Latent information in fluency lists predicts functional decline in persons at risk for Alzheimer disease

Abstract

Objective We constructed random forest classifiers employing either the traditional method of scoring semantic fluency word lists or new methods. These classifiers were then compared in terms of their ability to diagnose Alzheimer disease (AD) or to prognosticate among individuals along the continuum from cognitively normal (CN) through mild cognitive impairment (MCI) to AD. Method Semantic fluency lists from 44 cognitively normal elderly individuals, 80 MCI patients, and 41 AD patients were transcribed into electronic text files and scored by four methods: traditional raw scores, clustering and switching scores, “generalized” versions of clustering and switching, and a method based on independent components analysis (ICA). Random forest classifiers based on raw scores were compared to “augmented” classifiers that incorporated newer scoring methods. Outcome variables included AD diagnosis at baseline, MCI conversion, increase in Clinical Dementia Rating-Sum of Boxes (CDR-SOB) score, or decrease in Financial Capacity Instrument (FCI) score. ROC curves were constructed for each classifier and the area under the curve (AUC) was calculated. We compared AUC between raw and augmented classifiers using Delong’s test and assessed validity and reliability of the augmented classifier. Results Augmented classifiers outperformed classifiers based on raw scores for the outcome measures AD diagnosis (AUC 0.97 vs. 0.95), MCI conversion (AUC 0.91 vs. 0.77), CDR-SOB increase (AUC 0.90 vs. 0.79), and FCI decrease (AUC 0.89 vs. 0.72). Measures of validity and stability over time support the use of the method. Conclusion Latent information in semantic fluency word lists is useful for predicting cognitive and functional decline among elderly individuals at increased risk for developing AD. Modern machine learning methods may incorporate latent information to enhance the diagnostic value of semantic fluency raw scores. These methods could yield information valuable for patient care and clinical trial design with a relatively small investment of time and money.

Advancing Alzheimer’s research: A review of big data promises

“…Other biomarkers to detect AD and MCI included semantic fluency and eye movement. Clark et al constructed random forest classifiers using latent information in semantic fluency word lists to predict cognitive and functional decline (28). Lagun et al.…”

Paradigm shift: semantic memory decline as a biomarker of preclinical Alzheimer's disease

“…A similarly semantic, but somehow even more abstract mechanism, was recently reported in the longitudinal characterisation of the painting skills of a Chinese painter, who showed a progressive impoverishment in his work, concurrent with the progression of AD [16]. These studies are accompanied by other research that has looked at latent aspects of lexical performance in the spectrum from normal ageing to MCI and AD [17,18]. This latter research has shown that older adults outperformed people with clinical AD, its prodromal MCI stage and also young controls in category fluency, producing more words which are later acquired, less typical and less familiar [18].…”

Prediction of remission in obsessive compulsive disorder using a novel machine learning strategy

“…Most recently, support vector regression was used to predict OCD severity from brain structural magnetic resonance imaging (MRI) data (Hoexter et al ., 2013). Random Forests (RF), another machine learning algorithm, has been less widely used in clinical medicine and has seen very little application in neuroimaging or other psychiatric research (see exceptions: Tektonidou et al ., 2011; Arnold et al ., 2012; Gibbons et al ., 2013; Clark et al ., 2014). To our knowledge, RF has not yet been applied to predict remission outcomes in longitudinal psychiatric samples nor to analyze data in anxiety disorder samples.…”