[Clinic analysis of femtosecond laser-assisted cataract surgery].

OBJECTIVE: To evaluate the safety and effectiveness of femtosecond laser-assisted cataract surgery. METHODS: The interventional case series enrolled 314 eyes undergoing cataract surgery in Xiamen Ophthalmic Center between April and December 2013, patients were randomized to femtosecond laser-assisted cataract surgery(153 eyes) and conventional phacoemulsification cataract surgery(161 eyes). The clincic parameters of preoperation, during operation and postoperation were compared and statistically studied. Quantitative data were analyzed using the analysis of variance, independent t tests. Qualitative datawere analyzed using the crosstabs analysis chi-square test or the Fisher's exact test. RESULTS: The patient's demographics parameters(age, gender, IOP, axial length, mean keratometry, AC depth, Cataract grade) did not differ significantly between groups (P>0.05) . The CDE showing 4.78% (3.18%-8.88%) and EPT showing 14.05 s (10.07-20.85 s) in the laser group were significantly lower than the CDE showing 8.82% (6.01%-19.16%) and EPT showing 23.65 s (18.36-46.96 s) in the conventional group (z=2.30, 2.91; P<0.05) . The relative diameter and circularity in laser-assisted capusulotomies were significantly more accurate (t=2.58, 3.92; P<0.05). The corneal endothelial cell loss showing 73.50 (-69.51-111.03)/mm(2) was significantly lower in the laser group than the loss of 118.06 (53.55-299.03)/mm(2) in the conventional group 1 month postoperatively (z=2.44; P<0.05). Postoperative anterial chamber flare was significantly greater in the conventional group at 1 day of 18.81 (13.32-20.23) ph/ms in laser group, 24.51(16.38-32.18)ph/m in conventional group and at 1 month of 13.01(9.23-16.28) ph/ms in laser group, 18.05(12.37-24.97) ph/ms in conventional group than the laser group (z=2.40, 2.31; P<0.05). There were no severe surgical complications for both groups. CONCLUSIONS: The femtosecond laser-assisted cataract surgery was effective and safe. It reduced EPT and CDE during operation, therefore decreased endothelial damage and postoperative anterior chamber inflammation. It also provided more precise and reproducible capsulotomies.

Function approximation with spiked random networks.

This paper examines the function approximation properties of the "random neural-network model" or GNN. The output of the GNN can be computed from the firing probabilities of selected neurons. We consider a feedforward Bipolar GNN (BGNN) model which has both "positive and negative neurons" in the output layer, and prove that the BGNN is a universal function approximator. Specifically, for any f is an element of C([0, 1]s) and any epsilon>0, we show that there exists a feedforward BGNN which approximates f uniformly with error less than epsilon. We also show that after some appropriate clamping operation on its output, the feedforward GNN is also a universal function approximator.