The digital tree drawing test (dTDT) is a newly developed screening tool for the early detection of Alzheimer's disease. It is performed with a digitizing pen, recording each pen stroke with temporal and spatial precision. It was hypothesized that movement characteristics recorded during the painting process contribute to the identification of patients with mild cognitive impairment (MCI) and early dementia of the Alzheimer type (eDAT). The study population consisted of 187 participants (67 healthy controls, 64 MCI, and 56 eDAT patients) with a mean age of 68.6±10.6 years. Between-group comparisons of the dTDT-variables were conducted with analysis of variance. The diagnostic power of dTDT variables was analyzed with stepwise logistic regressions and areas under curve (AUC) of receiver operating control curves. Cognitively impaired persons used less colors and line widths and changed them less often than healthy subjects (p-values ≤0.05). Compared to control, eDAT patients had larger not-painting periods, were slower, and their pictures had less contrast, image size, and complexity (p-values ≤0.01). Logistic regression models of stepwise selected dTDT variables resulted in an AUC of 0.84 (95% confidence interval (CI) [0.79, 0.90], sensitivity = 0.78, specificity = 0.77) for discriminating healthy subjects from all cognitive impaired, an AUC of 0.77. (95% CI [0.69; 0.85], sensitivity = 0.56, specificity = 0.83) for discriminating healthy controls from MCI patients and an AUC of 0.90 (95% CI [0.84, 0.96], sensitivity = 0.86, specificity = 0.82) for discriminating controls from eDAT patients. The results suggest that digital recording of pen-stroke data during the drawing process can contribute to the screening of cognitive impaired patients.
Keywords: Digital device; digital tree drawing test; drawing characteristics; early alzheimer’s disease; logistic regression; mild cognitive impairment; neuropsychological drawing test; screening test.