SIMPLE AND CHOICE REACTION TIME PERFORMANCE FOLLOWING UNILATERAL STRIATAL DOPAMINE DEPLETION IN THE RAT

Transcription

1 Brain (1991), 114, SIMPLE AND CHOICE REACTION TIME PERFORMANCE FOLLOWING UNILATERAL STRIATAL DOPAMINE DEPLETION IN THE RAT IMPAIRED MOTOR READINESS BUT PRESERVED RESPONSE PREPARATION by VERITY j. BROWN and TREVOR W. ROBBINS {From the Department of Experimental Psychology, University of Cambridge, Cambridge, UK) SUMMARY Rats were trained to perform a visual spatial discrimination, where stimulus luminance provided information regarding the required direction of response. The visual stimuli were presented either in advance of a temporally unpredictable auditory imperative stimulus (simple reaction time condition) or simultaneously with the imperative stimulus (choice reaction time condition). Following unilateral striatal dopamine depletion, produced by intracerebral infusion of 6-hydroxydopamine, there was not only a marked spatial response bias towards the side of the dopamine depletion but also an abolition of the delay-dependent speeding of reaction time that reflects motor readiness, on the side contralateral to the lesion. Nevertheless, the rats continued to show a benefit in performance of the simple reaction time condition compared with the choice reaction time condition, indicating that they were able to use advance information to select and prepare responses. The results are discussed in the context of differential deficits shown by patients with Parkinson's disease in simple versus choice reaction time performance, and of the functions of parallel corticostriatal loops subserving the normal control of action. INTRODUCTION Some of the motor dysfunction in Parkinson's disease has been attributed to deficits in the advance construction or sequencing of a 'motor program' (Bloxham et al., 1984, 1987) or the automatic execution of a learned motor plan (Marsden, 1982). This suggestion is partly based on evidence that patients with Parkinson's disease do not show lengthened reaction times in choice reaction time tasks, but they are impaired in simple reaction time performance, thus failing to show the same speeding of reaction time as healthy subjects when advance information is available (Flowers, 1978; Evarts et al, 1981; Bloxham etal., 1984, 1987; Sheridan etal., 1987; Pullman etai, 1988). The neural basis for the seemingly paradoxical nature of the differential impairment of simple versus choice reaction time in Parkinson's disease is not well understood, although it presumably has implications for understanding the neural substrates of normal, as well as abnormal, action. Thus it is unclear to what extent the deficit in simple reaction time is either dopaminergic or indeed striatal in nature. This is an important consideration, Correspondence to: Dr Verity J. Brown, Laboratory of Sensorimotor Research, National Eye Institute, NIH Building 10, Room 10C101, Bethesda, MD 20892, USA. Oxford University Press 1991

2 514 V. J. BROWN AND T. W. ROBBINS given that parkinsonian patients exhibit neurochemical deficits in noradrenergic, serotonergic and cholinergic innervations of the neocortex (Scatton et al., 1982; Agid et al., 1987) as well as dopaminergic depletion in prefrontal cortex (Scatton et al., 1982). Furthermore, a significant relationship has been found between simple reaction time performance and CSF levels of the noradrenergic metabolite, 3-methoxy-4-hydroxyphenylglycol, in parkinsonian patients (Stern et al., 1984). Pullman et al. (1988) have recently attempted to resolve the neurochemical nature of the reaction time deficit by studying the effects of levodopa treatment on simple and choice reaction time tasks in patients with Parkinson's disease. They concluded that, whereas choice reaction time performance was improved with treatment, simple reaction time was not {see also Rafal et al., 1984). On this basis, they further concluded that simple reaction time paradigms may be at least partially dependent on 'non-dopaminergic physiologic mechanisms' and, in particular, upon the integrity of the supplementary motor area (SMA) and its noradrenergic connections, whereas choice reaction time depends on components of a 'functional loop' comprising a caudatothalamoprefrontal cortical pathway (cf. Alexander et al., 1986) the function of which is modulated by striatal dopamine. However, there are problems in inferring the neural substrate of reaction time performance from the data of patients with Parkinson's disease in this way. First, it is clear that the SMA also forms a functional loop with the striatum which is similarly modulated by dopamine, in particular within the putamen (Alexander et al., 1986). Secondly, the latter striatal region exhibits the greatest degree of dopamine depletion in Parkinson's disease (Hornykiewicz, 1982; Garnett et al., 1984). Thus according to the Pullman et al. (1988) scheme, it is difficult to understand why simple reaction time performance should not depend on striatal dopamine. With these problems of interpretation in mind, it seems likely that neither the neural basis of reaction time deficits in Parkinson's disease nor the role of striatal dopamine in simple and choice reaction time tasks will be easily resolved from studies of human patients. The present experiment therefore examined the effects of selective depletion of striatal dopamine (using the neurotoxin, 6-hydroxydopamine (6-OHDA)) on performance in a paradigm allowing measurement of both simple and choice reaction time in the rat. A unilateral preparation was used which enabled each animal to serve as its own control and avoided the debilitating consequences of a bilateral lesion (Carli et al., 1989). Previous experiments have examined the effects of unilateral striatal dopamine depletion in the rat in reaction time paradigms (Carli et al., 1985, 1989; Brown and Robbins, 1989a). In the former two studies, choice reaction time procedures that varied in the degree of stimulus-response compatibility were used, and in the third study simple and choice reaction time were compared directly. In all three studies, unilateral striatal dopamine depletion did not differentially affect reaction time performance, slowing occurring in all cases for responses contralateral to the side of the lesion. However, in none of these experiments has the effect on precueing, and hence advance information, been specifically investigated. The issue of possible differential effects of Parkinson's disease on simple and choice reaction time, resulting from a failure to use advance information efficiently, may be of central importance for understanding the nature of the parkinsonian deficit. Thus the present experiment was designed to examine whether striatal dopamine depletion, per se, produced a differential impairment in the ability

3 REACTION TIME AND STRIATAL DOPAMINE 515 to use advance information to prepare a response in simple versus choice reaction time procedures. METHODS Subjects The subjects were 9 male Lister Hooded rats (Olac, Bicester), weighing g at the beginning of the experiment. They were housed in pairs in natural lighting and maintained at 90% of their freefeeding body weight throughout the experiment on 15 g per day of standard laboratory chow, provided after the daily experimental session. Apparatus and procedure The apparatus used has been described in detail previously {see fig. 1, Carli et at., 1989; Brown and Robbins, 1989a). The operant chamber had an array of response holes, monitored by photoelectric cells, with a stimulus light at the rear of each hole. Three holes were open to the animal; the animal was required to hold its nose still in the central of the three holes, whilst the stimuli were presented in the adjacent side holes. Each trial was initiated by the animal, by making a response in the central hole. The light in the central hole was illuminated until the animal made such a response and was extinguished to signal that the trial had begun. The rats performed a brightness discrimination, in which the stimuli were not spatially localized, but presented to both sides of the animal's head. Bright stimuli signalled that a response in the left side hole was correct, whilst dim stimuli signalled that a response in the right side hole was correct. SRT Lights J Tnnp, CRT Lights i t Sustained nose poke i -0.3 s- Variable interval; r 0.6, 0.9, 1.2 or 1.5 s Variable interval; 0.6, 0.9, 1.2 or 1.5 s r i RT i i i t Withdrawal r. i.i. n i MT i I 1 t Response O FIG. 1. Task requirements of the simple (SRT) and choice (CRT) reaction time tasks. The upper lines show the simple reaction time condition in which the lights signifying the direction of the required response (either bright, for a left response, or dim, for a right response) were presented at the beginning of a trial. The lights were extinguished 0.3 s before the presentation of the auditory imperative signal. The choice reaction time condition is represented in the lower lines. The informative lights and the imperative tone were presented simultaneously and both remained on until a response was initiated.

4 516 V. J. BROWN AND T. W. ROBBINS In order to compare the effects of advance information, or simple reaction time, with no advance information, or choice reaction time, a tone was used in addition to the visual stimuli. The tone was the imperative signal and occurred after an unpredictable delay, staying on until the animal responded by withdrawing its nose from the central hole. The delay before the tone was 0.6, 0.9, 1.2 or 1.5 s. The lights specifying direction of responses were presented either simultaneously with the tone (choice reaction time condition) or presented as soon as the rat responded in the central hole, at the beginning of the variable delay, and were extinguished 0.3 s before the tone (simple reaction time condition). These task requirements are shown in fig. 1. In any one session, all the trials were either the simple or the choice condition. The rats were tested for 8 consecutive days, in the simple and choice reaction time conditions on alternate days, to obtain the preoperative baseline values, before receiving 6-OHDA lesions. Five rats were lesioned on the right and 4 rats were lesioned on the left. After a period of postoperative recovery lasting 4 days, they were tested for 8 consecutive days, again alternating the simple and choice reaction time conditions. They were tested again 30 and 60 days postoperatively, to examine the extent of recovery. Design and analysis The same rats performed both the simple and the choice variants of the reaction time task. Unilateral, rather than bilateral lesions were employed in order to compare performance on the lesioned and the unlesioned side in the same animal. The measures used were: percentage correct; time to initiate responses (reaction time) from the onset of the stimuli to the withdrawal of the head from the central hole; time to complete responses (movement time) from the withdrawal of the head from the central hole to the response in the side hole; spatial response bias (bias), a measure calculated as number of responses to the ipsilateral side divided by total number of responses made; frequency of responses made during periods of timeout, following incorrect or premature responses. Data were analysed by analysis of covariance (ANCOVA), with preoperative baseline values used as the covariate. The rationale for this was to test for a significant departure in the pattern of postoperative data compared with the preoperative pattern. Thus where the effects were not significant in the ANCOVA they could be assumed not to differ from the preoperative state, whilst significant effects indicated a departure in the pattern of data from the preoperative baseline. Polynomial regression analysis was performed on the reaction time data to test the significance of linear and other components, especially of the delay dependent effects. Significant interactions were further analysed by post-hoc Neuman-Keuls comparisons. Transformations were applied to percentage and reaction time data (angular and log ]0, respectively) to normalize the distributions in accordance with the ANOVA model (Winer, 1971). Surgery Following training to stable performance, the rats received unilateral lesions of the striatum by intrastriatal infusion of the neurotoxin 6-hydroxydopamine (6-OHDA). The rats were treated with 50 mg/kg of pargyline hydrochloride (Sigma Chemical Co.) dissolved in 0.9% sterile saline, injected i.p. 30 min before surgery, to enhance the effect of 6-OHDA (Breese and Traylor, 1971). Surgery was performed under Equithesin anaesthesia (0.3 ml/100 g, i.p.) in clean conditions. The animal was placed in a Kopf stereotaxic instrument, fitted with atraumatic earbars, and injected unilaterally with 6-OHDA (6-OHDA hydrobromide, Sigma Chemical Co.) freshly dissolved in 0.1 mg/ml ascorbic acid in 0.9% saline; 2 /xl of vehicle, containing 8 fig of 6-OHDA (base), were infused over 4 min via a 30-gauge stainless steel cannula, attached to a Harvard infusion pump with PP10 polythene tubing. The cannula was left in place for 2 min before and after the infusion. The coordinates used were 2.0 mm anterior to the bregma, ±3.5 mm from the midline and -5.5 mm from the cortical surface. Neurochemical assay Six weeks after surgery, the rats were killed under ether anaesthesia, the brains rapidly removed, and the tissue dissected on ice as previously described (Carli et ai, 1989) into the nucleus accumbens and the anterior caudate-putamen (taken from the same 2 mm thick slice). The posterior caudate-putamen and the tail of caudate were taken from the next 2 posterior 2 mm slices, respectively. The tissue was stored at 70 C until assay. Tissue concentrations of dopamine were measured by high performance liquid chromatography with

5 REACTION TIME AND STRIATAL DOPAMINE 517 electrochemical detection (HPLC-EC), based on the method described by Mefford (1981) and described in detail in Carli et al. (1989). RESULTS Neurochemical results The Table shows the results of the neurochemical assay of DA concentrations. There was profound (>85%) depletion of dopamine throughout the anterior, posterior and tail portions of the caudate-putamen, with mean depletions of less than 50% in the nucleus accumbens. TABLE, MEAN AND SEM OF CONCENTRATIONS OF DOPAMINE IN THE STRIATUM EXPRESSED AS ng/mg OF WET TISSUE, IPSILATERAL AND CONTRALATERAL TO THE SIDE OF INFUSION OF THE NEUROTOXIN, 6-OHDA, AND MEAN AND SEM OF PERCENTAGE DEPLETIONS OF DOPAMINE IN THE CONTRALATERAL STRIATUM Concentrations Mean % Area Ipsi Contra depletion Anterior caudate 1.60(1.3) 13.50(1.6) 87.8(10.5) Posterior caudate 1.10(1.0) 13.70(2.7) 91.1(10.0) Tail of caudate 1.10(1.5) 8.40(3.7) 84.6(18.4) Nucleus accumbens 4.65(2.3) 8.50(1.7) 48.8(25.5) Behavioural results Reaction time performance. Fig. 2 shows simple and choice reaction time, preoperatively and postoperatively, of responses to each side, plotted as a function of delay. Reaction times decrease as a function of delay preoperatively (F(3,21) = 26.14, P < 0.001), with faster reaction times at the longer delays due to the increasing probability of the occurrence of the stimulus as the delay elapses. This result was confirmed by finding a highly significant linear regression component (F( 1,21) = 75.54, P < 0.01). In addition to the delay-dependent speeding of reaction time, reaction time was also faster in the simple, as compared with the choice reaction time condition. This was evident as a Condition by Delay interaction (F(3,21) = 4.81, P = 0.011) which, when further analysed by post hoc Newman-Keuls comparisons, was found to result from there being a significant benefit to reaction time of the simple over choice reaction time condition at the two longest delays (1.2 and 1.5 s). There was a near significant difference at the delay of 0.9 s. Only at the shortest delay was there no significant difference between simple and choice reaction time. The Condition by Delay interaction had a highly significant linear (F(l,21) = 11.48, P = 0.003) but not quadratic (F(l,21) = 2.95, P ) or other (F(l,21) = 1.0, n.s.) component indicating that the two conditions mainly differed in the slope of the linear regression. The faster reaction time in the simple, as compared with the choice reaction time condition and the effect of delay remained unchanged postoperatively (F(3,21) = 0.75, n.s.). In particular, the crucial

6 518 V. J. BROWN AND T. W. ROBBINS Simple RT O IpsilateraJ Contralateral Choice RT A Ipsilateral A Contralateral \ c 0.40 o 2SED Delay (s) Preoperative Delay (s) Postoperative FIG. 2. Reaction time of responses to the ipsilateral and contralateral side in the simple and choice reaction time conditions, plotted as a function of delay. Simple reaction time (circles) was faster than choice reaction time (triangles) both preoperatively and postoperatively. Reaction time decreased as a linear function of delay preoperatively in both simple and choice conditions and on both sides. This effect was abolished on the side contralateral to the infusion of 6-OHDA (filled symbols). The vertical bar shows two standard errors of the difference of the means (2 SEDs) associated with the interaction between Surgery, Side and Delay. comparison, the postoperative interaction between Condition (simple versus choice reaction time) and Side of response (ipsilateral or contralateral to the lesion), which would have indicated a differential impairment in either simple or choice reaction time on the lesioned side, was not significant (F(l,7) = 0.98, n.s.). Independent of the advantage of simple over choice reaction time at the longer delays, postoperatively there was an interaction between the factors of Side of response and Delay (F(3,19) = 3.64, P = 0.03). Post hoc Newman-Keuls comparisons revealed that at the shortest delay, ipsilateral and contralateral reaction times were not different from each other. At the delay of 0.9 s, the difference approached significance and at the longer delays of 1.2 and 1.5 s, contralateral reaction time was significantly slower than ipsilateral reaction time. From fig. 2 it is clear that whilst ipsilateral reaction time continues to show a linear reduction as delay elapses, contralateral reaction time no longer shows the effect of the increasing probability of the stimulus. Post hoc Newman-Keuls comparisons revealed that ipsilateral reaction time was significantly faster at the longer delays of 1.2 and 1.5 s, compared with the 0.6 s delay, but contralateral reaction time was not significantly different at delays of 0.6 and 1.5 s. Polynomial regression analysis confirmed a significant linear (but no other) component for the Side by Delay interaction (F(l,19) = 9.21, P = 0.007). There was no significant difference between the reaction times of correct and incorrect responses (F(l,8) = 0.36, P = 0.57). A significant Side of response (ipsilateral or contralateral) by Response Outcome (correct versus incorrect) interaction (F(l,8) = 6.57, P = 0.037) was due to a difference in reaction time between ipsilateral and contralateral correct responses (0.32 as compared with 0.47 s, respectively), whereas no such

7 REACTION TIME AND STRIATAL DOPAMINE 519 difference was found for incorrect responses, which fell midway between the two and did not differ between sides (ipsilateral, 0.39 s and contralateral, 0.43 s). Response accuracy. Fig. 3 shows the percentage correct of responses made in the simple and the choice conditions preoperatively and postoperatively. There was a greater accuracy of performance in the simple reaction time condition which was confirmed by a significant main effect of Condition preoperatively (F( 1,7) = 42.3,.P < 0.001). This effect did not change postoperatively (F(l,6) = 0.55, n.s.); simple reaction time performance remained more accurate than choice reaction time performance, further demonstrating that the rats continued to use the advance information in performing this task I2SED a so- 25- Preoperative Postoperative Fio. 3. Percentage correct in the simple (open bars) and choice reaction (hatched bars) time conditions is shown preoperatively and postoperatively; there was a higher percentage correct in the simple reaction time condition which was not differentially affected by unilateral stratal dopamine depletion. The bar shows two standard errors of the difference of the means associated with the factor Condition (simple vs choice reaction time). Responses to the dim stimuli were less accurate than responses to the bright stimuli (F(l,7) = 8.77, P = 0.02). There was an average of 74.6% correct to dim stimuli and 78.6% correct to the bright stimuli. Once again, this effect was not changed postoperatively (F(l,6) = 0.03, n.s.). Response bias. Fig. 4 shows response bias for the simple and choice reaction time conditions postoperatively, plotted as a function of delay. From no bias (50%) preoperatively, bias towards the side of the lesion was increased to 74% postoperatively. The degree of spatial response bias was the same for simple and choice reaction time performance (F(l,8) = 0.06, n.s.) and bias also did not vary as a function of delay (F(3,24) = 0.74, n.s.), indicating that the apparent greater impairment in reaction time at the longer delays was not an effect of greater response bias, nor did it result in greater bias. Bias during timeout periods. Postoperatively, fewer responses were made in the hole contralateral to the side of dopamine depletion (mean of 17, compared with 38.5 ipsilateral responses), the significance of which was confirmed by the interaction between the factors of side of Lesion and Side of response (F(l,6) = 14.75, P = 0.009). Movement time. The average of 0.18 s for movement time preoperatively, was nonsignificantly increased to 0.20 s postoperatively. There was a main effect of Condition

8 520 V. J. BROWN AND T. W. ROBBINS 100 S 90 g 70 J2SED Delay (s) FIG. 4. Response bias to the ipsilateral side is plotted as a function of delay before the imperative signal. Postoperatively, animals were biased to the ipsilateral side in both the simple (open circles) and the choice (filled circles) reaction time conditions and bias did not change significantly as a function of condition or delay. The bar shows two standard errors of the difference of the means associated with the interaction between Condition and Delay. (F(l,6) = 9.38, P = 0.02) which was due to faster movement times in the simple reaction time condition (0.18 compared with 0.22 s), but this did not change after surgery. There was no effect of side of lesion on movement time (F(l,6) = 0.11, n.s.). Lack of behavioural recovery. The rats were tested again 30 and 60 days after surgery for 8 consecutive days. The impairments were still apparent up to 60 days postoperatively. Bias was not significantly reduced compared with the first postoperative test (F(l,7) = 1.59, n.s.); immediately postoperatively, bias was 73.7% and after 60 days it was still 69.1%. The reaction time effects also remained unchanged over the three postoperative tests (F(2,14) = 1.39, n.s.). DISCUSSION The results of the present experiment have clearly demonstrated that rats with profound dopamine loss in the striatum on one side are able to use advance information in the selection and preparation of responses to either side, and continue to show the same magnitude of speeding of reaction times and increases in performance accuracy when provided with advance information, while exhibiting specific deficits in reaction time performance. The present behavioural deficits arise from profound (up to 90%), and rather selective, depletion of dopamine from the caudate-putamen. Although 6-OHDA destroys noradrenergic as well as dopaminergic neurons, these neurons represent only a minor portion of striatal catecholaminergic innervation (McGeer et al., 1984) and therefore depletion of noradrenaline probably does not account for the behavioural impairments reported here. Moreover, substantially greater depletion of dopamine unilaterally from the nucleus accumbens, larger than reported here, has been shown to be without effect on rotational behaviour (Kelly and Moore, 1976). The behavioural impairments were long-lasting, persisting for at least 2 months postoperatively, thus contrasting with the many reports of only transient deficits in spontaneous behaviour following striatal dopamine depletion (e.g., Ungerstedt, 1971; Marshall, 1979).

9 REACTION TIME AND STRIATAL DOPAMINE 521 Motor programming: comparison of simple and choice reaction time performance The performance enhancements provided by advance information in the simple, as distinct from the choice, reaction time condition remained both for responses ipsilateral and contralateral to the side of dopamine depletion. These results have important implications for understanding not only the apparently selective deficit in simple reaction time seen in Parkinson's disease that was described in the Introduction, but also the neuropsychological substrates of action. The failure to find differential effects of striatal dopamine depletion on performance of these tasks extends the previous findings of no differential impairment in simple reaction time (Brown and Robbins, 1989a) to situations where the response is precued. Although our results may appear to contrast with some of those seen previously in patients with Parkinson's disease (e.g., Bloxham et al., 1987; Pullman et al., 1988), they are not necessarily inconsistent with the clinical findings. First, from the results of Pullman et al. (1988), it appears that L-dopa medication, in fact masks a deficit in choice reaction time in parkinsonian patients. Nevertheless, simple reaction time performance appears to be relatively impervious to L-dopa medication and this may yet indicate that simple and choice reaction time performance are mediated by distinct neural substrates. However, this argument may be fallacious because simple and choice reaction time performance may provide control baselines which are differentially sensitive to L-dopa medication because of scaling factors. Secondly, in the Pullman et al. (1988) study, it is clear that there is a trend for simple reaction time also to show improvement after L-dopa medication, which may have been significant had a larger sample of patients (than the 5 tested) been used. Therefore it may not be surprising that our results show no differential effect on simple and choice reaction time performance in rats with striatal dopamine depletion. In a group of parkinsonian patients that perhaps is of greatest relevance to the present experiment, namely largely unmedicated patients with hemiparkinsonism, Rafal et al. (1989) showed that the simple reaction time was faster than choice reaction time when tested using the 'bad' hand, showing clear evidence of clinical signs of parkinsonism (presumably contralateral to the side of greatest dopamine depletion) as well as when tested with the asymptomatic ('good') hand. In neuropsychological terms, the present results have failed to support the implication that the speeding of reaction time in simple reaction time tasks is not dependent on striatal dopamine (Pullman et al., 1988) but rather suggest that both simple and choice reaction time performance are striatally influenced. This is not to imply that we think that simple and choice reaction time could not be mediated by different neural substrates within the striatum, for example, by different segregated and parallel corticostriatal loops (Alexander et al., 1986), although the results do not provide direct evidence for this possibility. Motor readiness Although the reaction time advantage of the provision of advance information remained for contralateral responses, unilateral striatal dopamine depletion did impair contralateral responding, in agreement with earlier results (Carli et al., 1985, 1989; Brown and Robbins, 1989a; Robbins and Brown, 1990). Fewer responses were initiated to the contralateral side, with approximately 75% of all responses being directed ipsilaterally. This enhanced spatial bias has been argued to represent alterations in response set, the

10 522 V. J. BROWN AND T. W. ROBBINS prior assignment of probability of selection from the repertoire of available responses (Brown and Robbins, 1989ft; Carli et al, 1989). Contralateral reaction times were lengthened postoperatively, relative to ipsilateral reaction times, but as a function of stimulus presentation delay, with significant differences between the two sides only at the longest delays. This result appears different from that seen in an earlier study (Carli et al., 1989) which also showed delay-dependent changes in reaction time, but with differences at the shorter rather than the longer intervals. However, that study used a different paradigm for measuring reaction time in which the stimuli were presented only to one side of the animal's head at a time, rather than bilaterally. The previous results were consistent with an explanation in terms of lateralized attentional strategies resulting from exaggerated spatial response biases towards the side of the lesion (see Carli et al., 1989). By employing a bilateral mode of stimulus presentation, the present study prevented the development of such lateralized attentional strategies which would have contaminated the reaction time measures used here. We use the term 'motor readiness' to refer to the speeding of reaction time of responses as the delay elapses. In both the simple and choice reaction time conditions, the trials are initiated by the rat itself with a centre hole response. This response extinguishes the light in the central hole which marks the start of the delay. At the start of the delay, the rat cannot know for how long it must wait, but as the delay elapses the occurrence of the stimulus becomes increasingly likely. Thus the speeding of reaction time results from the gradual increase, as delay elapses, in the conditional probability of the occurrence of the visual stimulus (until certainty is reached at the longest delay, 1.5 s). The delaydependent speeding of reaction time has been reported previously in human reaction time studies (Naatanen and Merisalo, 1977; Frith and Done, 1986) where it appears to reflect some aspect of response preparation. In agreement with Frith and Done (1986), it was found in the present study that choice, as well as simple, reaction time showed a linear speeding as a function of delay. This strongly suggests that the effect does not arise from the preparation of a specific response, but rather from a more general readiness to respond. The attenuation of the delay-dependent speeding of reaction time of responses contralateral to the side of dopamine loss implies that this general readiness to respond is mediated at least in part by striatal dopamine. Furthermore, as it is only responses contralateral to the side of dopamine loss which show a failure to benefit from increasing stimulus probability, it appears that 'motor readiness' is lateralized and does not result from generalized bilateral activation. As stimuli of different intensities were used, the present experimental design also allowed a specific test to be made of the dopaminergic mediation of the nonspecific activation that can arise from the effects of stimulus intensity (see Posner, 1978, p. 113). Responses were more accurate to bright than to dim stimuli, and this effect was preserved postoperatively for both ipsilateral and contralateral responses, showing that there was still an effect of nonspecific activation following the reduction in striatal dopamine. Previously, it has also been shown that irrelevant white noise presented just before, or simultaneously with, a visual stimulus speeded to the same degree responses both ipsilateral and contralateral to the side of striatal dopamine depletion (M. Carli, T. W. Robbins, J. L. Evenden, unpublished observations; see Robbins and Brown, 1990), reminiscent of a similar finding in patients with Parkinson's disease (Heilman

11 REACTION TIME AND STRIATAL DOPAMINE 523 et al., 1976), and indicating that this generalized activation by an external stimulus is not mediated by striatal dopamine. The dual effects of impaired 'readiness' of contralateral responses and enhanced ipsilateral response bias do not necessarily reflect a common deficit, for example, of response set. It this were true, then it might be expected that, as the ipsilateral responses become increasingly primed by delay, bias to the ipsilateral side would also increase. In fact, no such effect was seen; bias was equal at all stimulus delays. This dissociation of bias and reaction time has also been observed by us in a different paradigm in which the two measures were differentially affected by cell body lesions of the lateral and medial striatum, respectively (Brown and Robbins, 198%). Thus it is clear that bias is neither a product of reaction time nor the cause of the reaction time effects, and bias and reaction time measures are functionally dissociable, although the same general deficit might underlie each. In summary, unilateral striatal dopamine depletion appears to result in lateralized changes in motor readiness generated by endogenous cues that are unrelated to motor programming per se. This is especially relevant to apparently similar findings in patients with Parkinson's disease by Bloxham et al. (1987) of a significant interaction between the foreperiod and reaction time, with reaction time decreasing as a function of delay in the control group but not in the parkinsonian patients. This deficit in the internal generation of motor readiness may reflect the well-known differential susceptibility of patients with Parkinson's disease to deficits when movements are initiated voluntarily, rather than in response to external cues (e.g., Martin, 1967; Flowers, 1976, 1978; Cooke et al., 1978; Stern et al., 1980). It may also correspond to the impairments in shifting cognitive set following internally derived, rather than externally provided, information (Brown and Marsden, 1988). The effects of striatal dopamine depletion on motor readiness are of particular interest as this is one of the functions attributed to the SMA, one component of a loop involving the basal ganglia and ventrolateral thalamus (Goldberg, 1985). Activation of this loop gives rise to the Bereitschaftspotential that is associated with the preparation of voluntary movement and which Deecke (1985) has described as having a 'priming function for our freely voluntary, fully endogenous movements and actions'. The Bereitschaftspotential, has recently been reported to be reduced both in duration and amplitude in cases of Parkinson's disease (e.g., Simpson and Khuraibet, 1987; Dick et al., 1989). These considerations and the results reported here on the effects of striatal dopamine depletion in the rat suggest that it would be of interest to examine the effects of temporal predictability and stimulus uncertainty in Parkinson's disease. ACKNOWLEDGEMENTS This work was supported by a Major Award from the Wellcome Trust. V.J.B. was in receipt of a Medical Research Council studentship. We would like to thank Dr R. D. Rafal for helpful comments on the manuscript. REFERENCES ACID Y, RUBERG M, DUBOIS B, PILIXJN B (1987) Anatomoclinical and biochemical concepts of subcortical dementia. In: Cognitive Neurochemistry. Edited by S. M. Slahl, S. D. Iversen and E. C. Goodman. Oxford: Oxford University Press, pp

12 524 V. J. BROWN AND T. W. ROBBINS ALEXANDER GE, DELONG MR, STRICK PL (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annual Review of Neuroscience, 9, BLOXHAM CA, MINDEL TA, FRITH CD (1984) Initiation and execution of predictable and unpredictable movements in Parkinson's disease. Brain, 107, BLOXHAM CA, DICK DJ, MOORE M (1987) Reaction times and attention in Parkinson's disease. Journal of Neurology, Neurosurgery and Psychiatry, 50, BREESE GR, TRAYLOR TD (1971) Depletion of brain noradrenaline and dopamine by 6-hydroxydopamine. British Journal of Pharmacology, 42, BROWN RG, MARSDEN CD (1988) Internal versus external cues and the control of attention in Parkinson's disease. Brain, 111, BROWN VJ, ROBBINS TW (1989a) Deficits in response space following unilateral striatal dopamine depletion in the rat. Journal of Neuroscience, 9, BROWN VJ, ROBBINS TW (1989i) Elementary processes of response selection mediated by distinct regions of the striatum. Journal of Neuroscience, 9, CARLI M, EVENDEN JL, ROBBINS TW (1985) Depletion of unilateral striatal dopamine impairs initiation of contralateral actions and not sensory attention. Nature, London, 313, CARU M, JONES GH, ROBBINS TW (1989) Effects of unilateral dorsal and ventral striatal dopamine depletion on visual neglect in the rat: a neural and behavioural analysis. Neuroscience, 29, COOKE JD, BROWN JD, BROOKS VB (1978) Increased dependence on visual information for movement control in patients with Parkinson's disease. Canadian Journal of Neurological Sciences, 5, DEECKE L (1985) Cerebral potentials related to voluntary actions: parkinsonian and normal subjects. In: Clinical Neurophysiology in Parkinsonism: Contribution to Assessment and Pathophysiology. Restorative Neurology Volume 2. Edited by P. J. Delwaide and A. Agnoli. Amsterdam and Oxford: Elsevier, pp DICK JPR, ROTHWELL JC, DAY BL, CANTELLO R, BURUMA O, GIOUX M et al. (1989) The Bereitschaftspotential is abnormal in Parkinson's disease. Brain, 112, EVARTS EV, TERAVAINEN H, CALNE DB (1981) Reaction time in Parkinson's disease. Brain, 104, FLOWERS KA (1976) Visual 'closed-loop' and 'open-loop' characteristics of voluntary movement in patients with parkinsonism and intention tremor. Brain, 99, FLOWERS K (1978) Some frequency response characteristics of parkinsonism on pursuit tracking. Brain, 101, FRITH CD, DONE DJ (1986) Routes to action in reaction time tasks. Psychological Research, 48, GARNETT ES, NAHMIAS G, FIRNAU G (1984) Central dopaminergic pathways in hemiparkinsonism examined by positron emission tomography. Canadian Journal of Neurological Sciences, 11, GOLDBERG G (1985) Supplementary motor area structure and function: review and hypotheses. Behavioral and Brain Sciences, 8, HEILMAN KM, BOWERS D, WATSON RT, GREER M (1976) Reaction times in Parkinson's disease. Archives of Neurology, Chicago, 33, HORNYKIEWICZ O (1982) Brain neurotransmitter changes in Parkinson's disease. In: Movement Disorders. Edited by C. D. Marsden and S. Fahn. London: Butterworth, pp KELLY PH, MOORE KE (1976) Mesolimbic dopaminergic neurones in the rotational model of nigrostriatal function. Nature, London, 263, MCGEER EG, STAINES WA, MCGEER PL (1984) Neurotransmitters in the basal ganglia. The Canadian Journal of Neurological Sciences, 11, MARSDEN CD (1982) The mysterious motor function of the basal ganglia: the Robert Wartenberg Lecture. Neurology, New York, 32, MARSHALL JF (1979) Somatosensory inattention after dopamine-depleting intracerebral 6-OHDA injections: spontaneous recovery and pharmacological control. Brain Research, Amsterdam, 177, MARTIN JP (1967) The Basal Ganglia and Posture. London: Pitman Medical. MEFFORD IN (1981) Application of high performance liquid chromatography with electrochemical detection to neurochemical analysis: measurement of catecholamines, serotonin and metabolites in rat brain. Journal of Neuroscience Methods, 3, NAATANEN R, MERISALO A (1977) Expectancy and preparation in simple reaction time. In: Attention and Performance VI. Edited by S. Dornic. Hillsdale, NJ: Lawrence Erlbaum, pp POSNER MI (1978) Chronometric Explorations of Mind. Hillsdale, NJ: Lawrence Erlbaum.

13 REACTION TIME AND STRIATAL DOPAMINE 525 PULLMAN SL, WATTS RL, JUNCOS JL, CHASE TN, SANES JN (1988) Dopaminergic effects on simple and choice reaction time performance in Parkinson's disease. Neurology, Cleveland, 38, RAFAL RD, POSNER MI, WALKER JA, FRIEDRICH FJ (1984) Cognition and the basal ganglia: separating mental and motor components of performance in Parkinson's disease. Brain, 107, RAFAL RD, FRIEDMAN JH, LANNON MC (1989) Preparation of manual movements in hemiparkinsonism. Journal of Neurology, Neurosurgery and Psychiatry, 52, ROBBINS TW, BROWN VJ (1990) The role of the striatum in the mental chronometry of action: a theoretical review. Reviews in the Neurosciences. In press. SCATTON B, ROUQUIER L, JAVOY-AGID F, AGID Y (1982) Dopamine deficiency in the cerebral cortex in Parkinson's disease. Neurology, New York, 32, SHERIDAN MR, FLOWERS KA, HURRELL J (1987) Programming and execution of movement in Parkinson's disease. Brain, 110, SIMPSON JA, KHURAIBET AJ (1987) Readiness potential of cortical area 6 preceding self paced movement in Parkinson's disease. Journal of Neurology, Neurosurgery and Psychiatry, 50, STERN GM, LANDER CM, LEES AJ (1980) Akinetic freezing and trick movements in Parkinson's disease. Journal of Neural Transmission, Supplement 16, STERN Y, MAYEUX R, C6T L (1984) Reaction time and vigilance in Parkinson's disease: possible role of altered norepinephrine metabolism. Archives of Neurology, Chicago, 41, UNGERSTEDT U (1971) Postsynaptic supersensitivity after 6-hydroxydopamine induced degeneration of the nigro-striatal dopamine system. Ada Physiologica Scandinavica, Supplement 367, WINER BJ (1971) Statistical Principles in Experimental Design. Second edition. New York: McGraw-Hill. (Received August 11, Revised March 6, Accepted April 4, 199G)

14