In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) organized in conjunction with the MICCAI 2012 and 2013 conferences. Twenty state-of-the-art tumor segmentation algorithms were applied to a set of 65 multi-contrast MR scans of low- and high-grade glioma patients—manually annotated by up to four raters—and to 65 comparable scans generated using tumor image simulation software. Quantitative evaluations revealed considerable disagreement between the human raters in segmenting various tumor sub-regions (Dice scores in the range 74%–85%), illustrating the difficulty of this task. We found that different algorithms worked best for different sub-regions (reaching performance comparable to human inter-rater variability), but that no single algorithm ranked in the top for all sub-regions simultaneously. Fusing several good algorithms using a hierarchical majority vote yielded segmentations that consistently ranked above all individual algorithms, indicating remaining opportunities for further methodological improvements. The BRATS image data and manual annotations continue to be publicly available through an online evaluation system as an ongoing benchmarking resource.

Here we present a scheme to separate particles according to their characteristic physical properties, including size, charge, polarizability, deformability, surface charge mobility, dielectric features, and local capacitance. Separation is accomplished using a microdevice based on direct current insulator gradient dielectrophoresis that can isolate and concentrate multiple analytes simultaneously at different positions. The device is dependent upon dielectrophoretic and electrokinetic forces incorporating a global longitudinal direct current field as well as using shaped insulating features within the channel to induce local gradients. This design allows for the production of strong local field gradients along a global field causing particles to enter, initially transported through the channel by electrophoresis and electroosmosis (electrokinetics), and to be isolated via repulsive dielectrophoretic forces that are proportional to an exponent of the field gradient. Sulfate-capped polystyrene nano and microparticles (20, 200 nm, and 1 μm) were used as probes to demonstrate the influence of channel geometry and applied longitudinal field on separation behavior. These results are consistent with models using similar channel geometry and indicate that specific particulate species can be isolated within a distinct portion of the device, whereas concentrating particles by factors from 10(3) to 10(6).

Insulator based dielectrophoresis is powerful tool for separating and charactering particles, yet it is limited by a lack of quantitative characterizations. Here this limitation is addressed by employing a method capable of quantifying the dielectrophoretic mobility of particles. Using streak-based velocimetry the particle properties are deduced from their motion in a microfluidic channel with a constant electric field gradient. From this approach the dielectrophoretic mobility of 1 μm polystyrene particles was found to be −2 ± 0.4 × 10−8 cm4/(V2·s). In the future, such quantitative treatment will allow for the elucidation of unique insights and rational design of devices.

The performance of fallible counters is investigated in the context of pacemaker-counter models of interval timing. Failure to reliably transmit signals from one stage of a counter to the next generates periodicity in mean and variance of counts registered, with means power functions of input and standard deviations approximately proportional to the means (Weber's law). The transition diagrams and matrices of the counter are self-similar: Their eigenvalues have a fractal form and closely approximate Julia sets. The distributions of counts registered and of hitting times approximate Weibull densities, which provide the foundation for a signal-detection model of discrimination. Different schemes for weighting the values of each stage may be established by conditioning. As higher order stages of a cascade come on-line the veridicality of lower order stages degrades, leading to scale-invariance in error. The capacity of a counter is more likely to be limited by fallible transmission between stages than by a paucity of stages. Probabilities of successful transmission between stages of a binary counter around 0.98 yield predictions consistent with performance in temporal discrimination and production and with channel capacities for identification of unidimensional stimuli.

This study presents an unbiased high-resolution separation and characterization of NSPC subpopulations using direct current insulator-based dielectrophoresis.