Attribute Visualisation for Computer-Aided Diagnosis: A Case Study

Digitalised spirography is a relatively new computer-assisted method for detection and evaluation of tremors. The task of the patient is to draw an Archimedean spiral on the tablet, and different quantitative parameters (attributes) of the spiral are provided by the computer. The neurologists or a computer decision support system (DSS) use these parameters to assess whether the spiral exhibits signs of Parkinsonian or essential tremor. The goal of the present pilot study is to provide the user (physician or a DSS developer) with a meaningful visualisation of the most important attributes for the particular case. The purpose of such visualisation is threefold: (a) it provides the physician with immediate visual clues to be aware of when assessing the spiral, (b) it can serve as a \"visual debugging aid\" for the developers of a DSS, and (c) it can trigger generation of new domain knowledge. In the paper we demonstrate a visualisation method and its application to all three cases using the Parkinson Check application.