AUTHOR=Garai Kitti , Adam Zoltan , Herczeg Robert , Katai Emese , Nagy Tamas , Pal Szilard , Gyenesei Attila , Pongracz Judit E. , Wilhelm Marta , Kvell Krisztian 

TITLE=Artificial Neural Network Correlation and Biostatistics Evaluation of Physiological and Molecular Parameters in Healthy Young Individuals Performing Regular Exercise

JOURNAL=Frontiers in Physiology

VOLUME=Volume 10 - 2019

YEAR=2019

URL=https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2019.01242

DOI=10.3389/fphys.2019.01242

ISSN=1664-042X

ABSTRACT=Studies support that regular physical activity decelerates senescence-related decline of physiological and molecular parameters in the elderly. We have addressed the other end of this spectrum: enrolled healthy and young, previously inactive individuals to a six-month long personal trainer-guided lifestyle program.
We have measured physiological and molecular parameters focusing on their segregation (differentiating high- and low responders) and their correlation with physical activity (sedentary status). Cluster analysis helped to both distinguish individuals with high- or low physical activity and differentiate high- and low-responders of each parameter. The assessed cardiovascular parameters (heart rate, blood pressure, six-minute walking distance, relative VO2max), body composition parameters (body fat percent, muscle mass percent) metabolic parameters (glucose, insulin, HDL, LDL), immune parameters (cortisol, CRP, lymphocyte counts, hTREC) all showed improvement. Artificial neural network analysis (ANN) revealed correlation efficiencies of physiological and molecular parameters using a concept-free approach. ANN analysis appointed physical activity as the mastermind of molecular level changes.  Besides sedentary status, insulin and hTREC showed the highest segregation among participants. Biostatistics evaluation also supported the schism of participants for their sedentary, insulin and hTREC status. In the future ANN and biostatistics, analysis and evaluation, respectively, may predict individual responses to regular exercise.
Our program reveals that high responder individuals of certain parameters may be low responders of others, highlighting distinct improvement patterns. Our data show that moderate regular physical activity is essential to counteract senescence in young and healthy individuals, despite individual differences in responsiveness and compliance. Such physical activity may not seem important in the everyday living of young and healthy adults, but shall become the basis for healthy ageing.