Talk2Me: Automated linguistic data collection for personal assessment.

Area of Science:

• Linguistics
• Cognitive Science
• Computational Linguistics

Background:

• Cognitive changes, particularly in language, are early indicators of neurological conditions.
• Longitudinal monitoring of language is crucial for understanding cognitive decline.
• Existing methods for linguistic data acquisition can be resource-intensive and geographically limited.

Purpose of the Study:

• To introduce Talk2Me, a novel web portal for remote, longitudinal linguistic data acquisition.
• To provide baseline data and scoring functions for various linguistic tasks to facilitate research.
• To release a comprehensive open-source software package for extracting a wide range of linguistic features.

Main Methods:

• Development of the Talk2Me web portal with diverse linguistic tasks.
• Establishment of baselines for scoring functions across and within tasks.
• Creation of an open-source package for extracting ~2000 lexico-syntactic, acoustic, and semantic features.

Main Results:

• Talk2Me enables remote and longitudinal data collection for language research.
• Provided baselines aid in the interpretation of linguistic data and normative modeling.
• The open-source package offers extensive linguistic feature extraction capabilities, applicable to various studies.
• Demonstrated utility in identifying dementia through baseline classification.

Conclusions:

• Talk2Me and its associated feature extraction software represent a significant advancement for cognitive and linguistic research.
• The publicly released resources facilitate large-scale, normative studies on language and cognitive change.
• This comprehensive toolkit empowers researchers to analyze linguistic data with unprecedented depth and breadth.