A 'Mini Linguistic State Examination' to classify primary progressive aphasia.

Area of Science:

• Neurology
• Linguistics
• Psychometrics

Background:

• Current methods for evaluating primary progressive aphasia (PPA) are often time-consuming, inaccurate, or unsuitable for specific patient groups.
• There is a need for a reliable and efficient clinical tool to assess and subclassify PPA presentations.

Purpose of the Study:

• To introduce and validate the Mini Linguistic State Examination (MLSE), a new clinical test for assessing and subtyping PPA.
• To evaluate the diagnostic accuracy of the MLSE using a novel error classification method and random forest analysis.

Main Methods:

• The MLSE assesses five language domains (motor speech, phonology, semantics, syntax, working memory) using 11 linguistic tasks.
• A novel error classification system was employed to derive a five-dimensional patient profile.
• Random forest classification was used to determine diagnostic accuracy for PPA subtypes.

Main Results:

• The MLSE achieved 96% overall diagnostic accuracy for PPA subtypes (logopenic, semantic, non-fluent variants).
• A decision tree derived from the model correctly classified 91% of unseen participants.
• The MLSE demonstrated excellent concordance with clinical gold standard diagnoses.

Conclusions:

• The Mini Linguistic State Examination is a validated, accurate, and efficient tool for PPA assessment and subtyping.
• Its adoption can enhance clinical consistency, facilitate research screening, and aid in monitoring language deficits.
• The MLSE's straightforward scoring and novel approach offer significant advantages for clinicians and researchers.