Open Access

Sibylle Robens, Thomas Ostermann, Petra Heymann, Stephan Müller, Christoph Laske, Ulrich Elbing

• Color characteristics in combination with gray level co-occurrence matrix (GLCM) texture features of digital drawings were compared between cognitive healthy subjects (Control, n = 67) and individualsColor characteristics in combination with gray level co-occurrence matrix (GLCM) texture features of digital drawings were compared between cognitive healthy subjects (Control, n = 67) and individuals clinically diagnosed with amnestic mild cognitive impairment (aMCI, n = 32) or early Alzheimer’s dementia (AD, n = 56). It was hypothesized that these variables contribute to the detection of cognitive impairments. Between subject groups comparisons of texture entropy, homogeneity, correlation and image size were conducted were performed with Chi-Square and Kruskal-Wallis tests. The diagnostic power of combining all texture features as explanatory variables was analyzed with a logistic regression model and the area under curve (AUC) of the corresponding receiver operating control (ROC) curve was calculated to discriminate best between healthy and cognitive impaired subjects. Texture and color features differed significantly between subject groups. The AUC for discriminating the control group from patients with early AD was equal 0.86 (95% CI [0.80, 0.93], sensitivity = .80, specificity = .79), and the AUC for discriminating between healthy subjects and all cognitive impaired equal 0.82. (95% CI [0.75; 0.89], sensitivity = .76, specificity = .79) for discriminating healthy controls from MCI patients. Although the study results are very promising, further validation is needed, especially with a larger sample size.…