Daily activities such as aerobic movements and athletic events found effective in mitigating bone loss as it promotes osteogenesis. Computational model considered normal strain, or strain energy density as a stimulus to predict site specific osteogenesis. This model, however, fails to predict site specific osteogenesis as cortical bone surfaces exhibit different remodelling rate to mechanical loading. Remodelling rate or mineral apposition rate depends upon the loading parameters such as loading cycle, frequency, and magnitude of strain. Therefore, the present study aims to develop an adaptive neuro-fuzzy inference system (ANFIS) model for finding a robust relationship between loading parameters like strain magnitude, frequency, and cycle, and a bone remodelling parameter i.e. mineral apposition rate (MAR). The model is trained, tested, and checked with the experimental data. The results indicate that ANFIS model outperformed state of the art Artificial Neural Network (ANN) models during the prediction of MAR at periosteal and endosteal surface. A strong corelation R2 = 0.92 and R2 = 0.97 was observed at 70% of the input data at periosteal and endosteal surface respectively. Result concludes that endosteal surface was more promisable as compared to periosteal surface in predicting accurate MAR. The outcomes of present study may be used to precisely predict site-specific osteogenesis in cortical bone as function of loading parameters.
Read full abstract