Abstract
It is widely accepted that information derived from analyzing speech (theacoustic signal) and language production (words and sentences) serves as auseful window into the health of an individual's cognitive ability. In fact,most neuropsychological batteries used in cognitive assessment have a componentrelated to speech and language where clinicians elicit speech from patients forsubjective evaluation across a broad set of dimensions. With advances in speechsignal processing and natural language processing, there has been recentinterest in developing tools to detect more subtle changes incognitive-linguistic function. This work relies on extracting a set of featuresfrom recorded and transcribed speech for objective assessments of cognition,early diagnosis of neurological disease, and objective tracking of diseaseafter diagnosis. In this paper we provide a review of existing speech andlanguage features used in this domain, discuss their clinical application, andhighlight their advantages and disadvantages. Broadly speaking, the review issplit into two categories: language features based on natural languageprocessing and speech features based on speech signal processing. Within eachcategory, we consider features that aim to measure complementary dimensions ofcognitive-linguistics, including language diversity, syntactic complexity,semantic coherence, and timing. We conclude the review with a proposal of newresearch directions to further advance the field.