Kinematic Correlates of Early Speech Motor Changes in Cognitively Intact APOE-ε4 Carriers: A Preliminary Study Using a Color-Word Interference Task

Mehrdad Dadgostar^1,2Lindsay C. Hanford³Jordan R. Green^1,4Brian D. Richburg¹Averi Taylor Cannon¹Nelson Barnett¹David H. Salat^2,5,6Steven E. Arnold⁷Marziye Eshghi^1*

• ¹MGH Institute of Health Professions, Boston, United States
• ²Massachusetts General Hospital Athinoula A Martinos Center for Biomedical Imaging, Charlestown, United States
• ³Harvard University Department of Psychology, Cambridge, United States
• ⁴Harvard University Speech and Hearing Biosciences and Technology, Cambridge, United States
• ⁵Harvard Medical School Department of Radiology, Boston, United States
• ⁶Department of Radiology, VA Boston Healthcare System, West Roxbury, United States
• ⁷Harvard Medical School Department of Neurology, Boston, United States

Alzheimer’s disease (AD) is the most prevalent form of dementia and a major public health challenge. In the absence of a cure, accurate and innovative early diagnostic methods are essential for proactive life and healthcare planning. Speech metrics have shown promising potential for identifying individuals with mild cognitive impairment (MCI) and AD, prompting investigation into whether speech motor features can detect elevated risk even prior to cognitive decline. This preliminary study examined whether speech kinematic features measured during a color-word interference task could distinguish cognitively normal APOE-ε4 carriers (ε4⁺) from non-carriers (ε4⁻). Lip movement properties were extracted across pre-, during-, and post-interference sentence segments. Descriptive statistics and independent t-tests were conducted to examine group-level trends in lip movement duration, average speed, and range. Although no group differences reached statistical significance after accounting for multiple testing, several features showed moderate effect sizes. These should be interpreted cautiously as preliminary trends that may reflect sampling variability rather than true group differences. A support vector machine model with a degree-2 polynomial kernel achieved 87.5% accuracy in classifying APOE-ε4 status using three features: lip movement duration prior to interference, average lip speed during interference, and the change in lip movement range from pre- to during-interference segments. These features reflect subtle differences between the two groups in baseline motor planning, susceptibility to cognitive-motor interference, and articulatory adaptability. The model’s performance metrics (precision 88.90%, sensitivity 88.90%, specificity 85.70%) suggest that speech kinematics may hold promise as an exploratory biomarker of early risk. However, these results are preliminary and require validation in larger, independent samples before any clinical utility can be established.