Heart Rate and its Variability From Short-Term ECG Recordings as Potential Biomarkers for Detecting Mild Cognitive Impairment.

Abstract

Background: Alterations in Heart Rate (HR) and Heart Rate Variability (HRV) reflect autonomic dysfunction associated with neurodegeneration making them biomarkers suitable for detecting Mild Cognitive Impairment (MCI). Methods: The study involves 297 urban Indian participants [48.48% (144) were male and 51.51% (153) were female]. MCI was detected in 19.19% (57) of participants and the rest, 80.8% (240) of them were healthy. ECG recordings spanning 10 s were collected and R-peaks were detected. Machine learning algorithms like were employed to further validate the features. Results: The mean of R-to-R (NN) intervals (P = .0021), the RMS of NN intervals (P = .0014), the SDNN (P = .0192) and the RMSSD (P = .0206) values differ significantly between MCI and non-MCI. Machine learning classifiers, SVM, DA, and NB show a high accuracy of 80.801% on RMS feature input. Conclusion: HR and its variability can be considered potential biomarkers for detecting MCI.