Disorders of the Nervous System
Author: Gonzalo Pérez | Email: perez.gonzalo.n@gmail.com
Gonzalo Pérez1°2°3°
1° Cognitive Neuroscience Center, Universidad de San Andrés, Buenos Aires, Argentina
2° National Scientific and Technical Research Council (CONICET), Buenos Aires, Argentina
3° School of Engineering, University of Buenos Aires, Buenos Aires, Argentina
INTRODUCTION: Verbal fluency assessments in mild cognitive impairment (MCI) are
traditionally limited to valid response counts. This subjective approach constraints analysis to
univariate methods and overlooks which semantic memory dimensions are affected. We
tackled these gaps with a novel automated framework. METHODS: We asked 106
participants (52 with MCI, 54 healthy controls) to perform phonemic and semantic fluency
tasks alongside standard cognitive tests. Word properties and timing features were
automatically extracted and used to (i) discriminate between groups via a generalized linear
model (GLM) and machine learning classification, and (ii) predict anatomo-functional brain
patterns. RESULTS: GLM revealed significant effects for frequency, granularity, length, and
imageability. Classification was maximal (AUC = .80) when combining all automated features,
surpassing cognitive measures (AUC = .71). Frequency and granularity correlated with the
volume of semantic-related regions commonly atrophied in MCI. DISCUSSION: Automated
fluency analyses facilitate MCI detection, capturing fine-grained neurocognitive patterns in the
condition.