Floating-Point Approximation Enabling Cost-Effective and High-Precision Digital Implementation of FitzHugh-Nagumo Neural Networks.

The study of neuron interactions and hardware implementations are crucial research directions in neuroscience, particularly in developing large-scale biological neural networks. The FitzHugh-Nagumo (FHN) model is a popular neuron model with highly biological plausibility, but its complexity makes it difficult to apply at scale. This paper presents a cost-saving and improved precision approximation algorithm for the digital implementation of the FHN model. By converting the computational data into floating-point numbers, the original multiplication calculations are replaced by adding the floating-point exponent part and fitting the mantissa part with piecewise linear. In the hardware implementation, shifters and adders are used, greatly reducing resource overhead. Implementing FHN neurons by this approximation calculations on FPGA reduces the normalized root mean square error (RMSE) to 3.5% of the state-of-the-art (SOTA) while maintaining a performance overhead ratio improvement of 1.09 times. Compared to implementations based on approximate multipliers, the proposed method achieves a 20% reduction in error at the cost of a 2.8% increase in overhead.This model gained additional biological properties compared to LIF while reducing the deployment scale by only 9%. Furthermore, the hardware implementation of nine coupled circular networks with eight nodes and directional diffusion was carried out to demonstrate the algorithm's effectiveness on neural networks. The error decreased to 60% compared to the single neuron of the SOTA. This hardware-friendly algorithm allows for the low-cost implementation of high-precision hardware simulation, providing a novel perspective for studying large-scale, biologically plausible neural networks.

Characteristic analysis of pulmonary ground-glass lesions with the help of 64-slice CT technology.

OBJECTIVE: With the help of new technologies like 64-slice spiral CT, including latest AW4.4, 2D nodule comparing and analyzing technology, MPR and 3D technology, MIP technology and the technology of analyzing pulmonary vascular density by the method of perfusion scanning, we performed characteristic analysis of ground-glass opacities (GGO) for the early diagnosis of lung cancer. PATIENTS AND METHODS: We selected 62 patients suspected of lung cancer, whose conventional CT showed that they were patients with GGO. With the help of the new technologies of 64-slice spiral CT provided by GE Company, prospective scans were made and 2 to 4 times of review were arranged. After that, the patients were treated with surgery or needle biopsy to get lesion's pathological results. After several scans, the results including lesion's form, density, blood supply, peripheral sign, doubling time and tissue perfusion were drawn to make a comparison. Based on the results, comparative analysis on GGO's characteristics was made from morphological and functional perspectives. RESULTS: 41 patients (66.1%) were pathologically diagnosed with cancer, 10 were diagnosed with inflammation, 7 with fibrosis, and 4 with edema, hemorrhage and other lesions. The comparisons were made between the tumor groups' clinical manifestations (sex, age, symptoms including smoking, coughing, and expectoration), and the difference had no statistical significance (p>0.05). Conventional CT scan showed that the shape of GGO was irregular and it showed spiculated sign and pleural indentation. The proportion of the patients with vessel convergence in the tumor group was significantly higher than that of the non-tumor group (p<0.05). However, the comparisons between lesions' number, location (superior lobe of the right lung), diameter, edge (blur) and lobulation were made to get a difference ratio (p>0.05) which had no statistical significance. Tumor group's doubling time was significantly short, and its perfusion parameters including BF, BV, MTT, and PS were increased significantly (p<0.05). CONCLUSIONS: The new 64-slice CT technology has great value in the diagnosis of the tumorous GGO.

[Experimental studies of hyperlipidemia in rats and mice given a hyperlipidemic diet].

Hyperlipidemia in rats and mice was induced by hyperlipidemic diet containing 1.5% cholesterol, 0.25% cholic acid, 7.5% lard and 90.75% essential diet. After feeding for 7 days, the serum cholesterol levels in the experimental group are about 5 times higher than that in the control group. It appears that there exist similarities between rats and mice and that these models may be used for screening tests of hypolipidemic drugs.