Fibrotic encapsulation of implanted biomedical devices represents the highest cause of device failure across numerous fields, including orthopedics, breast implants and neural electrodes. While most research has focused on the immune system for driving the foreign body response to devices, fibroblasts are emerging as another key regulator of tissue response. In the present study, we evaluated the effects of porous device design on the activation of primary human dermal fibroblasts into myofibroblasts, a contractile cell that is a hallmark of fibrosis, over four weeks in culture. In particular two factors were studied 1) the introduction of nanoparticles, as they are increasingly being used for device functionalization, and 2) the polymer matrix, which dictates chemistry, mechanics and degradation. The addition of tantalum oxide (TaOx) nanoparticles (0-20wt%) had a minimal effect on fibroblasts, significantly down-regulating expression of alpha smooth muscle actin (SMA) (0.31 +/- 0.1), likely driven by the increase in nanoscale surface roughness. The polymer matrix of the device caused significant changes to myofibroblast activation, with the fast degrading polymer, poly(lactide co-glycolide) (PLGA) 50:50, significantly up-regulating multiple myofibroblast markers, including SMA (5.16 +/- 2.5), vinculin, integrin 1, and integrin 5, compared to non-degrading polycaprolactone (PCL). The effect was due to the release of degradation products, namely lactic acid, which affects cellular metabolism. Together this highlights that device design affects biological response immediately post-implantation in ways that impact the ultimate success or failure of biomedical devices.
The Hidden Power of Normalization: Exponential Capacity Control in Deep Neural Networks
arXiv:2511.00958v1 Announce Type: cross Abstract: Normalization methods are fundamental components of modern deep neural networks (DNNs). Empirically, they are known to stabilize optimization dynamics and