Aim: We aimed to create a set of pictures unambiguously named by Czech one word and to divide them into two categories according to the difficulty of naming for patients with cognitive impairment. Patients and methods: First, a list of words accompanied by black and white drawings was created.
These pictures drawn in different ways were repeatedly presented to elderly persons and patients with cognitive impairment during five attempts of preliminary research. The result was expressed by the percentage of name agreement of the picture name and its expected name.
The pictures which did not meet certain criteria were excluded out of an intial set of 173 pictures. The last set of 70 pictures was presented to a sample of 305 elderly people in two forms, paper (n = 135) and electronic form (n = 170).
An additional 70 patients with a variety of cognitive disorders, primarily with Alzheimer's disease, named the same set of pictures in written paper form for comparative purposes with the rest of the sample. They were also examined by the Mini-Mental State Examination (MMSE) (21 +/- 5 points) or the Montreal Cognitive Assessment (MoCA) (20 +/- 4 points).
Results: The main outcome includes 14 pictures that were difficult to name by patients with cognitive impairment and easily named by elderly people without cognitive deficit - Eskimo, crane, thermometer, totem, mirror, cake, stamp, magnifying glass, traffic light, robot, waterfall, airship, submarine and chessboard. Pictures with more than 90% agreement of correct naming in both healthy and cognitively impaired samples were identified as easy to name.
They included barrel, pencil, moon, glasses, comb, key, pram, table, roller, chair and 27 other pictures. Conclusion: Pictures that were difficult to name will be an important source for creating new screening tests and aids for discrimination of cognitively normal elderly individuals from those with cognitive impairment.
On the contrary, easy-to-name pictures will be suitable for testing visual memory where the learning phase will not be affected by naming errors.