When performing explicit timing tasks, we make an explicit use of temporal information (e.g., estimates of the duration of different stimuli or of their inter-stimulus intervals [ISI]) in order to represent precise temporal durations through a sustained or periodic motor act (Coull and Nobre, 2008). Functional magnetic resonance imaging (fMRI) (Bengtsson et al., 2004, 2005; Lewis et al., 2004; Jahanshahi et al., 2006; Bueti et al., 2008) and non-invasive brain stimulation (Koch et al., 2003; Dusek et al., 2011; Vicario et al., 2013) studies of motor timing have consistently identified brain regions crucial to time processing, in particular supplementary motor area (SMA), basal ganglia (BG), cerebellum and right-inferior frontal and parietal cortices.

The synchronization-continuation paradigm has been frequently used to study the neural substrates of explicit motor timing. This paradigm allows for the study of both externally triggered and internally triggered movements as it involves: (i) a synchronization phase, in which subjects are asked to tap in synchrony with a train of tones separated by a constant ISI, and (ii) a continuation phase, in which subjects are requested to continue tapping at the previous rate in the absence of the auditory cue. Using fMRI paradigms, Rao et al. (1997) showed that the continuation (internally triggered) phase, but not the synchronization (externally triggered) phase, activates the SMA, the left-caudal putamen, and the left ventrolateral thalamus. These findings support the existence of a hypothetical “internal clock” that beats the rhythm when the movement is internally generated (Pastor et al., 1992b; Meck and Benson, 2002; François-Brosseau et al., 2009), and involves the sensorimotor circuit of the BG (Figure 1). Indeed, it has been proposed that several parallel, segregated loops through the BG connect back to distinct cortical areas (namely motor, associative, and limbic network) (Alexander et al., 1986). The sensorimotor loop connects the motor areas in the cortex with the caudolateral striatum territories playing a role in stimulus-response habitual control (Redgrave et al., 2010).

The role of the BG in timing was further substantiated by demonstrating that they are activated during motor, but also perceptual, time processing tasks (Bueti et al., 2008) involving both sub-second and supra-second temporal intervals (Jahanshahi et al., 2006). Noteworthy, the associative basal ganglia circuit is particularly active during externally triggered tasks (Figure 1).

The cerebellum, instead, is more often activated (i) during motor than perceptual explicit timing tasks (Bengtsson et al., 2005; Jahanshahi et al., 2006; Bueti et al., 2008), (ii) when a synchronization to an external rhythm is required (Del Olmo et al., 2007), and (iii) when processing involves sub-second rather than supra-second intervals (Lewis and Miall, 2003a,b).

Timing is implicit when temporal information can optimize performance on a non-temporal task. The processing of time-dependent features of a movement is crucial in predicting whether the outcome of the movement will be consistent with its ultimate goal, or will result in an execution error. When temporal information is processed to predict the outcome of self-executed or externally-driven movements, cerebellar outflow pathways are primarily involved. Indeed, cerebellar networks optimize self-executed actions by recalibrating predictions, capturing the sensory consequences of the same actions (Bell et al., 2008; Synofzik et al., 2008; Izawa et al., 2012). When subjects use temporal information inherent to the spatial-temporal trajectory of a dynamic visual stimulus to predict its final position, fMRI studies revealed activation in the left inferior parietal cortex (Assmus et al., 2005), sensorimotor regions of the premotor and parietal cortices (Field and Wann, 2005), and cerebellum (O'Reilly et al., 2008). We recently showed that the cerebellum is engaged when temporal information is processed to predict the temporal outcome of a motor act (Avanzino et al., 2015a). We developed an ad hoc task in which participants were required to observe a movement in a video and then predict the end of the same movement (Avanzino et al., 2013, 2015a; Martino et al., 2015). Crucially, a few seconds after its onset, the video was darkened for a given time interval, thus the task could be performed only by extrapolating time-related features of observed motion sequence. By this task, we have shown that inhibiting the activity of the lateral cerebellum with 1 Hz-repetitive transcranial magnetic stimulation induced a deterioration of timing performance selectively when subjects were asked to estimate the duration of a body segment movement (handwriting) and not of a movement involving an inanimate object (Avanzino et al., 2015a). This finding may suggest that cerebellar sub-regions are specifically involved in processing temporal features of movements belonging to the human motor repertoire.

The notion that the mechanisms underlying time processing are abnormal in PD has been suggested by some pivotal observations. Pastor et al. (1992a,b) reported that the accuracy in timing of repetitive alternating wrist movements, time estimation and time reproduction on a verbal task were impaired in PD patients “off” medication, indicating timing mechanisms abnormalities in PD already at a perceptual level and possibly related to dysfunction of an “internal clock” that is modulated by dopamine.

Afterwards, impaired time reproduction was confirmed during paced finger tapping in PD (O'Boyle et al., 1996; Elsinger et al., 2003) and assessed in “on” and “off” conditions (Jones et al., 2008). Using the synchronization-continuation task to investigate both de novo and treated PD subjects, Jones et al. (2011) showed that treated PD patients tapped ahead of the beat at the 250 ms rate, respect to de novo and healthy subjects without any difference between “off” and “on” conditions. This observation, resembling the clinical phenomenon of festination that typically occurs in PD for fast repetitive movements, suggested an influence of either chronic pharmacological treatment or disease progression. Altogether, the continuation phase highlighted major differences likely reflecting the difficulties of PD subjects with internally generated movements.

More recently, using a modified version of the synchronization-continuation paradigm with a motor imagery task, we showed that PD patients exhibited a selective deficit in motor timing for internally triggered sequential movements separated by a supra-second interval, and that this deficit was better explained by a defect of motor planning (Avanzino et al., 2013). Bienkiewicz and Craig (2015) showed that abnormalities in sensorimotor synchronization in PD are not limited to finger tapping task but rather extend in the context of a beat interception task based on aiming movements. The type of task required prospective motor control (i.e., coupling movement to neural based dynamic information that helps anticipate when the beat is going to sound) suggesting that BG circuitry might undermine also the temporal prediction ability.

Overall, the motor timing abnormalities observed in PD indicate that the BG are mainly involved in explicit timing mechanisms (see also Martinu et al., 2012). Accordingly, a recent study showed that the implicit timing of salient events seems largely unaffected by PD (de Hemptinne et al., 2013). It is worthwhile to underline that the tight interaction between difficulties with cognitive timing and spatio-temporal control of movement emerges not only from experimental findings, but also from the clinical aspects of PD. Bradykinesia (slowness of movement initiation and execution) is particularly evident for internally generated sequential movements (McIntosh et al., 1997; Heremans et al., 2012), commonly occurs in gait and can manifest as slow shuffling strides, an accelerating gait, or highly variable and random stride times (Baltadjieva et al., 2006; Almeida et al., 2007).

Related to the neural network supporting motor timing abnormalities in PD, evidence demonstrated decreased activation within the sensorimotor cortex, cerebellum, and medial premotor system in PD patients compared to controls during paced finger tapping (Elsinger et al., 2003). Dopamine supplementation did not improve task performance, but partially “normalized” brain activation patterns. In a PET study PD patients presented a significantly enhanced activation of the cerebellum, thalamus and substantia nigra reticulata during a synchronization-continuation task, compared to healthy subjects (Jahanshahi et al., 2010). Moreover, pallidal over-activation and cortical regions under-activation were observed in PD subjects in the “off” state, but switching to the “on” state was associated with increased striato-frontal functional connectivity. This study and others (Husárová et al., 2013, 2014) supported the idea that motor timing networks are modulated by dopaminergic stimulation and that PD patients activate an alternative timing network relying more on cerebellar activation than healthy controls. Summarizing the behavioral and imaging evidence on the influence of dopaminergic medication on motor timing, so far results are mixed and therefore conclusions are unclear. The role of the BG oscillatory activity in motor timing has been evidenced in a group of 12 PD patients implanted for subthalamic nucleus deep brain stimulation (STN-DBS) (Wojtecki et al., 2011). Supra-second time reproduction performance was significantly improved during “on” stimulation (130-Hz) compared to no stimulation, whereas milliseconds timing was not affected.

Huntington's disease (HD) is a hereditary neurodegenerative disorder, caused by an expansion of CAG triplet repeats in the Huntingtin gene on chromosome 4. The monogenic etiology of HD allowed investigating timing in both symptomatic and pre-symptomatic individuals. Explicit timing abilities progressively deteriorate in mutation carriers as they approach clinical disease onset. Early-moderate HD patients exhibited marked irregularity of tapping rates on the synchronization-continuation paradigm (Freeman et al., 1996). In pre-symptomatic HD mutation carriers, reproduction of supra-second intervals accuracy decreases with estimated time to onset, Hinton et al. (2007) and Rowe et al. (2010). A self-paced variant of the same tapping task was also used to investigate the relationship between timing performance and disease progression. Performance on self-paced finger tapping was associated with disease progression (Michell et al., 2008; Tabrizi et al., 2009; Bechtel et al., 2010), showing high sensitivity in separating mutation carriers from non-carriers already from the earliest stages (Tabrizi et al., 2009; Bechtel et al., 2010). A similar pattern of timing deterioration was observed on non-rhythmic motor reproduction of supra-second intervals, which became less precise as disease onset approached and progressed further as disease advanced (Beste et al., 2007; Wild-Wall et al., 2008; Rao et al., 2014).

Functional magnetic resonance imaging (fMRI) confirmed that performance on paced finger tapping was associated with reduced neural activation in subcortical regions (putamen, caudate, thalamus) in pre-symptomatic HD mutation carriers closer to disease onset (Zimbelman et al., 2007). This subgroup of pre-symptomatic patients under-activated the SMA, left-anterior insula and right-inferior frontal gyrus during the finger tapping task. Conversely, those far from onset exhibited more efficient compensatory activation of other brain areas (sensorimotor cortex, medial frontal cortex and cerebellum) (Zimbelman et al., 2007).

Overall, the progressive involvement of cortico-basal ganglia sensorimotor and associative loops in HD patients is reflected by a similarly progressive timing deficit that is more selective for explicit, cognitive-controlled timing tasks rather than implicit timing tasks.

Dystonia is a network movement disorder associated with maladaptive plasticity of motor output coupled to disordered sensorimotor integration. Dysfunction in both cortico-basal ganglia and cerebello-thalamo-cortical networks is believed to give rise to the involuntary tonic and phasic muscle contractions that characterize dystonia. Timing has been assessed mainly in task-specific forms of dystonia, triggered by complex movements like musical playing or writing. Kinematic analyses of scales or finger tapping performed on a digital piano by pianists with dystonia showed inaccuracies in tone and interval duration and rhythmic inconsistency (Jabush et al., 2004; Furuya and Altenmüller, 2013). This may represent a mere consequence of the motor overflow during musical performance, since part of these measures improved with botulinum toxin therapy (Jabush et al., 2004; Furuya et al., 2014). Further, explicit timing performance in these patients appeared similar to control subjects (Van Der Steen et al., 2014), suggesting that direct extrapolation of temporal properties while performing timed movements may be preserved in dystonia. On the other hand, we showed that patients with either task-specific (writer's cramp) (Avanzino et al., 2013) or non-task-specific (cervical) dystonia (Martino et al., 2015) are less accurate in predicting the temporal outcome of a visually perceived movement, and that this is observed for human body, but not inanimate object, motion. This supports the view that patients with dystonia manifest subtle differences in extrapolating temporal properties during the perception of hand/arm movements. Such deficit may be linked to an abnormal internal model of motor commands reflecting dysfunction of cerebellar outflow pathways. The selective implicit timing task abnormalities support the notion that dystonia is a broader network disorder, in which the crucial nodes are located not only in the BG and the sensorimotor cortex, but also in the cerebellum.

Tourette's syndrome (TS) is associated with the abnormal maturation of cortico-basal ganglia loops. Timing abilities are underexplored in TS. In an early report, discrimination between different time intervals in untreated patients was comparable to healthy subjects (Goldstone and Lhamon, 1976). More recently Goudriaan et al. (2006) reported that time estimation and reproduction of time intervals spanning from 2 to 20 s did not differ between TS and healthy adults. However, these authors measured time intervals using a manual stopwatch, which is poorly accurate. A study by Vicario et al. (2010) has documented higher accuracy in the processing of supra-second intervals in TS children with lower tic severity, compared to age-matched healthy volunteers (Vicario et al., 2010). This result was explained by suggesting the existence of an effective compensatory modulation phenomenon, in correspondence of the prefrontal cortex, leading to better tic inhibition and increased precision in explicit processing of supra-second time intervals. Dopaminergic modulation seems also relevant to temporal discrimination in TS patients, as dopamine receptor blockers like pimozide may improve patients' variability on this task (Vicario et al., 2016).

Overall, TS has been investigated mainly for perceptual aspects of timing, revealing greater accuracy of temporal discrimination of supra-second intervals associated with lower tic severity. As already suggested for other behavioral abnormalities in TS (Mueller et al., 2006; Jackson et al., 2007), this behavioral gain appears to be more linked to adaptive mechanisms facilitating tic suppression rather than to core pathogenic mechanisms.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.