Gait and Posture 19 (2004) 172–183 Comparison between subthalamic nucleus and globus pallidus internus stimulation for postural performance in Parkinson’s disease  Laura Rocchi a,∗ , Lorenzo Chiari a , Angelo Cappello a , Anne Gross b , Fay B. Horak b a Biomedical Engineering Unit, Department of Electronics, Computer Science and Systems, DEIS-University of Bologna, Viale Risorgimento, 2, 40136 Bologna, Italy b Neurological Sciences Institute, Oregon Health and Science University, West Campus-Bldg, 1505 N.W. 185th Avenue Beaverton, OR 97006, USA Received 12 April 2003; accepted 12 April 2003 Abstract Nine subjects with Parkinson’s disease, five subjects with electrodes implanted in the subthalamic nucleus (STN) and four with electrodes in the globus pallidus internus (GPi), were compared with a population of ten age-matched control subjects. The measures studied include a set of summary statistic scores, two stochastic parameters, the distribution of the center of pressure (CoP) displacement angles under each foot, and the distribution of bilateral patterns of CoP displacement angles. A Bayes classifier was designed to monitor the trend of postural performance in patients, with different treatments. Results suggested that the selected measures were sensitive to Parkinsonian postural sway abnormalities and highlight differences in response to treatments. Deep brain stimulation restored a more normal postural sway and levodopa increased sway abnormalities. Furthermore, the selected measures appear to detect different responses to levodopa between the STN and GPi groups: the negative side effects of levodopa on posture were less severe for STN than for GPi patients, perhaps due to the decreased need for levodopa intake in STN subjects. The measures proposed in this preliminary study may be useful adjuncts to evaluate balance and postural control strategies in patients with Parkinson’s disease and may allow the comparison of DBS electrode sites, on stance posture. © 2003 Elsevier B.V. All rights reserved. Keywords: Parkinson’s disease; Posture control; Deep brain stimulation; Centre of pressure; Bilateral measures 1. Introduction Natural sway and corrective surface reactive forces dur- ing stance are important signs of postural function and can be quantified by the measurement of the center of pressure (CoP). The time- or frequency-domain analysis of CoP time-series can provide several measures of postural steadi- ness. The most common CoP analysis, particularly in clinical practice, estimates statistical properties of CoP displace- ment [1]. In addition, model-based parameters postulate the time-scale dependence of CoP statistical properties [2,3]. Quantification of CoP parameters can be especially use- ful to evaluate postural control in a population that have a disordered postural-motor control system, such as subjects with Parkinson’s disease (PD) [4,5]. Postural problems in  SIAMOC best paper award 2002. ∗ Corresponding author. Tel.: +39-051-209-3067; fax: +39-051-209-3540. E-mail address: lrocchi@deis.unibo.it (L. Rocchi). patients with PD become increasingly severe as the disease progresses, despite levodopa replacement therapy [6]. The worsening of posture with the severity of PD may be due to the increased involvement of non-dopaminergic pathways, as suggested by the declining effectiveness of levodopa re- placement therapy as PD progresses [7,8]. High frequency deep brain stimulation (DBS) in the basal ganglia is now being used to augment levodopa replacement therapy in subjects with severe PD [8–11]. DBS consists of tonic, high frequency stimulation bilaterally in the sub- thalamic nucleus (STN) or in the globus pallidus internus (GPi). Although there is a consensus that levodopa replace- ment does not help the balance problems of subjects with PD [12], the effectiveness of DBS on balance is still un- known. The relative effectiveness of DBS in STN and GPi is still controversial [13–15]. A previous study in our laboratories [5] suggested that quantitative posturography is useful in understanding motor control in subjects with PD and effect of anti-Parkinsonian treatments. Using traditional, statistical measures of CoP 0966-6362/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0966-6362(03)00059-6