Elicitation of Neurological Knowledge with ABML Vida Groznik 1 , Matej Guid 1 , Aleksander Sadikov 1 , Martin Moˇ zina 1 , Dejan Georgiev 2 , Veronika Kragelj 3 , Samo Ribariˇ c 3 , Zvezdan Pirtoˇ sek 2 , and Ivan Bratko 1 1 Artificial Intelligence Laboratory, Faculty of Computer and Information Science, University of Ljubljana, Slovenia 2 Department of Neurology, University Medical Centre Ljubljana, Slovenia 3 Faculty of Medicine, University of Ljubljana, Slovenia Abstract. The paper describes the process of knowledge elicitation for a neurological decision support system. To alleviate the difficult problem of knowledge elicitation from data and domain experts, we used a recently developed technique called ABML (Argument Based Machine Learning). The paper demonstrates ABML’s advantage in combining machine learn- ing and expert knowledge. ABML guides the expert to explain critical special cases which cannot be handled automatically by machine learn- ing. This very efficiently reduces the expert’s workload, and combines it with automatically learned knowledge. We developed a decision sup- port system to help the neurologists differentiate between three types of tremors: Parkinsonian, essential, and mixed tremor (co-morbidity). The system is intended to act as a second opinion for the neurologists, and most importantly to help them reduce the number of patients in the “gray area” that require a very costly further examination (DaTSCAN). 1 Introduction and Motivation Essential tremor (ET) is one of the most prevalent movement disorders. [2] It is characterized by postural and kinetic tremor with a frequency between 6 and 12 Hz. ET usually starts in one hand and then spreads to the neck and vo- cal cords, giving the characteristic clinical picture of the disease. Parkinsonian tremor (PT), on the other hand, is a resting tremor classically described as “pill rolling” tremor with a frequency between 4 and 6 Hz. It is one of the major signs of Parkinson’s disease (PD), which also includes bradykinesia, rigidity and postural instability. Although distinct clinical entities, ET is very often misdiag- nosed as PT. [10] Co-existence of both disorders is also possible [9], additionally complicating the differential diagnosis of tremors. Digitalized spiralography is a quantitative method of tremor assessment [5], based on spiral drawing on a digital tablet. In addition to precise measurement of tremor frequency, spiralography describes tremors with additional parameters — these, together with physical neurological examination, offer new means to differentiate between numerous types of tremors [4,5], including ET and PT. M. Peleg, N. Lavraˇ c, and C. Combi (Eds.): AIME 2011, LNAI 6747, pp. 14–23, 2011. c  Springer-Verlag Berlin Heidelberg 2011