Related Experiment Video
Updated: Jul 7, 2026

Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
Published on: December 23, 2025
Changes in conversational pattern as a clinical trial outcome: a proof-of-concept study using the I-CONECT data
Liu Chen1, Chao-Yi Wu1, Hiroko H Dodge1
1Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, United States.
Natural language processing can detect cognitive decline. A 6-month conversational intervention significantly reduced semantic noise in individuals with mild cognitive impairment, bringing their scores closer to those with normal cognition.
Area of Science:
- Natural Language Processing
- Cognitive Science
- Computational Linguistics
Background:
- Natural language processing (NLP) offers novel linguistic measures for detecting subtle cognitive decline.
- Current NLP analyses often rely on constrained tasks like picture descriptions, limiting real-world applicability.
- Semantic noise (SN) presents a novel approach to quantify conversational patterns and monitor cognitive changes.
Purpose of the Study:
- To investigate the efficacy of a behavioral intervention in altering semantic noise (SN) in individuals with mild cognitive impairment (MCI).
- To assess the sensitivity of SN as a linguistic measure for tracking cognitive changes over time.
- To explore the potential of using spontaneous conversation analysis for cognitive monitoring.
Main Methods:
- Analysis of text from semi-structured conversations within the I-CONECT project.
- Participants included cognitively normal (NC) and mild cognitive impairment (MCI) individuals in an intervention group.
- Mean semantic noise (MSN) was measured at baseline (week 2) and 6 months (week 24) of a 6-month conversational intervention (4 sessions/week).
Main Results:
- Baseline MSN was significantly higher in the MCI group (2.09) compared to the NC group (1.87) (p=.024).
- After 6 months, MSN in the MCI group significantly decreased (p=.009), approaching levels observed in the NC group (1.95).
- No significant change in MSN was observed in the NC group (p=.426), and post-intervention MCI and NC groups did not differ significantly (p=.280).
Conclusions:
- A 6-month conversational intervention normalized semantic noise (SN) in individuals with mild cognitive impairment (MCI), making their conversational patterns resemble those of cognitively normal (NC) individuals.
- Linguistic measures derived from naturalistic conversations are sensitive to cognitive changes.
- Further research into extracting linguistic measurements from spontaneous conversations can aid in daily cognitive monitoring.
More Related Videos
08:36Dynamic Inter-subject Functional Connectivity Reveals Moment-to-Moment Brain Network Configurations Driven by Continuous or Communication Paradigms
Published on: March 21, 2019
07:12Using Informational Connectivity to Measure the Synchronous Emergence of fMRI Multi-voxel Information Across Time
Published on: July 1, 2014
Related Concept Videos
Clinical Trials
There are four phases in a clinical trial. A phase one...
Clinical Trials: Overview
Longitudinal Studies
Therapeutic Communication
Verbal communication depends on language or a prescribed way of using words so that people can share information effectively. The critical aspects of verbal...
Observational Studies
There are three types of observational studies – Prospective, retrospective, and cross-sectional.
Prospective Study
Prospective studies, also known as longitudinal or cohort studies, are carried out by collecting future data from groups sharing similar characteristics. One example of...
Longitudinal Research