Exploring the Interplay between Digital Physiological Signals and Cognitive Functioning: A Long-term Observational Study
Introduction
Cognitive decline, an alarming concern among aging populations, often manifests as amnestic mild cognitive impairment (MCI), a precursor to Alzheimer’s disease and other dementias. Early detection and intervention are paramount in managing cognitive decline, and advancements in wearable sensor technology offer promising avenues for monitoring cognitive health. This comprehensive study delves into the relationship between digital physiological signals collected from a wearable wristband and cognitive performance in older adults with MCI.
Methods
Data and Participants
The study involved 30 older adults (aged 50-70) diagnosed with amnestic MCI and 10 age-matched cognitively normal individuals. Data were meticulously gathered from a single-arm clinical trial conducted in Singapore between November 2021 and August 2022. Participants were equipped with an Empatica-E4 wristband, a sophisticated device that continuously recorded physiological signals, including triaxial acceleration, skin temperature, electrodermal activity (EDA), and blood volume pulse (BVP). Cognitive assessments were diligently conducted at baseline and after a 10-week intervention.
Wearable Signal Processing and Feature Extraction
Wearable sensor data underwent meticulous processing and transformation into a comprehensive set of physiological features using NeuroKit2, a specialized Python package designed for neurophysiological signal processing. The extracted features encompassed time-domain and frequency-domain measures of heart rate variability (HRV), EDA, temperature, and PPG-based measures. Following careful scrutiny, redundant and unreliable features were eliminated, resulting in a refined set of 106 features ready for further analysis.
Design of Statistical Analysis and Predictive Modeling
To capture the dynamic evolution of cognitive performance over time, cognitive performance scores were imputed for regular 10-day intervals using a Gompertz curve, a mathematical function that effectively models growth and decay processes. Physiological features were skillfully centered around individual median values, a crucial step in mitigating between-individual differences and ensuring data comparability. Pearson and Spearman correlations, robust statistical techniques, were employed to assess the intricate associations between physiological measures and cognitive functioning. Furthermore, linear mixed-effect regression models, sophisticated statistical models that account for both fixed and random effects, were meticulously employed to rigorously test the associations between cognitive measures and digital physiological features. Additionally, supervised machine learning models, powerful algorithms capable of learning from data, were diligently trained to accurately predict cognitive performance scores based on physiological features.
Results
Correlations between Physiological Measures and Cognitive Functioning
Correlational analyses, statistical methods that unveil relationships between variables, revealed a symphony of significant associations between several digital physiological features and cognitive performance measures. Notably, higher HRV measures, indicative of a healthy heart, were harmoniously linked with enhanced cognitive functioning. Skin temperature, a reflection of thermoregulatory processes, and EDA, a measure of sympathetic nervous system activity, also exhibited positive correlations with cognitive performance. Conversely, higher resting heart rate, a marker of cardiovascular strain, and PPG-based measures of HRV were found to be in discord with cognitive performance, suggesting a potential link between cardiovascular health and cognitive decline.
Predictive Ability of Digital Physiological Features
Machine learning models, harnessing the power of data-driven algorithms, demonstrated the remarkable predictive ability of digital physiological features in determining cognitive performance scores. The meticulously crafted models achieved Pearson correlation coefficients ranging from 0.5 to 0.7, indicating a moderate to strong relationship between physiological features and cognitive functioning. This predictive prowess was comparable to that of a baseline model based solely on age and sex, underscoring the additional predictive value offered by digital physiological features beyond basic demographics.
Conclusion
This groundbreaking study provides compelling evidence for the intricate association between digital physiological signals and cognitive functioning in older adults with MCI. The findings illuminate the potential of wearable sensor technology in monitoring cognitive health and paving the way for early detection of cognitive decline. While these results are promising, further research with larger sample sizes and longitudinal follow-ups is warranted to validate these findings and explore the potential of digital physiological features as reliable biomarkers for cognitive impairment.
Call to Action
As we continue to unravel the complex relationship between digital physiological signals and cognitive functioning, we invite you to join the conversation. Share your thoughts, insights, and experiences in the comments section below. Together, we can contribute to a deeper understanding of cognitive health and empower individuals to take proactive steps in preserving their cognitive well-being.