Motor Systems NeuroReport 8, 1209–1213 (1997)

1

1

1

1

1

p

SACCADES of patients with mild Parkinson’s disease (PD) are said to be abnormal in the absence of a concurrently visible target or when they are part of a rapid sequence of eye movements. We tested this hypothesis using a predictive saccade paradigm in which target visibility is withdrawn for a period. Three rates of target alternation were used (0.25 Hz, 0.5 Hz and 1.0 Hz). Withdrawal of target visibility brought out the extremes of primary saccade gain for both the controls and the patients with PD, most undershoot being displayed at the lowest frequency, whereas the gain was greatest at the highest frequency, actually overshooting the target location. These results demonstrate that the spatial error of parkinsonian saccades does not invariably take the form of hypometria when part of a rapid sequence of eye movements.

Abnormalities of predictive saccades in Parkinson’s disease

Key words: Parkinson’s disease; Predictive saccades

CA

Introduction

and antisaccades normally, but not remembered or predictive saccades.4 Lueck et al.5 suggested that the critical feature shared by these paradigms was the absence of a visible target at the time of programming the saccade, the concurrent target visibility (CTV) hypothesis. However, while the absence of CTV invariably resulted in saccadic hypometria, the presence of CTV did not guarantee normality. It also seemed that Parkinsonian saccades were abnormal when they were part of a rapid sequence of eye movements.3,5–10 There appear, therefore, to be two features of target presentation which in PD result in saccadic hypometria; namely, absence of a visible target at about the time the saccade was summoned or the requirement to execute a rapid sequence of saccades.5 Crawford et al.3 studied predictive saccades in patients with PD employing a target which regularly alternated every 2 s between two fixed locations, 22.5° apart. The hypometria observed in this task was more severe than had been found in the standard remembered paradigm. However, a study of predictive saccades in hemi-Parkinson’s disease has found saccadic abnormalities to be more prominent at lower frequencies of target stepping11 and PET evidence suggests that basal ganglia activation is not rate sensitive.12 In the present study, our principal aim was to investigate the effects of rate of target presentation on predictive saccades in Parkinson’s disease. This we combined with a manipulation that had previously amplified the difference between patients with PDand controls,3 in which target visibility is suddenly withdrawn for a period but the subject is required to continue tracking as if the target were still visible.

The highly selective impairments which sometimes result from focal brain damage can provide information about the relationship between neural pathways in the brain and the various cognitive processes that underlie a particular behaviour. For saccadic eye movements, such a prospect seems to be offered by the pattern of abnormality found in the early stages of Parkinson’s disease (PD), when the pathology is still fairly circumscribed to the dopaminergic nigrostriatal pathway. In these patients the saccadic abnormality is doubly selective, being restricted to a specific parametric feature of saccades as well as to specific behavioural paradigms. It is mainly confined to the initial saccade to the target (primary saccade) which tends to be abnormally reduced in amplitude (hypometria), while the latency and peak velocity are normal. For large saccadic excursions the final eye position (FEP) may be reached in multiple steps (multistepping hypometria), but the FEP is usually normal. In patients with mild to moderate idiopathic PD saccades were normal when directed to a novel target (reflexive saccade)1,2 or even to the mirror image of the target location (antisaccade).2 Multistepping hypometria, however, occurred whenever saccades had to be directed towards the remembered location of a target which had recently been visible (remembered saccades)1 or where a target moving repetitively between two locations was expected to appear (predictive saccades).3 Similar results have been demonstrated in patients with bilateral lentiform nucleus lesions, who can perform reflexive saccades © Rapid Science Publishers

Eoin P. O’Sullivan, Sandip Shaunak, Leslie Henderson, Malcolm Hawken, Trevor J. Crawford and Christopher KennardCA Department of Clinical Neuroscience, Charing Cross and Westminster Medical School, Fulham Palace Road, London W6 8RF, UK Corresponding Author

Vol 8 No 5 24 March 1997

1209

E. P. O’Sullivan et al 1

Since previous studies of repetitive eye movements in the absence of visual targets have found considerable hypermetria (target overshoot) in normal subjects,13–15 we were particularly interested in assessing the performance of the patients with PD in this condition.

Subjects and Methods 11

11

11

11

11

1p

Subjects: Nine patients with mild to moderate PD (Hoehn and Yahr grade 1, n = 6; grade 2, n = 1; grade 3, n = 2) were recruited. Their ages ranged from 49 to 79 years with a mean of 66 years and a mean disease duration of 2.6 years. All were free from dementia (none scoring below 28 on the Mini Mental State test). One patient was untreated; the remainder were receiving only levodopa preparations and selegeline. The seven normal age-matched control subjects ranged in age from 40 to 76 years with a mean of 59 years. All subjects gave informed consent. Eye movement recordings: Eye movements were recorded in total darkness, using either the magnetic scleral search coil technique (CNC Engineering, Seattle, Washington) or infrared oculography (Scalar, IRIS). Subjects were comfortably seated 150 cm from a flat, translucent screen in which red light emitting diode (LED) targets were embedded. A buzzer was mounted behind the subject’s head, which was supported by a head rest designed to minimize movements of the head. Each recording session was preceded by a calibration sequence in which subjects fixated horizontal targets with known positions. Stimulus presentation was controlled by a PDP 11/73 computer with a CED 502 interface. Predictive paradigm: Two targets positioned 11.25° to the left and right of the mid-line were illuminated alternately at one of three fixed frequencies, 0.25, 0.5 and 1.0 Hz. On every trial a buzzer situated directly behind the subject’s head provided an audible cue, synchronous with each target step. At each frequency subjects were tested in three blocks of 12 trials, these blocks being run as a continuous sequence. Two blocks with presentation of the visual targets (V1 and V2) were separated by a block in which no visual target was presented (NV) but the buzzer continued, as before, to supply timing information. Instructions to the subjects informed them fully about the various manipulations. Specifically, they were informed about the unvarying regularity of target alternation at each rate and were asked to move their eyes in time with the stepping of the target. Subjects were also informed about the sudden withdrawal of the visual target at the commencement of the NV block and were requested to continue 1210 Vol 8 No 5 24 March 1997

executing saccades back and forth, exactly as if the target were still present. Data analysis: Saccadic data was analysed using interactive computer programmes (one written locally using the ASYST language, Asyst Software Technology, Rochester, NY and a commercial programme EYEMAP, Amtech GmbH, Heidelberg). The gain of the primary saccade (defined as the first saccade occurring prior to or just following a target step in that direction) and the gain of the FEP were measured. Latency of the primary saccade was also measured to ensure that the task was being correctly performed. The number of secondary saccades leading up to FEP was calculated. The first step in each block was not analysed as in Block V1 it was a half step and in the other blocks it marked a transition period. This resulted in 11 steps being analysed per block.

Results Since the pattern of performance was similar in blocks V1 and V2 for the gain and latency of the primary saccade as well as for FEP, we collapsed the data over blocks V1 and V2, to produce a V/NV factor balanced for block order and used this factor in all further analyses. Separate ANOVAs were performed on latency and gain of the primary saccade and FEP gain, with group as a between subject factor and target rate and visibility as within subject factors. Primary saccade latency (see Table 1): Inspection of trial by trial data revealed that both the controls and the patients with PD developed predictive behaviour within the first two cycles of each rate condition. The main effects of both rate and visibility on the latency of primary saccades were significant (F(2,28) = 7.98, p < 0.005 and F(1,14) = 9.75, p < 0.01 respectively). Moreover, rate interacted significantly with group (F(2,28) = 3.52, p < 0.05) and with visibility (F(2,28) = 14.68, p < 0.005), reflecting the fact that the tendency of latencies to be more anticipatory (i.e. negative), in the patient with PD group and the NV condition, was largely confined to the slowest rate of target alternation. Overall, the patients with PD were more rate-sensitive. Primary saccade gain and final eye position (see Table 1): Across all trials both the primary saccades and the FEPs of the patient with PD group had lower gain than in the control group (F(1,14) = 12.08, p < 0.005 and F(1,14) = 5.71, p < 0.05, respectively). However, between the two groups this effect was qualified by target visibility only for FEP (F(1,14) = 11.58, p < 0.005). Whereas withdrawal of the visual

Abnormalities of parkinsonian saccades Table 1. Mean values for primary saccade gain, FEP gain, latency (ms) and number of secondary saccades for controls and Parkinson’s disease (PD) patients across each frequency and in the presence (V) and absence (NV) of visual targets. 0.25 Hz V

1

0.5 Hz

1.0 Hz

NV

V

NV

V

NV

Primary saccade gain Control 0.90 PD 0.63

0.83 0.47

0.90 0.67

1.03 0.71

0.86 0.79

1.15 0.93

FEP gain Control PD

1.31 0.97

1.03 0.90

1.26 1.00

1.01 0.87

1.26 1.02

–276 –209

–51 –43

–87 24

0.95 0.95

Latency (ms) Control 55 PD –230

–371 –643

–180 –269

Number of secondary saccades per primary saccade Control 0.60 1.06 0.60 0.45 0.27 0.09 PD 1.34 1.45 0.92 1.03 0.22 0.24

1

1

1 FIG. 1. Mean primary saccade (black bars) and final eye position gain (open bars) for controls and Parkinson’s disease (PD) patients with standard errors in the presence (V) and absence (NV) of the visual targets. The data are collapsed across all three frequencies.

1

p

target raised the patient with PD group’s FEP from 0.91 to 1.00, in the normal controls it raised the gain from the perfect accuracy (gain of 1.0) of their performance when the target was visible to an overshoot by 28% when the visual target was withdrawn (see Fig. 1). No significant interaction was found between rate, target visibility and group. However when the data were collapsed over both groups the primary saccade gain increased markedly with rate of target

FIG. 2. Group mean primary saccade gain with s.e.m. by trial number for each rate of target presentation; (A) 0.25 Hz, (B) 0.5 Hz and (C) 1.0 Hz. V/NV indicates the presence/absence of the visual target. Open symbols signify controls and closed symbols signify patients with Parkinson’s disease.

alternation (F(2,28) = 7.88, p < 0.005), and this effect was much greater in the absence of a visible target (F(2,28) = 10.61, p < 0.001). In the NV block, gain was positively related to frequency whereas FEP appeared wholly insensitive to frequency (F < 1.0). In the NV block at the fastest rate of target presentation, controls and patient with PD made progressively larger saccades so that by the end of the block both groups were making hypermetric saccades (see Fig. 2). Vol 8 No 5 24 March 1997

1211

E. P. O’Sullivan et al 1

Increase of target frequency was accompanied by a reduction in the mean number of secondary saccades per target excursion from 1.4 in patients with PD and 0.8 in controls at 0.25 Hz to 0.2 in both groups at 1.0 Hz (see Table 1).

Discussion 11

11

11

11

11

1p

This study confirms that there is a rate-dependent effect for primary saccade gain. Contrary to our expectations, however, it actually takes the form of increasing gain with increasing frequency of target presentation. This effect did not differentiate between patients with PD and controls. The only factor, an interaction between target visibility and group, which differentiated between patients and controls for gain was the effect of target visibility for the FEP gain. In the presence of visual targets, controls and patients with PD attained a similar FEP independent rate of target presentation. However, in the absence of the targets the controls overshot the target location whereas the patients with PD continued to have an accurate FEP. It is, therefore, the absence of visible targets and not the effect of rate of target presentation which demonstrates the differences between the groups. In the predictive saccade task, the gain of a primary saccade probably reflects the motor programming of a saccade. Our findings suggest that although the actual motor programming is influenced by rate, since the effect was common to both patients with PD and controls, it is probably not a specific function of the dopaminergic nigrostriatal system. This is in keeping with the PET finding that regional cerebral blood flow in the basal ganglia is independent of the rate of limb movement.12 The findings also serve to refute the suggestion that fast rates of target presentation in themselves lead to parkinsonian hypometria.5 It is the visibility of the target that differentiates between groups. The fact that the FEP in patients with PD is not significantly reduced relative to the controls in the presence of the visual target suggests that the effect is not merely an overall decrease in the parkinsonian saccade gain, nor is it due to the reduced amount of time available to make secondary saccades. Instead the effect appears to be dependent on the absence of visible targets. Somatomotor studies of prehension have also demonstrated an abnormal dependence on vision in patients with PD.16 Lueck et al.5 hypothesized that the requirement for concurrent target visibility may have been to allow subjects to encode spatial co-ordinates in a retinotopic frame. In the absence of visible targets it may not be possible to use a retinotopic co-ordinate frame, co-ordinates having to be mapped in another reference frame such as a craniotopic one. Studies in non-human primates 1212 Vol 8 No 5 24 March 1997

have suggested that there may be divergence of cortical control for these mapping systems, the frontal eye fields performing retinotopic mapping and the supplementary eye fields, which are located within the supplementary motor area (SMA), craniotopic mapping,17,18 although this is controversial.19 Studies on patients with cortical lesions appear to add some support for this idea (see Ref. 20 for review). This separation of control may explain the parkinsonian abnormality as it has been demonstrated that patients with PD have decreased regional blood flow in the SMA,21 and this has been used to explain the increased numbers of errors made by patients with PD when performing sequences of memory-guided saccades after withdrawal of levodopa therapy.22 An alternative explanation is that the target locations are remembered in retinotopic space and that the performance of the patients with PD is a function of the delay since the target was last visible (i.e. the length of the memory period). If this was indeed the case, the primary saccade and FEP gains would be reduced both at the lower rates of target alternation and in the NV blocks relative to the V blocks. An interaction between absence of visual targets and rate would be expected and indeed a highly significant interaction was demonstrated in our study for primary saccade but not FEP gain. However, several points need to be considered. The parkinsonian abnormality in both remembered and predictive tasks takes the form of reduced primary saccade gain.1–3 The effect of the length of the delay period in remembered saccades is unclear. In one study, which used delays of up to 5 s, there was no evidence of a reduced FEP gain, nor was the reduction of primary saccade gain influenced by the length of the delay,23 whereas in another study a delay of 6 s resulted in reduced FEP but only for saccades to the left.22 In our study both groups acted in a similar manner so there was no three-way interaction between target visibility, rate and group for either primary saccade or FEP gain. If it were true that the longer the delay period the more hypometric the saccades, a progressive decline in gain over the course of the trials in NV blocks would be expected, but this was not observed. Indeed at 1 Hz there was an increase in primary saccade gain in the NV block (see Fig. 2C). Although the patients with PD were hypometric relative to the controls this is the first time that primary saccade hypermetria relative to the target location has been demonstrated in these patients. The hypermetria demonstrated by controls has been previously described when subjects made sequences of saccades of >20° in the absence of visual targets13–15 and remembered saccadic tasks to radial targets in humans and monkeys, where the hypermetria was confined to upward saccades.24 A possible reason for

Abnormalities of parkinsonian saccades

1

this was that the remembered saccade task requires a spatial transformation process which has an inherently imprecise spatial resolution.24 This study was not concerned with the latency of primary saccades other than to use it as an indicator of task performance. Subjects performed the task well and it is interesting to note that at the slower rates patients with PD have more negative latencies than the controls. This is probably because at the slower rates the patients with PD made more secondary saccades and so would have to initiate their eye movements earlier to keep in time with the target, whereas at the 1.0 Hz rate they made fewer secondary saccades.

Conclusion

1

We have demonstrated that it is the visibility of the target rather than the rate of target presentation which differentiates between patients with PD and controls. The abnormal dependence on visual targets demonstrated by the patients with PD suggests that it might be the co-ordinate frame in which targets are stored that is important. These results also demonstrate that the spatial error of Parkinsonian saccades does not invariably take the form of hypometria when part of a rapid sequence of eye movements and that they can be hypermetric relative to the target location.

References 1. Crawford TJ, Henderson L and Kennard C. Brain 112, 1573–1586 (1989). 2. Lueck CJ, Tanyeri S, Crawford TJ et al. J Neurol Neurosurg Psychiatry 53, 284–288 (1990) 3. Crawford T, Goodrich S, Henderson L et al. J Neurol Neurosurg Psychiatry 52, 1033–1042 (1989). 4. Vermesch A-I, Müri RM, Rivaud S et al. J Neurol Neurosurg Psychiatry 60, 179–184 (1996). 5. Lueck CJ, Crawford T, Henderson L et al. Q J Exp Psychol 45A, 211–233 (1992). 6. Lueck CJ, Tanyeri S, Crawford T et al. Q J Exp Psychol 45A, 193–210 (1992). 7. DeJong JD and Melvill Jones MG. Exp Neurol 32, 58–68 (1971). 8. Bronstein AM and Kennard C. Brain 108, 925–940 (1985) 9. Shibasaki H, Tsuji S and Kuroiwa Y. Arch Neurol 36, 360–364 (1979). 10. Teravainen H and Calne DB. Acta Neurol Scand 62, 137–148 (1980). 11. Ventre J, Zee D , Papageorgiou H and Reich S. Brain 115, 1147–1165 (1992). 12. Jenkins IH, Pasingham RE, Frackowiak RSJ et al. Movement Disord 9 S1, 118 (1994). 13. Becker W and Fuchs AF. Vision Res 9, 1247–1258 (1969). 14. Ohtsuka K, Sawa M and Takeda M. Ophthalmologica 198, 53–56 (1989). 15. Zambarbieri D, Schmid R and Ventre J. Saccadic eye movements to predictable visual and auditory targets. In: O’Regan JK, Levy-Schoen A, eds. Eye Movements: From Physiology to Cognition. North-Holland: Elsevier Science, 1987: 131–140. 16. Jackson SR, Jackson GR, Harrison J et al. Exp Brain Res 105, 147–162 (1995). 17. Schall JD. J Neurophysiol 66, 530–558 (1991). 18. Schall JD. J Neurophysiol 66, 559–579 (1991). 19. Russo GS and Bruce CJ. J Neurophysiol 76, 825–848 (1996). 20. Pierrot-Deseilligny C, Rivaud S, Gaymard B et al. Ann Neurol 37, 557–567 (1995). 21. Jenkins H, Fernadez W, Playford ED et al. Ann Neurol 32, 749–757 (1992). 22. Vermersch A-I, Rivaud S, Vidailhet M et al. Ann Neurol 35, 487–490 (1994). 23. Shaunak S, O’Sullivan EP, Crawford TJ et al. J Neurol 241, S149 (1994). 24. Gnadt JW, Bracewell RM and Andersen RA. Vision Res 31, 693–715 (1991).

ACKNOWLEDGEMENTS: This work was supported by the Wellcome Trust.

Received 18 December 1996; accepted 23 January 1997

1 General Summary Under certain experimental conditions, the saccades (the fast eye movements which change the direction of gaze) of patients with Parkinson’s disease fall short of the intended target location. Such circumstances are the absence of a visible target when the saccade is being generated or the requirement to make a rapid sequence of saccades. We investigated the interaction of these factors by using an experiment in which saccades were made at various rates and in which the targets disappeared. Subjects were told that the targets would disappear and were asked to continue executing saccades, as if the targets were still present. Withdrawal of target visibility brought out the extremes of performance, most undershoot being displayed at the lowest frequency, whereas the amplitude of the saccades was greatest at the highest frequency. These results demonstrate that Parkinsonian saccades need not always undershoot the target location when part of a rapid sequence of eye movements.

1

1

p Vol 8 No 5 24 March 1997

1213