Cerebral and cerebellar regions involved in the control of smooth pursuit - A functional magnetic resonance imaging study

C. Moschner1, M.F. Nitschke1, W. Heide1, F. Binkofski2, G. Buccino3, S. Posse4, D. Kömpf1, H.J. Freund2, R.J. Seitz2

1Dept. of Neurology, Medical University of Lübeck, Ratzeburger Allee 160, D-23538 Lübeck, Germany (e-mail:moschner_c@neuro.mu-luebeck.de);
2Dept of Neurology, University of Düsseldorf, Germany;
3Dept. of Physiology, Parma University, Parma, Italy;
4Institute of Medicine, Research Center Jülich, Germany

Introduction: Electrophysiological and lesion studies in primates as well as studies in human patients have demonstrated complex cerebro-cerebellar networks that initiate and maintain smooth pursuit eye movements (SP) and correct for positional errors by means of small saccades. SP-related cortical regions in primates include the striate cortex, the middle temporal area (MT) and middle superior temporal area (MST) of the superior temporal sulcus, the posterior parietal cortex as well as the frontal eye fields (FEF) and supplementary eye fields (SEF). Prior fMRI studies on SP were focussed on the relevant regions in the cerebral cortex. One goal of this study was to demonstrate corresponding activation in cerebellar areas that are known to be involved in SP control, e.g. the flocculus, paraflocculus, or the posterior vermis.

Methods: Seven right-handed healthy volunteers were asked to track a laser dot projected on a screen in a darkened room. During the activation period the following conditions were tested:

  1. Continuous sinusoidal stimulation (0.5 Hz; amplitude 10 deg) which normally leads to steady-state SP without many catch-up saccades.
  2. Sequential step-ramps with a ramp movement from the left into the right hemifield (velocity 16 deg/s). This triggers the initiation of SP (and frequent catch-up saccades).
  3. Sequential step displacements to elicit pure visually-guided saccades. Constant fixation in centre position served as control.

Functional mapping of the whole brain was performed at 1.5 T (echoplanar sequences, TR/TE/flip angle = 3 s/66 ms/90 deg, voxel size 3x3x4 mm). The protocol consisted of 6 cycles alternating between task and control condition (25 ms each). After realignment, spatial normalisation and smoothing, individual and group analysis was performed using SPM 96 (P < 0.05).

Results: In both SP tasks, activation was found in the striate cortex bilaterally. Additional activation in the FEF and SEF, in the posterior vermis and the paramedian cerebellar hemispheres was more pronounced during step-ramps compared to sinusoidal stimulation, possibly due to an increased number of catch-up saccades in the step-ramp condition. In line with that, pure saccades also activated the FEF, SEF and the posterior vermis. Furthermore, a right hemispheric preponderance of activity during step-ramps was found in the area V5, the precuneus, and the intraparietal sulcus (IPS). This seems to reflect the rightward direction of SP.

Conclusions: By means of fMRI, we were able to demonstrate SP-related activation in distinct regions of the cerebral cortex, and, additionally, in certain cerebellar areas, especially in the posterior vermis. From animal studies these location are known to be part of the cortico-cerebellar networks that control SP. However, some of the activation (e.g., in FEF, SEF and the posterior vermis) might be due to co-stimulation of saccade-related neuronal activity.