Prediction of cognitive impairment using higher order item response theory and machine learning models

Datos Bibliográficos

ID15516751
AutoresLihua Yao (0000-0002-6039-3446, Northwestern University, autor de correspondencia), Yusuke Shono (0000-0002-7006-1816, Claremont Graduate University), Cindy J Nowinski (0000-0001-5608-909X, Northwestern University), Cindy Nowinski, Elizabeth M Dworak (0000-0003-4589-1663, Northwestern University), Aaron Kaat, Aaron J Kaat (0000-0001-8147-1899, Northwestern University), Shirley Chen (0000-0002-8232-3772, Aurora St. Luke's Medical Center), Rebecca Lovett (0000-0003-0169-9485, Northwestern University), Emily Ho (0000-0002-3522-3779, Northwestern University), Laura M Curtis (0000-0003-2380-2201, Northwestern University), Laura Curtis, Michael Wolf (0000-0001-6700-2201, Northwestern University), Richard Gershon (0000-0003-0085-0112, Northwestern University), Julia Yoshino Benavente (0000-0003-0549-1809, Northwestern University)
Año2024
Volumen14
Páginas1297952-1297952
Fecha de publicación2024-03-01
Peer ReviewedSí
Open AccessSí
TipoARTICLE
RevistaFrontiers in Psychiatry (JOURNAL)
Identificadores de la revistaISSN: 1664-0640 • E-ISSN: 1664-0640
EditorialFrontiers Media (PUBLISHER • CH)
DOI10.3389/fpsyt.2023.1297952
PMID38495777
OpenAlexW4392351953
IdiomaEN
Referencias citadas26

Timely detection of cognitive impairment (CI) is critical for the wellbeing of elderly individuals. The MyCog assessment employs two validated iPad-based measures from the NIH Toolbox ® for Assessment of Neurological and Behavioral Function (NIH Toolbox). These measures assess pivotal cognitive domains: Picture Sequence Memory (PSM) for episodic memory and Dimensional Change Card Sort Test (DCCS) for cognitive flexibility. The study involved 86 patients and explored diverse machine learning models to enhance CI prediction. This encompassed traditional classifiers and neural-network-based methods. After 100 bootstrap replications, the Random Forest model stood out, delivering compelling results: precision at 0.803, recall at 0.758, accuracy at 0.902, F1 at 0.742, and specificity at 0.951. Notably, the model incorporated a composite score derived from a 2-parameter higher order item response theory (HOIRT) model that integrated DCCS and PSM assessments. The study's pivotal finding underscores the inadequacy of relying solely on a fixed composite score cutoff point. Instead, it advocates for machine learning models that incorporate HOIRT-derived scores and encompass relevant features such as age. Such an approach promises more effective predictive models for CI, thus advancing early detection and intervention among the elderly

Cognition · Cognitive psychology · Item response theory · Machine learning · Psychometrics · Random forest · Recall · Clinical Psychology · Cognitive Functions and Memory · Computer Science · Dementia and Cognitive Impairment Research · Functional Brain Connectivity Studies · Psychology · Artificial Intelligence