Automated Scoring of the Mini-Cog Psychological Test

Abstract

Neuropsychological test are used to identify the presence of brain injuries or any associated function deficits. These tests evaluate functioning in a number of areas which include intelligence, attention, memory and language etc. Psychologists generally conduct these tests with a single person in an office environment that is free from distractions and disturbances. Drawing tasks have achieved a central location in neuropsychological assessment and contains information about the presence of neurological disorders. Drawing tests have been long used by doctors and researchers for early detection of psychological and neurological disorders. Mini-Cog test is a combination of the Recall test and Clock Drawing Test and the combined scoring of these tests identifies whether the person is demented or not. This test is divided into three different steps. Firstly, the patient is shown three different words. Furthermore, the patient has to draw digits on a piece of paper where a clock circle is already drawn. In the last phase, patient is asked to recall the words which were shown in the first step. This test is currently scored and interpreted manually which increases the work load of practitioners and also consumes a significant amount of time. The purpose of the proposed project is to develop an efficient, reliable and accurate diagnostic tool which will reduce the work load of psychologists by automatically screening the subjects using the Mini-Cog test. This study proposes the application of image analysis techniques to automatically score a subset of hand drawn clock images in the test. The developed system evaluated on a number of hand-drawn clocks realized promising results.