Evolving Concepts on Bradykinesia

Matteo Bologna; Giulia Paparella; Alfonso Fasano; Mark Hallett; Alfredo Berardelli


Brain. 2020;143(3):727-750. 

In This Article

Neurophysiological Characterization of Bradykinesia

As a refinement of clinical observations, neurophysiological measures have clearly shown that bradykinesia includes a broad range of motor abnormalities, involving both the preparation and execution phases of voluntary movements, and variably affects different body segments. Many neurophysiological studies on bradykinesia have been performed in Parkinson's disease, conversely a relatively low number of studies have been performed in atypical parkinsonisms.

In Parkinson's disease, the majority of studies have consistently shown that the movement preparation is abnormally prolonged due to a delay in muscle activation, as evidenced by investigations on simple and choice reaction times (Wiesendanger et al., 1969; Heilman et al., 1976; Stelmach et al., 1986; Dubois et al., 1988; Pullman et al., 1988; Rafal et al., 1989; Jahanshahi et al., 1992a, b, 1993; Brown et al., 1993; Revonsuo et al., 1993; Pascual-Leone et al., 1994; Torrecillos et al., 2018). Abnormalities of simple reaction times and no (or less) alterations of choice reaction times (Evarts et al., 1981; Bloxham et al., 1984; Sheridan et al., 1987) suggested that slowness in motor execution in Parkinson's disease was not due to a global deficiency in motor preparation and that the contribution of prolonged simple reaction time in generating bradykinesia could be interpreted as a failure in programming a motor response (Berardelli et al., 2001). The addition of a warning signal improved reaction times both in healthy controls and in Parkinson's disease, demonstrating that the slowness in motor preparation was not due to an 'arousal defect' (Bloxham et al., 1987). Other observations also suggest that Parkinson's disease patients use different strategies for dealing with different stimulus-response relationships; for example, their performance was worse in simpler tasks with compatible stimulus-response relationship (Brown et al., 1993). The effects of levodopa on reaction times have also been shown to be variable, although choice reaction times were shown to be more sensitive than simple reaction times to levodopa administration (Bloxham et al., 1987; Pullman et al., 1988; Jahanshahi et al., 1992b; Brown et al., 1993).

Studies on motor execution have extensively characterized bradykinesia of the upper limbs. Parkinson's disease patients performing fast single-joint movements display an inability to generate an adequate initial agonist burst so that the exerted force is insufficient to reach the desired target. The movement is completed by a series of small amplitude movements characterized by repeated agonist and antagonist EMG bursts or by a prolonged continuous discharge. Despite this, if patients aim to perform a movement of a larger amplitude, they succeed in increasing the size of the first burst. Further studies demonstrated that Parkinson's disease patients choose a smaller size than is appropriate for the motor task. This means that bradykinesia is generated from an inappropriate scaling of the dynamic muscle forces, the so-called scaling effect (Flowers, 1976; Hallett et al., 1977; Hallett and Khoshbin, 1980; Baroni et al., 1984; Berardelli et al., 1986, 1996a; Pfann et al., 2001; Seidler et al., 2001). Complex (multi-joint) movements are also slower in Parkinson's disease (Phillips et al., 1994; Bennett et al., 1995; Castiello and Bennett, 1997; Majsak et al., 1998; Alberts et al., 2000; Castiello et al., 2000; Jackson et al., 2000; Farley et al., 2004). The simultaneous execution of two different tasks in patients with Parkinson's disease leads to a more severe slowing in movement than that observed when each task was performed alone (Schwab et al., 1954; Benecke et al., 1986; Stelmach and Worringham, 1988; Brown and Marsden, 1991; Castiello and Bennett, 1997; Johnson et al., 1998; Delval et al., 2017). Movement slowness is even worse in the execution of sequential movements (Table 1). The slowness of sequential movements is more evident when patients rely on internal control processes (i.e. internally generated movements) in comparison to externally cued movements (Georgiou et al., 1994; Currà et al., 1997; Fasano and Bloem, 2013). This is particularly evident during gait, thus meaning that Parkinson's disease patients experience difficulties walking in an automatic manner and must progressively rely on a goal-directed control of their gait (Nonnekes et al., 2019). External cues trigger goal-directed behaviour that is better performed, which bypasses the defective basal ganglia (Redgrave et al., 2010).

One component of bradykinesia is the sequence effect, i.e. the decrement of movement speed and amplitude during movement repetition (Agostino et al., 1992, 1994, 2003; Iansek et al., 2006; Lee et al., 2014; Postuma et al., 2015; Bologna et al., 2016b; Tinaz et al., 2016) (Table 2). Patients with Parkinson's disease take progressively longer times to perform sequential segments of a geometric sequence (Agostino et al., 1992, 1994; Lee et al., 2014). A progressive decrease in movement speed has also been documented during finger tapping (Agostino et al., 2003; Espay et al., 2009, 2011; Bologna et al., 2016a, 2018), the Purdue pegboard test (Kang et al., 2010, 2011), during writing, i.e. the so-called progressive micrographia (Wu et al., 2016), and walking (Iansek et al., 2006; Delval et al., 2017). When severe, the sequence effect can terminate with a freeze (i.e. motor block) (Giladi et al., 1992). The sequence effect may indeed be a prominent feature of the early stages of Parkinson's disease (Lee et al., 2014; Bologna et al., 2016a, b), and it tends to be less frequent in advanced Parkinson's disease (Bologna et al., 2016a, b). One possibility why the sequence effect is less frequent during arm movements in advanced patients is likely because in advanced patients the initial EMG burst is already of low amplitude (Bologna et al 2016a, b). During gait, a sequence effect is noticeable during a straight path, expressed as progressive reduction of the step length. A sequence can be terminated by a motor block (akinesia), which is clinically defined as freezing of gait (Iansek et al., 2016). Alternatively, it can be transient as the patient restores the ability to generate step length, which often results in another sequence ('oscillatory variability') (Fasano and Bloem, 2013). Rarely the sequence effect does not cease, and it is not interrupted by a motor block, as seen in festination. The phenomenon of sequence effect during straight walking has been interpreted as a maladaptive motor behaviour, whereby the brain tries to compensate for gait asymmetry in the presence of defective inter-limb coordination and amplitude generation (Fasano et al., 2016). In PSP, Ling et al. (2012) found that repetitive finger tapping, as assessed by a 3D motion analysis system, were 'hypokinetic', i.e. characterized by low amplitude movement, with no decrement in amplitude as movements continued. Djuric-Jovicic and colleagues (2016) confirmed the lack of progressive reduction in amplitude during the finger tapping in patients with PSP-Richardson syndrome compared to Parkinson's disease or MSA patients (Table 3). It has been argued that atypical parkinsonisms as well as advanced Parkinson's disease have a too severe impairment of amplitude generation that further decrements (i.e. the sequence effect) are not possible. Our own observation, however, is that in some cases of PSP the sequence effect can be present.

Although there are no systematic studies investigating the effect of levodopa administration in Parkinson's disease on each motor abnormality, it seems that levodopa improves but does not normalize the abnormal movement parameters. Levodopa improves the speed and amplitude of movement by modifying the amplitude and temporal scaling of the agonist and antagonist bursting pattern (Baroni et al., 1984; Berardelli et al., 2001; Vaillancourt et al., 2004; Espay et al., 2009, 2011; Suppa et al., 2017). However, the improvement of kinematic measures is lower in subjects markedly hypokinetic (Espay et al., 2011). Moreover, levodopa does not alleviate the sequence effect (Kang et al., 2010; Lee et al., 2014; Wu et al., 2016; Suppa et al., 2017). Further evidence of the peculiarity of the sequence effect and of its responsiveness to levodopa came from a study on micrographia in Parkinson's disease (Wu et al., 2016). Patients were divided in two groups, the first showing consistent micrographia, i.e. a global reduction but without significant progressive reduction in writing size, the second presenting progressive micrographia, i.e. a gradual reduction in size during writing. While consistent micrographia can be considered a manifestation of hypokinesia (smallness of movement), progressive micrographia is a manifestation of the sequence effect. Patients underwent functional MRI, revealing different neural correlates of the two types of micrographia. Dysfunction of the basal ganglia motor circuit contributed to consistent micrographia, whereas dysfunction of the basal ganglia motor circuit plus disconnections among motor cortical areas and the cerebellum was likely involved in progressive micrographia. Consistent but not progressive micrographia improved with the treatment (Wu et al., 2016). Additional characterization of the sequence effect on Parkinson's disease has been provided by Tinaz et al. (2016), who approached the sequence effect as a central problem of motor energy. They tested Parkinson's disease patients performing a dynamic isometric task with or without a visual feedback in two conditions (on and off therapy) and compared them with healthy subjects. They confirmed the poor effect of levodopa on the sequence effect, pointing to the possible involvement of other systems and neurotransmitters in generating the sequence effect. Second, they pointed out that energetic cost of performance in the first 15 s of the sequence was significantly higher in Parkinson's disease. Finally, visual feedback, which provided an external reference, improved the performance in both groups. The sequence effect may then reflect the difficulty in sustaining motor performance when the required effort has to be motivated and generated internally (Tinaz et al., 2016).

Although limb bradykinesia must be documented to establish a diagnosis of Parkinson's disease, bradykinesia also occurs in the face, voice and axial/gait domains (Postuma et al., 2015, 2018). At facial level, one of the most striking features of Parkinson's disease is hypomimia, ranging from minimal masked face and spontaneous blink rate reduction to lower face involvement and lips parted when the mouth is at rest (Goetz et al., 2008). Spontaneous blink rate reduction has consistently been demonstrated in Parkinson's disease, and it has been demonstrated that it is strictly associated with low central dopaminergic activity (Karson et al., 1984; Deuschl and Goddemeier, 1998; Kimber and Thompson, 2000; Altiparmak et al., 2006; Korosec et al., 2006; Agostino et al., 2008; Bologna et al., 2013). Studies based on voluntary facial movements, however, showed normal velocity and amplitude of blinking, although the duration of the inter-phase pause was longer in Parkinson's disease than in control subjects (Korosec et al., 2006; Agostino et al., 2008; Bologna et al., 2012). A reduction in spontaneous lower face (perioral) movements has been observed in Parkinson's disease (Deuschl and Goddemeier, 1998). Different from upper face, reduced velocity and amplitude has been documented in the lower face during repetitive syllable productions or spontaneous and posed smile (Caligiuri, 1987; Katsikitis and Pilowsky, 1988; Connor et al., 1989; Jacobs et al., 1995; Smith et al., 1996; Simons et al., 2004; Marsili et al., 2014; Bologna et al., 2016c). Also, although levodopa improves spontaneous blinking, it seems to have negligible effects on voluntary movements of both the upper and lower face in Parkinson's disease (Bologna et al., 2013; Suppa et al., 2017) (Table 4). In contrast, DBS of subthalamic nucleus (STN) has been found to increase beyond normal the velocity and amplitude of voluntary blinking while prolonging the duration of the inter-phase pause (which was reduced by the concomitant administration of levodopa) (Bologna et al., 2012). Recently, it has been pointed out that hypomimia is highly associated with the likelihood of striatal dopaminergic denervation more than limb bradykinesia (Mäkinen et al., 2019).

Loss of spontaneous facial expression has been consistently reported in patients with PSP and MSA (Romano and Colosimo, 2001; Tison et al., 2002; Fabbrini et al., 2009). The spontaneous blink rate is markedly reduced in atypical parkinsonisms, particularly in PSP (Altiparmak et al., 2006; Bologna et al., 2009, 2013, 2014). In PSP, reduced velocity and amplitudes have been documented during voluntary movements of both the upper and lower face (Bologna et al., 2013) (Table 3).

In summary, earlier neurophysiological studies in Parkinson's disease patients focused on upper limb bradykinesia and documented abnormalities of movement preparation and execution, resulting in abnormal prolongation of simple and choice reaction times and movement slowness. Subsequent studies have documented the sequence effect and its variable presence or association with other movement abnormalities in Parkinson's disease and atypical parkinsonisms. Finally, an increasing number of studies now document bradykinesia features in the face, voice, and axial/gait domains. Because some of these abnormalities (i.e. hypomimia) may be associated with a higher likelihood of striatal dopaminergic denervation compared to limb bradykinesia (Mäkinen et al., 2019), the assumption that limb bradykinesia must be documented to establish a diagnosis of Parkinson's disease (Postuma et al., 2015, 2018) is questionable.