Saccadic latency distributions in Parkinson's disease and the effects of L-dopa

Abstract

Parkinson's disease (PD) is associated with a loss of central dopaminergic pathways in the brain leading to an abnormality of movement, including saccades. In PD, analysis of saccadic latency distributions, rather than mean latencies, can provide much more information about how the neural decision process that precedes movement is affected by disease or medication. Subject to the constraints of intersubject variation and reproducibility, latency distribution may represent an attractive potential biomarker of PD. Here we report two studies that provide information about these parameters, and demonstrate a novel effect of dopamine on saccadic latency, implying that it influences the neural decision process itself. We performed a detailed cross-sectional study of saccadic latency distributions during a simple step task in 22 medicated patients and 27 age-matched controls. This revealed high intersubject variability and an overlap of PD and control distributions. A second study was undertaken on a different population specifically to investigate the effects of dopamine on saccadic latency distributions in 15 PD patients. L-dopa was found to prolong latency, although the magnitude of the effect varied between subjects. The implications of these observations for the use of saccadic latency distributions as a potential biomarker of PD are discussed, as are the effects of L-dopa on neural decision making, where it is postulated to increase the criterion level of evidence required before the decision to move is made.

Figures

Fig. 1

Saccadic latency distribution plotted on reciprobit axes. An example cumulative frequency plot of saccadic latencies is shown, using a probit scale as a function of reciprocal latency: a reciprobit plot. When saccadic latencies are plotted in this way, most points lie on line a, intercepting the infinite latency axis at an intercept I. In some cases, as here, there may be an additional sub-population of early saccades, which tend to lie on a distinct line b, which passes through the point (inf,50). A change in expectation or urgency causes swivel about the point I, generating a new line of best fit such as the dashed line c (see text and Reddi et al. 2003). Three parameters (median latency, main slope, and slope of the early component) are sufficient to characterise the entire distribution

Fig. 2

a Three example reciprobit plots from patients ‘on’ (blue circle) versus ‘off’ (red square) L-dopa. Note that the shape of the distribution of saccadic latencies is reproducible and tends to swivel anticlockwise on l-dopa (see also Fig. 1). The magnitude of the effect varies between subjects. b Three example reciprobit plots from control subjects. There is intersubject variability in median latency and the shape of the distribution varies between subjects. Note that small differences in the tails of reciprobit plots have highly visible effects on their shape despite representing a small proportion of the saccades (P, PD patient codes refer to Table 2; C control)

Fig. 3

The intersubject variability in saccadic latency is high in all groups. a Mean ± SEM error bars. There is a significant difference between the control and ‘PD on’ groups (t=2.20, df=40, P=0.03), and between the ‘PD on’ and ‘PD off’ groups (t =−2.15, df=14, P=0.05). b Latencies of each subject are plotted individually to demonstrate the intersubject variability

Fig. 4

The effect of l-dopa on saccadic latency distributions: it tends to slow reaction time. a There is a significant difference in reciprocal median saccadic latencies on versus off l-dopa (P=0.05), although it varies between subjects (P1–P15, see Table 2). b The Kolmogorov–Smirnov test showed that subjects 1, 2, 4, 10 and 13 had significantly different median saccadic latencies on and off l-dopa (P<0.05). The COMT Val158Met and BDNF Val66Met polymorphism genotype is shown for each subject: V valine; M methionine; a dash indicates that no sample was available. (A full colour version of this figure is available online at http://www.dx.doi.org/10.1007/s00221-006-0412-z

Fig. 5

When tested on medication the reciprocal median saccadic latency correlated with the patients' UPDRS motor subsection score (regression line fitted, r=−0.55, P=0.03)