Klinische Neurophysiologie 2004; 35 - 312
DOI: 10.1055/s-2004-832224

Cerebellar Involvement in Visual On-Line Adaptation of Vertical Stroke Size: A PET Activation Study

M Weiss 1, T Schenk 2, M Peller 3, A Drzezga 4, M Schwaiger 5, HR Siebner 6
  • 1Kiel
  • 2Durham
  • 3Kiel
  • 4München
  • 5München
  • 6Kiel

Introduction: Behavioral data suggest that visual feedback can be effectively used for on-line adjustment of letter size during open-loop automated handwriting (Mai et Marquart Exp Brain Res 1999; 128: 224–228). Objective: Our intent was to pinpoint the involvement of the cerebellum in visual on-line adaptation of letter size. Methods: In eight male right-handed subjects (age: 25–39 years), we used H2 15O positron emission tomography (PET) to measure regional cerebral blood flow (rCBF) during a handwriting task and a control condition (holding the pen). The written trace was recorded using a graphics tablet. Visual feedback of handwriting was continuously provided on a computer screen. In each trial, participants wrote four pairs of „l„ („ll ll ll ll“) every six seconds. Subjects were asked to match the letter size to the distance between two horizontal lines (˜ 9mm) using the feedback on the screen. The critical manipulation was an unpredictable change in the gain of vertical feedback during each trial. After the subject had written the first pair of „l“, the gain of vertical feedback was either reduced to 66% (1/3 of the trials), enlarged to 133% (1/3 of the trials) or remained the same (1/3 of the trials). Subjects were asked to maintain stroke size as specified by the horizontal lines. Kinematics of handwriting were analyzed using CS software (MedCom, Munich, Germany). The mean number of inversions in velocity (NIV) was calculated for each scan to quantify the degree of automation. Three scans per condition were acquired in each subject. PET data were analyzed with SPM99 (Wellcome Dept. of Neuroscience, ION, UCL, UK). Using a general linear model, the NIV were treated as a regressor to model the changes in rCBF associated with the mean NIV per scan (p<0.001, uncorrected). Results: Regression analysis revealed a positive linear relationship between the rCBF in several cerebellar clusters and the degree of automation during handwriting (i.e., as indexed by the NIV). These cerebellar areas were located bilaterally in the medial and paramedian aspects of the superior cerebellum. Conclusion: Our findings demonstrate that a high degree of open-loop performance is associated with an increased cerebellar activity during a task requiring rapid integration of visual feedback. This activity pattern points to a significant role of the cerebellum in on-line adjustments of movement amplitude during skilled hand movements.