Differentiating Patients at the Memory Clinic With Simple Reaction Time Variables: A Predictive Modeling Approach...

Area of Science:

• Neuroscience
• Cognitive Science
• Machine Learning Applications in Healthcare

Background:

• Mild Cognitive Impairment (MCI) and dementia are increasing conditions with overlapping symptoms, necessitating accurate differential diagnostics.
• Distinguishing between Subjective Cognitive Impairment (SCI), MCI, and dementia is crucial for prognoses, treatment, and patient management.
• Current diagnostic methods can be costly, highlighting the need for more efficient and accessible diagnostic tools.

Purpose of the Study:

• To investigate the efficacy of simple reaction time (SRT) variables, alone and combined with psychometric tests, in differentiating cognitive conditions.
• To develop and evaluate machine learning models for classifying Subjective Cognitive Impairment (SCI), Mild Cognitive Impairment (MCI), and dementia.
• To assess the cost-efficiency and potential clinical utility of SRT-based diagnostics.

Main Methods:

• A supervised machine learning approach was employed using nine variables extracted from simple reaction time (SRT) data.
• One hundred twenty elderly patients (65-95 years) from a memory clinic were assessed using a freely available SRT task and standard psychometric tests.
• Support vector machine models were trained and validated using recursive feature elimination and cross-validation to classify SCI, MCI, and dementia.

Main Results:

• Machine learning models incorporating SRT variables achieved good accuracy in classifying MCI/dementia (Accuracy = 0.806) and some merit for SCI/MCI/dementia (Accuracy = 0.650).
• Five out of seven selected predictors in the final models were derived from SRT data, indicating its significant contribution.
• The developed models are available in a free application for research and educational purposes.

Conclusions:

• Simple reaction time variables show promise as a supplementary tool, alongside psychometric tests, for differentiating MCI, dementia, and SCI in memory clinic patients.
• SRT-based diagnostics offer a potentially cost-efficient approach to differential diagnosis in aging populations.
• Further external validation is required, but SRT integration into diagnostic support systems is a promising avenue for cognitive assessment.