اثر بیماری پارکینسون در برآورد زمان به عنوان یک تابع از محدوده طول مدت محرک و روش

The present research sought to investigate the role of the basal ganglia in timing of sub- and supra-second intervals via an examination of the ability of people with Parkinson’s disease (PD) to make temporal judgments in two ranges, 100–500 ms, and 1–5 s. Eighteen non-demented medicated patients with PD were compared with 14 matched controls on a duration-bisection task in which participants were required to discriminate auditory and visual signal durations within each time range. Results showed that patients with PD exhibited more variable duration judgments across both signal modality and duration range than controls, although closer analyses confirmed a timing deficit in the longer duration range only. The findings presented here suggest the bisection procedure may be a useful tool in identifying timing impairments in PD and, more generally, reaffirm the hypothesised role of the basal ganglia in temporal perception at the level of the attentionally mediated internal clock as well as memory retrieval and/or decision-making processes.

In recent years, there has been increasing interest in the neural mechanisms subserving the processing of time, and, in particular, the question of whether the neural substrates and circuitry involved in the temporal processing of intervals of very brief duration (milliseconds) differ from those underlying longer timing intervals (seconds-to-minutes range; Ivry and Keele, 1989, Ivry and Spencer, 2004, Lewis and Miall, 2003b and Rammsayer, 1997). One enduring issue concerns the contribution of frontal-striatal circuits to timing and time perception across these different time intervals (for reviews see Ivry and Spencer, 2004 and Meck and Benson, 2002). Timing research has been heavily influenced by the prominent information-processing model, the Scalar Timing Theory (Gibbon, 1977 and Gibbon et al., 1984). The model entails three distinct stages in which temporal information about an event is abstracted, encoded and acted upon (clock, memory, and decision). The essence of the model follows thus: an internal pacemaker with an attention mediated switch emits pulses stored by an accumulator to form an expressed interval of subjective time that approximately corresponds to objective (‘real’) time. After the designated passage of time has lapsed, the sum value of these accumulated pulses is stored in reference memory which frees the counting procedure to begin again. Expected time values, or pulse totals, are accrued in memory over many trials and inherent pulse values are placed upon certain corresponding events. Any subsequent interval of time discrimination is then compared to existing remembered time references and matched accordingly to the targeted event (Gibbon, 1977 and Gibbon et al., 1984). According to this model, each stage is independent and capable of selective readjustment or realignment, and alterations to each stage of temporal processing give rise to specific patterns of accuracy and variability in duration judgments (Meck, 1983 and Meck, 1996). In its simplest form, the scalar property holds that errors of time estimation are strictly proportional to the target times being estimated. This scalar variability can reflect changes in either the clock, the memory, or decision stages. Non-scalar variability (where time estimation errors are disproportionate to target times), however, is assumed to arise from dysfunctional memory encoding or decoding (Meck, 1983 and Meck, 1996). A central role for the basal ganglia in the temporal processing of intervals has been strongly implicated by animal research primarily focussed on the internal clock and memory stages of the model. Meck (1983) demonstrated that clock speed and memory stages can be dissociated, both with respect to the brain circuitry that underlies them, and in the time course of their responsiveness to pharmacological manipulations. Whereas administration of cholinergic drugs in rats previously trained to estimate a target duration appeared to modify time representations in reference memory (Meck & Church, 1987), the administration of dopaminergic drugs was shown to alter rats’ internal clock speeds, with dopamine antagonists (e.g., haloperidol) typically leading to a ‘slow clock’, as manifested by underestimations of time intervals (i.e., perceive less time as having elapsed than has actually passed) and overestimations when required to reproduce the timespan, and dopamine agonists typically having the opposite effect (Meck, 1983 and Meck, 1996; but see Harper, Bizo, & Peters, 2006). Studies with human subjects have also reported impaired temporal duration discrimination after the administration of the dopamine antagonist haloperidol, indicating that temporal information processing in humans depends on central dopaminergic activity (Lustig and Meck, 2005, Meck, 1996, Rammsayer, 1997 and Rammsayer, 1999). The fact that dopaminergic systems thought to be closely associated with aspects of interval timing are impaired by Parkinson’s disease, make this disorder of special interest. Parkinson’s disease (PD), a common neurological disorder affecting some 1% of people over the age of 50, is perhaps the most widely studied human model of basal ganglia dysfunction (Saint-Cyr, 2003). Pathological features include marked degeneration and atrophy of the substantia nigra and a consequent major reduction of the dopaminergic projection to the striatum (Agid, Javoy-Agid, & Ruberg, 1987). The resulting dysfunction of striatal circuits is expressed in PD by motor and cognitive symptoms, including the disordered timing of movement, usually manifesting in the form of bradykinesia and/or akinesia (Barbosa et al., 1997 and Morris, 2000), and the lengthening of normal information processing time that parallels patient bradykinesia, referred to as bradyphrenia (Rogers, Lees, Smith, Trimble, & Stern, 1987). Cognitive changes in PD may also be closely linked to prefrontal dysfunction with an emergence of neuropsychological impairments in the form of working memory impairments and attentional deficits (Owen et al., 1992 and Taylor et al., 1986). Empirical findings in PD research have tended to reveal a range of specific motor-timing deficits in PD, including increased reaction time (RT) and movement time (Bloxham, Dick, & Moore, 1987), reduced ability to maintain fixed rhythms in tapping tasks (Elsinger et al., 2003 and O’Boyle et al., 1996), and impaired speech time processing (Breitenstein, Van Lancker, Daum, & Waters, 2001). Dopaminergic-related slowing in PD has also been related to a disturbance in the perception and production of a range of time intervals (for a review, see Meck & Benson, 2002). Time perception studies offer the advantage that the experimental tasks usually have very little or no motor requirements, so that, at least on the surface of it, they are likely to be relatively pure reflections of timing processes. These studies typically use a variety of different methods including verbal time estimation tasks, in which participants are asked to estimate the duration of a given time period (e.g., the length of time a stimulus is presented); duration production tasks, which requires participants to produce a target duration (e.g., press a button after an experienced duration of 30 seconds); and reproduction tasks, in which participants have to evaluate and reproduce a target duration by comparing time during an evaluation phase with one elapsing during the reproduction phase. Nevertheless, there is considerable controversy about whether the timing processes impaired by the disrupted nigrostriatal dopaminergic system in PD affect temporal processing for both short and long durations (Ivry and Spencer, 2004, Rammsayer, 1997 and Rammsayer, 1999) and whether such deficits reflect a failure of the basal ganglia to perform its timekeeping functions (i.e., a disturbance in the internal clock speed of the scalar model) as opposed to impairment of other more cognitively mediated processes (i.e., disruption to memory and/or decision stages of the scalar model). A large number of studies of people with Parkinson’s disease, many conducted within the framework of scalar timing theory, have almost uniformly reported temporal processing difficulties for intervals within the seconds-to-minutes range (Lange et al., 1995, Malapani et al., 2002, Malapani et al., 1998, Pastor et al., 1992, Perbal et al., 2005 and Riesen and Schnider, 2001), leaving little doubt that frontal-striatal circuits are crucially involved in time estimation of longer intervals. The precise nature of this deficiency, however, remains less obvious. For instance, two early studies showed that compared with controls, non-medicated PD patients underestimated durations of different lengths in verbal estimation tasks and overestimated the same durations in reproduction tasks, with the greater magnitude of overestimation occurring at longer interval (scalar variability) (Lange et al., 1995 and Pastor et al., 1992). Both sets of authors interpreted the results as evidence that the internal clock was running at a slower rate in patients than control participants. Notably, in both these studies, the administration of dopaminergic replacement therapy to PD patients led to time judgments and estimations that were similar to those of control subjects, adding weight to the hypothesised role of neostriatal dopamine in the modulation of the internal clock. More recently, using a peak-interval procedure, Malapani and colleagues (1998) also observed Parkinsonian deficits in timing experiments in which participants were trained to learn two target durations although again only when patients were off dopaminerigic replacement therapy. In this instance, however, non-medicated patients tended to overestimate the shorter interval (8 s) and underestimate the longer one (21 s). This non-scalar variability was termed a “migration effect” and, according to the authors, was indicative of memory dysfunction rather than a change in clock speed. Interestingly, when trained on the longer interval only, patients overestimated the interval duration, consistent with the hypothesis of a slower internal clock in non-medicated PD patients. A subsequent investigation by the same group that fully manipulated the medication status of patients across training and test phases revealed that the migration effect occurs only when patients are tested off medication (regardless of whether they have been trained in an “on” or “off” state), indicating that the migration of two time intervals in Parkinson’s disease may stem from memory retrieval difficulties, a consequence of dysfunction of the decoding system in comparison with the currently elapsing time (Malapani et al., 2002). Interestingly, a recent study reported a similar migration effect in Parkinson’s disease patients, but only for patients with left hemi-PD implicating a critical role for the right basal ganglia in memory retrieval in time reproduction in the range of seconds (Koch, Brusa, Oliveri, Stanzione, & Caltagirone, 2005). In an attempt to disentangle mnemonic deficiencies from changes in the rate of the internal clock, Perbal et al. (2005) compared time performance of medicated PD and control participants on a memory-dependent reproduction task and a duration production task assumed to reflect changes in the speed of internal time-keeping mechanisms. Perbal and colleagues reported that PD patients’ performance in the reproduction task was more variable than controls and this variability was related to both measures of memory and disease severity. In contrast to the hypothesised slower internal clock rate in PD, patients’ duration judgments in the production task were significantly shorter than those of controls, the authors attributing this to the dopaminergic restorative effect of anti-PD medication. In summary, then, recent PD studies have suggested that PD patients do make inaccurate and variable duration judgments in the seconds-to-minutes range. Depending on the precise timing task employed (verbal estimation, production, or reproduction), and, more specifically, whether PD patients are required to encode/decode one or two temporal memories, the observed timing deficits stem from changes occurring either at the internal clock stage or at the memory retrieval stage of temporal processing. Whether or not the basal ganglia are involved in timing of short intervals is less clear. In an early PD study, Ivry and Keele (1989) did not detect any impairment in medicated patients when they made judgments about the relative duration of two brief tone intervals. In contrast, patients with cerebellar lesions evidenced a clear deficit in perception on the same task, leading the authors to hypothesise that the basal ganglia is only crucially involved in temporal perception of durations in the second-to-minutes range (see also Ivry & Spencer, 2004). A number of subsequent studies, however, have implicated a role for the basal ganglia and nigrostriatal dopaminergic pathways in temporal processing of very brief durations. Artieda, Pastor, Lacruz, and Obeso (1992) tested PD patients’ ability to discriminate pairs of short duration stimuli separated by periods ranging from 1 to 200 ms in auditory, visual, and somaesthetic modalities and found impairments for all three modalities when patients were withdrawn from medication. The authors reported a significant improvement in temporal discrimination performance after receiving dopaminergic replacement therapy although it is unclear if performance returned to normal levels. In line with this finding, Rammsayer and Classen (1997) reported pronounced performance decrements in PD patients’ temporal discriminations of very brief auditory stimuli (50–98 ms), impairments that were strongly associated with dosage of dopaminergic replacement therapy, and explained these findings by suggesting dopamine depletion in PD leads to the slowing of the internal clock. Interestingly, Harrington, Haaland, and Hermanowicz (1998) reported that a medicated PD group was impaired on a duration perception task in which participants judged the relative duration of the interval between a pair of tones that was either 300 or 600 ms. Another study also reported a deficit in brief duration discrimination for auditory stimuli, although the impairment was limited to female patients with Parkinson’s disease (Hellstrom, Lang, Portin, & Rinne, 1997). Although not unequivocal, taken as a whole, these findings suggest that basal ganglia circuitry may also be involved in processing temporal information in the range of milliseconds. However, unlike temporal processing of time intervals in the range of seconds or more, timing of sub-second intervals is assumed to be relatively automatic process and, to a large extent, beyond cognitive control. Therefore, observed discrepancies in PD timing for brief durations are unlikely to be a result of memory dysfunction commonly observed in PD and, within a scalar model of time at least, likely to reflect changes (decrease) in the speed of the internal clock mechanism (Rammsayer, 1997 and Rammsayer, 1999). Despite an accumulating volume of work focussed on investigating the neural mechanisms involved in timing across different stimulus durations, only a few relevant studies have directly compared short and long duration timing discrimination tasks in a single procedure. In two pharmacological studies of temporal information processing and memory in healthy subjects, Rammsayer, 1997 and Rammsayer, 1999 observed that whereas the administration of haloperidol (a dopamine antagonist that effects multiple dopaminergic brain systems) led to impaired temporal discrimination of both short and long durations, administration of remoxipride (a dopamine anatagonist that blocks D2 receptors in the mesolimbocortical dopaminergic system only, see Awad, Lapierre, & Jostell, 1990) disrupted time estimation for durations in the range of seconds only. Rammseyer attributed the deleterious effects of haloperidol and remoxipride in the long duration condition to a drug-induced impairment of working memory, but suggested the selective impairment of timing in the short duration ranges caused by haloperidol was consistent with the specialised role of basal ganglia (or, more specifically, the mesostriatal dopaminergic system) in processing temporal information in the range of milliseconds. Interestingly, however, although significant striatal activation has been reported in the temporal discrimination of both short and long duration ranges of healthy subjects across (separate) brain imaging studies (Ferrandez et al., 2003, Nenadic et al., 2003 and Rao et al., 2001), the one fMRI study that has directly compared activation in the basal ganglia in short (600 ms) and long (3 s) visual stimulus durations observed no significant activation in either condition (Lewis & Miall, 2003a). Only one study of people with PD has directly compared the temporal discrimination performance of short and long intervals. In this study, Riesen and Schnider (2001) reported that medicated PD patients made more errors than controls in the discrimination of short visual stimulus durations ranging from 200 ms to 2 s, but performed normally on a verbal estimation task for durations ranging from 12 to 48 s. The authors explained the findings by pointing towards the differential cognitive demands of each of the two duration tasks and suggested the PD deficit in the short stimulus duration condition arose not from impaired timing per se, but rather was related to the requirement to simultaneously process serial stimuli in the visual stimulus discrimination task, symptomatic of patients’ impaired working memory and/or (divided) attentional difficulties. Reisen and Schnider argued that by contrast, attentional demands in the long duration task were less and preserved long-term memory allowed the group of PD patients to accurately estimate longer durations. Both within and across the reviewed studies, then, the comparison of PD timing performance in short and long intervals is complicated by the fact that the tasks employed were often very diverse with respect to range of procedural and methodological factors, including response requirements, the extent to which performance was cognitively mediated, the stimulus modality used, whether the scalar model was used as a methodological structure for measurement, and the administration of medication in the patient samples. What is required is an approach by which the effect of duration range can be systematically examined in a single experiment using the same procedural task. The goal of this research was to examine the involvement of the basal ganglia in duration perception as a function of range (milliseconds vs multiseconds) and modality (visual vs auditory). The examination of short and long interval timing in a single group of PD patients, where brain structures assumed to be involved in some aspects of temporal processing are damaged, provides a further opportunity to investigate the possibility that these forms of timing may be neurally independent. Furthermore, comparisons of signal modality in timing experiments can be very informative given their potential to reveal associations between impaired temporal processing and mnemonic/attentional mechanisms (Melgire et al., 2005 and Penney et al., 2005). For this purpose, the present study employed a duration-bisection procedure to examine interval timing (Droit-Volet and Wearden, 2002 and Meck, 1996). Briefly, this experimental paradigm entails participants perform a training phase, in which participants are taught to discriminate between probe signals of a short and long duration, and a subsequent test phase, in which participants are required to classify probe signals as short or long relative to the anchor durations learned in training. Although the probe signals in the test phase can be the same as the anchor durations participants were exposed to in training, the majority are of intermediate duration. Based on previous findings, we expected patients with Parkinson’s disease to evidence a deficit in time perception for both visual and auditory stimuli, with the greater impairment for stimuli in the supra-second range, given the hypothesised involvement of the basal ganglia in both the timekeeping and mnemonic stages of temporal processing.