Related Experiment Video
Updated: Oct 5, 2025
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Screening for cognitive impairment with the Montreal Cognitive Assessment in Spanish patients with minor stroke or
J M Ramírez-Moreno1, S Bartolomé Alberca2, P Muñoz Vega3
1Departamento de Ciencias Biomédicas, Universidad de Extremadura, Badajoz, Spain; Centro de Ictus, Sección de Neurología, Hospital Universitario Infanta Cristina, Badajoz, Spain; Grupo de Investigación Multidisciplinar de Extremadura (GRIMEX), Villanueva de la Serena, Spain.
Objective:
The symptoms of minor stroke and transient ischemic attack (TIA) are temporary and mild. Despite the transient nature of the focal symptoms and the absence of visible brain lesions in some patients, many experience persistent cognitive problems subsequently. We aimed to establish the discriminant capacity of the Montreal Cognitive Assessment (MoCA) in screening for cognitive impairment (CI) within 90 days of TIA.
Methods:
A total of 50 patients with minor stroke or TIA were recruited. Patients were administered the MoCA test and a formal neuropsychological test battery. CI was defined clinically according to neuropsychological test findings.
Results:
The average age of recruited patients was 57.7±8.0 years; 70.0% were men; all patients had completed at least primary education. Thirty-seven patients (74.0%) presented CI. Receiver operating characteristic curve analysis obtained an optimal MoCA cut-off point of 25 for discriminating between patients with CI and those without, with an area under the curve of 0.835 (95% confidence interval [CI], 0.720-0.949), sensitivity of 78.4% (95% CI, 62.8%-88.6%), specificity of 76.9% (95% CI, 49.7%-91.8%), positive predictive value of 90.6% (95% CI, 81.0%-95.6%), and negative predictive value of 55.6% (95% CI, 39.5%-70.4%).
Conclusions:
More than half of the patients presented CI as determined by the formal battery of neuropsychological tests. A MoCA cut-off point of 25 is sufficiently sensitive and specific for detecting CI after minor stroke or TIA, and may be implemented as a screening technique in routine clinical practice.
