- Matteo Bologna1,
- Giovanni Fabbrini1,2,
- Luca Marsili2,
- Giovanni Defazio3,
- Philip D Thompson4,
- Alfredo Berardelli1,2
- 1Neuromed Institute IRCCS, Pozzilli (IS), Italy
- 2Department of Neurology and Psychiatry, ‘Sapienza’ University of Rome, Rome, Italy
- 3Department of Neurological and Psychiatric Sciences, University of Bari, Bari, Italy
- 4University Department of Medicine and Department of Neurology, Royal Adelaide Hospital, The University of Adelaide, Adelaide, South Australia, Australia
- Correspondence to Professor A Berardelli, Department of Neurology and Psychiatry, ‘Sapienza’ University of Rome, Viale dell'Università, 30, Rome 00185, Italy;
- Received 23 August 2012
- Revised 17 October 2012
- Accepted 13 November 2012
- Published Online First 12 December 2012
The aim of this paper is to summarise the main clinical and pathophysiological features of facial bradykinesia in Parkinson's disease (PD) and in atypical parkinsonism. Clinical observation suggests that reduced spontaneous and emotional facial expressions are features of facial bradykinesia in PD and atypical parkinsonism. In atypical parkinsonism, facial bradykinesia is complicated by additional dystonic features. Experimental studies evaluating spontaneous and emotional facial movements demonstrate that PD is characterised by a reduction in spontaneous blinking and emotional facial expression. In PD, neurophysiological studies show that voluntary orofacial movements are smaller in amplitude and slower in velocity. In contrast, movements of the upper face (eg, voluntary blinking) are normal in terms of velocity and amplitude but impaired in terms of switching between the closing and opening phases. In progressive supranuclear palsy (PSP), voluntary blinking is not only characterised by a severely impaired switching between the closing and opening phases of voluntary blinking, but is also slow in comparison with PD. In conclusion, in PD, facial bradykinesia reflects abnormalities of spontaneous, emotional and voluntary facial movements. In PSP, spontaneous and voluntary facial movements are abnormal but experimental studies on emotional facial movements are lacking. Data on facial bradykinesia in other atypical parkinsonism diseases, including multiple system atrophy and corticobasal degeneration, are limited. In PD, facial bradykinesia is primarily mediated by basal ganglia dysfunction whereas in PSP, facial bradykinesia is a consequence of a widespread degeneration involving the basal ganglia, cortical and brainstem structures.
Bradykinesia refers to the slowness in executing simple and complex voluntary movements, whereas hypokinesia and akinesia refer to a reduction in amplitude or loss of voluntary and spontaneous movement, such as reduced arm swing and decreased stride length, or slow reaction time.1 ,2 A characteristic feature of bradykinesia in Parkinson's disease (PD) is the progressive slowness in speed or a decrease in amplitude of sequential movements (ie, ‘sequence effect’ or decrement).3 Bradykinesia and hypokinesia have been extensively investigated in the upper limbs in PD and in atypical parkinsonism (ie, progressive supranuclear palsy (PSP), multiple system atrophy (MSA) and corticobasal degeneration (CBD)) in clinical2 and neurophysiological studies.1–3 In contrast to the many studies of limb bradykinesia, few studies have investigated facial bradykinesia in PD and in atypical parkinsonism. This is surprising as facial motor impairment (ie, reduction or loss of spontaneous and emotional facial expression—hypomimia) is an important clinical sign of PD, PSP and MSA, and is commonly rated in scales assessing severity of motor impairment in these conditions.
In comparison with upper limbs, the facial motor control system has several distinctive physiological features. Facial motor neurones receive neither reciprocal nor recurrent inhibition.4 ,5 Unlike limb muscles, facial muscles do not act on joints and have few or no proprioceptors.6 Equally important, the moving part shows lower inertia during facial movements than during arm movements.7 Facial and arm movements differ in terms of electromyographic (EMG) muscle activation in that the triphasic EMG pattern that characterises arm movements is not a feature of facial movements.1 ,7 Moreover, other factors that can potentially contribute to bradykinesia, including rigidity and tremor, scarcely affect the face.1 Finally, facial movements may be not only voluntary, but also spontaneous and emotional in contrast with the predominantly voluntary limb movement. For these reasons, the clinical and pathophysiological features of facial bradykinesia differ from those in the limbs in PD and atypical parkinsonism.
In this paper, we first review the major findings about the physiological mechanisms underlying facial motor control. Then we present clinical and experimental evidence describing facial bradykinesia in PD and other causes of parkinsonism, highlighting the pathophysiological mechanisms.
Physiological basis of facial motor control
In the early 1950s, Penfield and Jasper first provided a comprehensive classification and localisation of the cortical motor areas, including the somatotopic description of the facial motor areas. More recent studies of non-human primates revealed five distinct cortical facial representations corresponding to defined facial nucleus subsections.8 The primary motor cortex (M1), the ventral lateral premotor cortex (LPMCv) and the dorsal lateral premotor cortex (LPMCd) lie on the frontal lobe convexity. The other facial areas lie on the medial surface of the cerebral hemisphere (supplementary motor cortex, M2; rostral cingulate cortex, M3; caudal cingulate cortex, M4). Functional neuroimaging and neurophysiological studies in humans provided evidence for a somatotopic organisation of the human striatum9 and subthalamic nucleus (STN).10 Using functional neuroimaging, Gerardin et al9 found ventral and medial putamen activation during face (lip) movements, while using single cell recordings, Rodriguez-Oroz et al10 reported facial representation in the dorsolateral two-thirds of STN. These data provide evidence of a somatotopic representation of the face at the level of cortical motor areas and at the level of subcortical structures, including the striatum and STN.
Animal studies have investigated the distribution of the corticofacial projection and the musculotopic organisation of the facial nucleus. It was reported that M1 and LPMCv give rise to the main corticofacial projection; in contrast, M2, M3, M4 and LPMCd give only a minor contribution to facial muscle innervations; M1, LPMCv, M4 and LPMCd primarily innervate the contralateral lower facial muscles whereas M2 and M3 bilaterally innervate the upper face.8 In humans, transcranial magnetic stimulation (TMS) studies provided direct insight into facial muscle innervations. When single pulse TMS was applied to activate the M1 facial area, motor evoked potentials (MEPs) were recorded in the upper and lower facial muscles.5 ,11–16 Single pulse TMS applied to activate the M1corticobulbar projections elicited MEPs in the contralateral upper and lower facial muscles.5 ,11 ,12 When single pulse TMS was given over the midline frontal region, MEPs were also evoked bilaterally in the upper facial muscles.11 Further investigations showed that single pulse TMS over M1 evoked contralateral and ipsilateral motor responses in the lower facial muscles.13–16 In summary, TMS observations in humans confirm the results provided by animal studies, collectively suggesting that the medial frontal cortex controls the upper facial muscles and that the M1 projections control the contralateral lower facial muscles.5 ,11 ,12 There is experimental evidence, however, challenging the traditional view of contralateral cortical innervation of the lower facial motor neurones and providing evidence for M1 projections to the ipsilateral lower facial muscle in humans.13–16 The observation that latencies of the ipsilateral and contralateral MEPs were similar makes unlikely the possibility that MEPs ipsilateral to the stimulated hemisphere were mediated through interhemispheric connections or other polysynaptic circuits crossing back over the midline at the brainstem level.15 A number of methodological issues and possible confounding factors due to possible contamination of the blink reflex responses, or volume conduction, during the facial MEPs recordings have also been taken into account.12 ,13 More recently, using single unilateral deep brain electrical stimuli, Costa et al17 recorded MEPs in cranial muscles compatible with activation of the corticobulbar tract. The authors found similar latencies in ipsilateral and contralateral MEPs, confirming an ipsilateral corticobulbar projection to upper and lower facial muscles.17 Demonstration of an important input from both hemispheres on either side of the lower face is relevant for the pathophysiological interpretation of facial palsies due to cortical lesions.15 ,17 The clinical observation that upper facial muscles are often spared in central facial palsies may be due to a greater amount of subcorticobulbar input to these muscle.15 Again, demonstration of uncrossed cortical projections to the lower facial muscle does not support the assumption of the hemispheric lateralisation of emotions.15
Facial muscles mediates a variety of involuntary (spontaneous) and voluntary facial movements, ranging from stereotyped eyelid and eyebrow movements (ie, eyelid blinking, raising eyebrows and frowning) to more complex movements of the lower face, such as grimacing, grinning and lip pursing, which contribute to facial mimicry and to orofacial motor functions (ie, chewing, swallowing, pronouncing syllables and talking). A peculiar feature of the facial motor system is that it allows widely differing prototypical emotional expressions (ie, happiness, sadness, anger, fear and disgust) or more ambiguous and subtle expressions.
In an early study, Monrad-Krohn18 proposed a distinction between the neural pathways controlling involuntary (spontaneous and emotional) and voluntary facial movements. For example, lesions of cortical motor areas and the pyramidal tract led to volitional facial paresis but did not affect spontaneous and emotional activation.19 ,20 In these conditions, spontaneous and emotional control is maintained because the corticofacial projection from the mesial frontal areas is preserved.19 ,20 In contrast, emotional facial paresis is thought to result from lesions involving the white matter of the frontal lobe, operculum, mesial temporal lobe, insula, the striatocapsular territory and the thalamus.19 ,20
Facial movements in healthy human subjects have been investigated using neuroimaging and neurophysiological techniques. Although functional MR imaging studies and electro-oculogram suggest that the mesial frontal areas may be involved in generating spontaneous eye blinking,21 other neurophysiological studies suggest alternative explanations.22 For example, it was reported that the spinal trigeminal complex is a major element in the spontaneous blink generator.23 It is likely, therefore, that the structures generating spontaneous blinking include cortical and subcortical structures. Again, neuroimaging investigations indicate that regions associated with the spontaneous and posed emotional facial expressions, like smiling, are different. It has been suggested that a complex neural network mediates spontaneous emotional facial expression—namely deactivation of the frontal areas mediated by activation of other cortical areas (ie, left temporo-occipitoparietal junction, left lateral prefrontal lobe and basal temporal lobes) contributes to disinhibition of emotional expression.18 ,24 On the other hand, posed expressions (ie, posed smiling) and other voluntary movement of the face (ie, blinking) are controlled by cortical areas of the frontal lobe convexity and of the mesial frontal region, including the primary and non-primary motor cortex.11–16 ,25 In conclusion, clinical observation, along with neuroimaging and neurophysiological studies overall, indicate that separate, although partially overlapping, cortical and subcortical areas are involved in the generation of spontaneous, emotional and voluntary facial movements.
Facial movement abnormalities in PD
One of the most distinctive clinical features in PD is hypomimia, a reduction or loss of spontaneous facial movements and emotional facial expression.2 In patients with PD, facial expression resembles that of a person without interest in the surrounding environment (‘masked/poker face’). The palpebral fissures are wider (staring expression), the nasolabial folds are flattened, wrinkles on the orbicularis oris are reduced and the mouth is unintentionally opened.2 Hypomimia is usually bilateral and symmetrical. In up to ∼5% of patients with PD, one half of the face is more affected than the contralateral half (hemihypomimia).26 Hemihypomimia is homolateral to the most affected side of the body and to the side of the body where symptoms first started.26 Hypomimia and limb bradykinesia improve in rapid eye movement sleep and during rapid eye movement sleep behaviour disorder.27 The improvement in hypomimia during sleep is possibly due to activation of the primary motor cortex and lower motor neurones, through descending projections which bypass the extrapyramidal system.27 Dopaminergic treatment improves hypomimia whereas some studies reported that STN–deep brain stimulation (DBS) may deteriorate facial motor control.28–30 It is still unclear, however, whether the detrimental effects of STN–DBS on facial movements simply reflects postoperative modification of dopaminergic medication29 or might be a result of current spread and changes in corticobulbar projection excitability,29 or changes in cortical–basal ganglia activity30
In PD, facial movement abnormalities have often been characterised in neurophysiological studies in terms of decreased frequency of spontaneous blinking rate.7 ,30–35 Although the pathophysiology of reduced spontaneous blinking in PD has not been fully elucidated, it seems to be related to central dopamine deficiency and improves following dopaminergic replacement.31 More recently, it has been observed that the reduced spontaneous blinking rate in PD is increased by STN–DBS.30 As STN–DBS also modified voluntary, but not reflex, blinking, it was suggested that the effect of STN–DBS on the spontaneous blinking rate was mediated through changes in cortical–basal ganglia activity.30 In advanced stages of PD, however, spontaneous blinking may be increased, suggesting a form of ‘off period’ dystonia (blepharospasm).35 In PD patients with an increased blink rate, dopaminergic replacement tended to reduce it.7 ,35 Hence dopaminergic replacement stabilises the spontaneous blink rate frequency.7 ,35
Loss of emotional facial expression in PD may reflect problems with spontaneous smiling, including reduced frequency and reduced degree of mouth opening during smiling.36–38 The reduction in spontaneous smiling in PD is not caused by diminished appreciation of the humorous content but it is unclear whether it relates to depression.36 ,37 It has also been suggested that loss of spontaneous emotional facial expression in PD patients might be secondary, at least in part, to impairment in facial emotion imagery39 or recognition.40 The main deficit in decoding emotions in PD concerns the emotion of disgust, fear and sadness, but not happiness.41 ,42 It is also possible that impaired facial emotion recognition in PD patients depends on cognitive abnormalities.42 In an early study investigating how patients with PD exercise control over facial movements, Rinn43 proposed that although these patients fail to make spontaneous emotional facial expressions they have no difficulty in posing facial emotions (emotional miming) when asked to do so voluntarily. The posed smile has been evaluated using subjective rating scales or using objective computerised methods.37–39 Investigations regarding posed expressions in PD, however, have reported contradictory results. Although some reported normal posed facial expressions in patients with PD,37 ,43 impairment of all voluntarily posed emotional expressions was also reported, both in the intensity and quality of movements.38 ,39 PD patients also find it difficult to make incongruent expressions—for example, when asked to show disgust while watching an amusing video.38 The impairment in emotional facial expression (including emotional miming) in PD may depend on pathological changes (ie, atrophy and Lewy body formation) occurring at the level of the amygdala.44 Other observations suggest that PD patients also have a deficit in performing non-emotional voluntary facial movements such as blinking7 ,30 ,34 and mouth movements,45 ,46 suggesting that the deficit in emotional facial expression in PD is a pure motor problem.
A number of neurophysiological studies have also investigated voluntary facial movements in PD.7 ,30 ,34 ,45 ,46 Bradykinesia was evident in the lower face in studies of orofacial movement with reduced velocity and amplitude of movement as well as increased labial tone during repetitive syllable productions.45 ,46 Labial rigidity was consistently associated with decrements in the range of lip movements and EMG from the orbicularis oris and mentalis muscles, suggesting a possible relationship between rigidity and hypokinetic movement abnormalities of the lower face in PD.47 Concerning movements of the upper face, kinematic analysis of blinking allows a detailed evaluation of the execution of closing and opening phases and of the switching time between the two phases (ie, the inter-phase pause during voluntary, spontaneous or reflex blinking).7 ,30 A kinematic analysis of blinking in PD has shown normal velocity and amplitude of the closing and opening phases during voluntary blinking7 but the duration of the inter-phase pause was longer in PD than in control subjects and was only partially reduced by dopaminergic treatment.7 On the other hand, there was no evidence of impaired kinematics or duration of the inter-phase pause in PD during reflex blinking.7 These findings are in line with previous studies in patients and in animal models of PD showing that although enhanced in terms of excitability, the brainstem circuits of the blink reflex are preserved.48–51 Hence altered switching between the closing and opening phases of voluntary blinking is likely a consequence of the abnormal basal ganglia output in PD, leading to functional deafferentation of the mesial frontal region areas, including the supplementary motor areas, which subserve sequential movements.7 Our recent observation that STN–DBS in PD patients prolongs the inter-phase pause duration for voluntary blinking without modifying this parameter during spontaneous and reflex blinking30 supports the hypothesis that the basal ganglia and interconnected cortical structures are primarily involved in the abnormal coordination of the timing and reciprocity of the orbicularis oculi and levator palpebrae superioris muscle activation in PD.7 ,30
In summary, in PD, the reduction in spontaneous and emotional facial expression contributes to the less intense facial expressions of PD. In addition, voluntary orofacial movements are bradykinetic, whereas voluntary eyelid movements are only impaired in terms of switching between the closing and opening phases of blinking (table 1).
Facial movement abnormalities in atypical parkinsonism
Hypomimia generally manifests early in the course of PSP as the loss of spontaneous facial expression and dystonic contraction of the facial muscles. Facial dystonia in PSP encompasses deepened nasolabial folds, dystonic vertical wrinkles in the glabellar region and bridge of the nose secondary to contraction of the frontalis, procerus and corrugator muscles (ie, the procerus sign).52 In contrast with patients with PD, who have flattened nasolabial folds, patients with PSP show deep nasolabial folds (dystonic) in an expressionless face.53 Patients with PSP may also show a progressive widening of the palpebral fissures, giving them a staring gaze (‘reptilian’ stare) or a surprised and astonished facial expression.53 Another feature that differs in hypomimia of PD and PSP is the response to dopaminergic treatment. In contrast with PD, dopaminergic treatment does not affect hypomimia in PSP and may even cause blepharospasm and other dystonic facial features to worsen.53
Hypomimia is also present in MSA. A study comparing MSA with PD patients matched for age and disease duration reported that hypomimia was more common in early PD but that at the final visit, hypomimia measured by the Unified Parkinson's Disease Rating Scale-III was more severe in patients with MSA.54 Interestingly, patients with MSA had more expressive faces during rapid eye movement sleep behaviour disorder in comparison with facial expression during wakefulness.55 Hypomimia in MSA may be accompanied by orofacial and platysma dystonia, as in PSP,53 and also resemble risus sardonicus in tetanus.56 In MSA, dopamine replacement is usually ineffective and may induce pronounced dystonic–dyskinetic contraction of the facial muscles.57 This susceptibility to facial dystonia could be associated with the relative sparing of orofacial ventral striatopallidal circuitry in MSA, which is somatotopically related to facial muscles.
The spontaneous blink rate is reduced in PSP, often dramatically (to 3/min).33 ,58 Peak velocities and amplitudes during both closing and opening phases of spontaneous blinking are also reduced.58 The reduction in the blink rate, as well the abnormal kinematic properties of spontaneous blinking observed in patients with PSP, could reflect not only reduced central dopaminergic activity31 but also primary degeneration in cerebral structures, generating spontaneous blinking, including the mesial cortical areas in the frontal lobe and brainstem.21–23
To date, only one study has investigated the voluntary control of facial movements in atypical parkinsonism using a kinematic technique.58 It was reported that the closing and opening phases during voluntary blinking are markedly bradykinetic (ie, last longer in patients with PSP than in healthy controls). On the other hand, the closing phase of reflex blinking is normal.58 These observations suggest that impaired execution of voluntary blinking in PSP may primarily result from degeneration of the frontal areas.58 Supporting this hypothesis, some imaging evidence, obtained with positron emission tomography, showed glucose hypometabolism in the midline frontal regions and caudate nucleus of patients with PSP.58 In addition, the functional deafferentation of the frontal lobe areas involved in the so-called basal ganglia motor loop, due to basal ganglia degeneration, might play a role.3 ,58 Finally, the observation that in patients with PSP the inter-phase pause is prolonged not only during voluntary blinking but also during reflex blinking,58 suggests that brainstem damage related to PSP may play a predominant role in lengthening the switching between the closing and opening phases during voluntary blinking.
In summary, hypomimia in PSP and MSA is often complicated by facial dystonia. In PSP, spontaneous and voluntary blinking is abnormal but there are no data on voluntary movements of the lower face and emotional facial movements. In addition, data on spontaneous, emotional and voluntary facial movements in MSA and CBD are lacking (table 1).
Conclusions and perspectives
Facial bradykinesia in PD reflects abnormalities of spontaneous and emotional facial expression and of voluntary movements of both the upper and lower face. In the upper face, bradykinesia manifests as a markedly low blink rate, as impaired switching between the closing and opening phase during voluntary blinking, and in the lower face as reduced spontaneous and posed smiling and impaired voluntary orofacial movements. In PD, the impaired switching during the opening and closing phases of voluntary blinking and during coordinated lower facial movement synergies resembles the switching abnormalities described in studies on limb movements.1 ,7 In PD, the abnormalities of facial movement control largely reflect basal ganglia dysfunction.7 ,30 Sparing of some elements of upper facial movement (ie, voluntary blinking kinematics) in PD may relate to the distinctive physiological features of the facial motor control system, including dual innervation of the upper face from the medial frontal cortical areas or, alternately, the lower inertia of the eyelid during blinking which might be a safety factor for developing bradykinesia.7 A further possibility is that normal kinematics of voluntary blinking may be due to the pattern of activation of the orbicularis oculi and levator palpebrae superioris muscles that, in contrast with limb muscle activation, lack the triphasic EMG pattern typical of arm movements.1 Unlike patients with PD, those with PSP exhibit more severe abnormalities in both spontaneous and voluntary facial movements which are mediated by both cortical and subcortical structures. Differences among PD and PSP are likely to be due to the distribution of pathology affecting different levels of the cortical and subcortical control of facial movement and by a significant degeneration of the frontal lobe and brainstem structures in PSP.58
A number of issues, however, still need to be addressed before we will fully understand the pathophysiology of facial bradykinesia in PD and in atypical parkinsonism. For example, some of the conclusions on the pathophysiology of blinking in PD and atypical parkinsonism are based on the assumptions that the execution of the closing and opening blink phases and switching between them are mediated by different mechanisms. Moreover, data on facial bradykinesia in MSA and CBD are limited. Also, whether the impaired emotional muscle activation in PD is a specific deficit or simply reflects an abnormality in the voluntary activation of facial muscles is still unclear. Evidence of facial bradykinesia in PD and in atypical parkinsonism, and in particular evidence of impaired voluntary blinking, indicates the need to broaden our perspectives and extend our knowledge on facial bradykinesia and possibly incorporate the clinical evaluation of voluntary blinking abnormalities in the commonly used scales to assess severity of motor impairment.
Contributors MB: conception and design, and drafting the article. GF: conception and design, drafting the article and revising it critically for important intellectual content. LM: drafting the article and revising it critically for important intellectual content. GD: revising the article critically for important intellectual content. PDT: drafting the article and revising it critically for important intellectual content. AB: conception and design, drafting the article and revising it critically for important intellectual content, and final approval.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.