/or be based on speech diadochokinetic repetition of single (papapapa…; tatatata….) or alternating (patakapataka….) syllables in different rhythms (papaPAA papaPAA…; PAApapa…PAApapa…, etc). These facilitate quantification of rate control, but also sustainability of rhythm and coordination and integrity and maintenance of articulatory contacts (lips for pa, tongue tip for ta, tongue back for ka). Simple instrumental means can supplement naked ear/stop-watch judgements (figures 1 and 2). Given that speech motor control engages functionally different networks to non-speech control of the tongue, lips, and so forth, ideally tasks should be based on real words,28 rather than for example ‘blow out your cheeks’, ‘wiggle your tongue from side to side’.

People with Parkinson’s disease without dementia generally fare well on standard aphasia batteries, since these are designed to measure the aphasia as seen after stroke. They are more challenged by online tasks that look at speed and complexity of processing, tax attention and short-term memory and activation and inhibition of (in)appropriate words.16 17 Alternating category naming tests bring naming/word fluency difficulties into focus.21


NICE (National Institute for Heath and Care Excellence) Guidelines for Parkinson’s disease recommend intervention for communication changes ( Intervention aims to achieve at any one stage of Parkinson’s disease, the optimum level of communication to allow people to participate successfully and meaningfully in family and social life. Clinically relevant outcomes can only be gauged by measuring these factors. Diadochokinetic rate or acoustic measures provide important objective speech information, but there is poor correlation between changes on these measures and satisfaction with and success in daily living—just as grip force or quadriceps strength are not necessarily clear predictors of whether the person can pick up their cup of tea or stand at the sink.

Rehabilitation targets both speaker and listener. Targets may be underlying impairments (eg, quiet voice and dysprosody). Other techniques seek compensatory or alternative strategies to maximise understandability—but always to functional ends: simply being able to produce a louder ‘ah’ or with greater pitch range does not automatically translate into improved communication. Important ingredients are education of the speaker and their family about communication, how it changes in Parkinson’s disease, how to anticipate and manage change and to create an ideal communicative environment. To facilitate a preventive and anticipatory role it is helpful when referral happens as soon as possible after diagnosis.

Key approaches entail interventions that emphasise self-monitoring and attention to effort, focusing on rescaling of vocal intensity.29–31 Rate control, especially when combined with articulatory exaggeration may improve listener perceptions. Rhythmic cueing represents a possibility to aid initiation of speech, maintenance of fluency and control of rate.32 33

Gains in a calm, unhurried clinic room are relatively easy to attain. The constant challenge is to transfer and maintain gains into the to and fro of fast moving, inherently distractive situations outside. Therapy programmes commonly incorporate systematic methods to achieve this. More recently apps have been piloted to enable better self-management of transfer and maintenance of loudness, rate and intelligibility levels in naturalistic contexts.24 34 Other devices have been trialled that augment or compensate for impaired voice and speech. If spoken output becomes unviable alternative or augmentive means of communication are introduced.

Several studies address the effects on voice of acting or singing in a choir,35 but no study has established unequivocal lasting benefits for speech-voice parameters directly related to the breathing and voice exercises of these pastimes. Furthermore, the outcomes are generally biased by recruiting self-selecting participants. However, the studies are unanimous in concluding there are social and psychological advantages and in turn this may translate into less depression, more confidence and, through this, better speech and social participation.

Other interventions attend to the pragmatic and discourse aspects of communication. They (re)train skills for entering and retaining a place in conversations, recognising breakdowns in understanding and how to successfully repair these. Therapeutic management of language processing issues is under-researched, though studies that have examined the effects of cognitive training in patients with Parkinson’s disease may be relevant here.

Repetitive transcranial magnetic stimulation offers a safe non-invasive intervention to treat deficits in patients with Parkinson’s disease, though with variable indications regarding what is ameliorated or not, and short and longer term effectiveness. We have much to learn regarding the variables of site of stimulation and its intensity, depth and frequency. There is little work exploring the effects on speech and voice; weak designs and low numbers cloud conclusions for most studies.36 Readers shoud be aware that one study on TMS and speech has been retracted 37 .

L-dopa effects on speech

There is a longstanding debate on whether dopaminergic therapy has similarly positive effects on speech as it does on limb movement. Speech motor decline correlates only weakly with limb motor changes,1 27 and is largely unaffected by dopaminergic intervention.

Studies and systematic reviews27 38–40 conclude that while individual voice-speech parameters (eg, velocity, intensity of movement and pitch range) may be positively influenced, the benefits do not transfer to improved intelligibility or communication. This likely reflects the fact that speech motor control depends on more than just dopaminergic pathways; the complexity of speech motor control means that there is typically only a weak relationship between the non-verbal oral tasks or the isolated acoustic parameters usually employed to test the articulators and actual live speech performance; and speech is intimately tied up with language processing that involves cognitive processes beyond dopaminergic influence and that may even be depressed via L-dopa ‘overdosing’

DBS and speech

The heterogeneity of patient groups reported and variety of measures used hamper definitive conclusions on the effects of DBS on voice, speech and language. General indications are that sites and settings of stimulation compatible with improvement to limb function are neutral or even detrimental to speech intelligibility and language. Pre-existing significant speech problems serve as a contraindication for DBS in some centres.41 Dysarthria and even dysphasia represent adverse effects of DBS, especially bilateral.42–44

While individual acoustic or speech motor variables may improve, or there is differential impact on underlying processing networks (eg, reaction time versus set switching in language output), these do not carry through to improved intelligibility or perceived benefit of communication.43 DBS may provoke added speech impairments, with greater spasticity and increased dysfluency besides the existing hypokinetic picture.45 Speech–language therapy interventions successful with patients without DBS may not succeed or require modification for DBS treated individuals.46

Some adverse effects on speech may relate to device insertion. Other effects may relate to current diffusion into neighbouring corticobulbar, thalamo-cerebellar or periaqueductal tracts. Pulse width and frequency can also influence speech outcomes.47 48 Little et al 49 found adaptive DBS less detrimental to speech and maybe even capable of benefits.


This overview has emphasised that communication change in Parkinson’s disease is much more than just the quiet voice, and even when voice intensity and intelligibility appear unaffected there are many more barriers to successful communication. Motor, non-motor and cognitive changes permeate all aspects of message formulation and expression. Aspects of communication difficulty link to more far-reaching impairments in auditory perception, attentional control, attentional switching and interpersonal behaviour. Thus, it is more realistic to view communication change in Parkinson’s as a receptive/perceptual and cognitive-linguistic disorder, not just an isolated dysarthrophonia. Vitally, assessment and management must also acknowledge the pragmatic and psychosocial consequences of communication changes. Pharmacological and surgical interventions offer little for speech. This is likely to alter soon with advances in DBS and for those for whom DBS is indicated, but currently behavioural interventions remain the core methods for rehabilitation.

Key points

  • Communication changes are pervasive in their occurrence and impact in Parkinson’s disease.

  • Changes extend well beyond a quiet voice and imprecise speech to encompass language and social interaction variables and receptive as well as expressive challenges.

  • Absence of obvious changes to the naked ear does not mean the person with Parkinson’s disease and/or their family are not affected by communication changes.

  • Speech and language changes are largely uninfluenced by medical therapy, and deep brain stimulation (DBS) currently can result in detrimental effects.

  • Behavioural therapies therefore remain the first choice for rehabilitation of communication changes, especially those that focus on attention to effort, self-monitoring of output and transfer of practice into naturalistic settings.


View Abstract


  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed. Reviewed by Simon Lewis, Sydney, Australia.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Editors' commentary
    Phil E M Smith Geraint N Fuller