Our aim was to determine the biomechanical properties of the normal human cervical spine under physiological static loads. The three-dimensional displacements under three pure moments: flexion-extension, left-right lateral bending and left-right axial torsion--were measured in 56 intact functional spinal units (FSUs) taken from between C2 and C7 in 29 human cadavers. For each mode of loading, load-displacement curves were plotted. Then we calculated each neutral zone, range of motion, neutral zone ratio, ratio of coupled motion, limit moment and secant stiffness. The effects of intervertebral disc degeneration and the disc level were also taken into account by the analysis of variance. Our results adequately demonstrated both the non-linearity of load-displacement curves and the neutral zone of the cervical spine in three-dimensional space. At the same time, we found statistically that the stiffness in the three planes are significantly different, as are the stiffnesses in lateral bending of successive different FSUs. However, significant differences of stiffness in different states of disc degeneration were only found in right lateral bending. There were significant differences between levels in ratio of coupled motion under both lateral bending and axial torsion. The loading cycle conditions and the biomechanical responses of principal motion of C1-2 are also reported.

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

We wanted to determine the biomechanical properties of the human cervical spine with ligamentous injuries. The three-dimensional displacements under four pure moments--flexion, extension, left-right lateral bending and left-right axial torsion--were measured in 18 functional spinal units (FSUs) taken from 9 human cadaveric cervical spines. The experimentation was first performed with intact FSUs, then a series of ligamentous injuries were artificially created. The same measurements for the same FSUs were repeated after each step of ligamentous injury. For each mode of loading and each step of injury, three-dimensional load-displacement curves were plotted. From these curves, we calculated the following parameters for the principal motion: the neutral zone, the range of motion and the flexibility coefficients. A statistical analysis of these parameters was performed between intact FSUs and different ligamentous injury situations. The analysis of ligamentous injuries shows the consequences of different ligamentous injuries and will help us to assess certain assumptions about clinical stability.