Kyung-Ho Shin

Nicotine-induced behavioral sensitization is associated with extracellular dopamine release and expression of c-Fos in the striatum and nucleus accumbens of the rat

It is well known that repeated injections of nicotine produce progressively larger increases in locomotor activity, an effect referred to as behavioral sensitization. This study was carried out to investigate the neural mechanisms underlying nicotine-induced behavioral sensitization using in vivo microdialysis and Fos-like immunohistochemistry (FLI). Rats were given repeated injections of saline or nicotine (0.4 mg/kg s.c., twice daily for 7 days) followed by one challenge injection on the 4th day after the last daily injection. Systemic challenge with nicotine produced a much larger increase in locomotor activity in nicotine-pretreated rats (659.1+/-94.9 counts/2 h) than in saline-pretreated rats (218.1+/-61 counts/2 h). A direct local challenge of nicotine (1 or 5 mM) via a microdialysis probe in the nucleus accumbens or striatum induced a much greater dose-dependent increase of dopamine (DA) output in nicotine-pretreated rats than in saline-pretreated rats. Furthermore, in parallel with the behavioral and biochemical data, systemic challenge with nicotine produced marked Fos-like immunohistochemistry in the nucleus accumbens and the striatum in the nicotine-pretreated rats. Taken together, this study demonstrates that behavioral sensitization is clearly associated with an increase in DA release and activation of Fos-like immunoreactive cells in the striatum and the nucleus accumbens produced by repeated nicotine treatment. Our results strongly suggest that the striatum and the nucleus accumbens may play a major role in nicotine-induced behavioral sensitization. The present results are discussed in terms of the development and expression of nicotine-induced behavioral sensitization.

High fat diet altered the mechanism of energy homeostasis induced by nicotine and withdrawal in C57BL/6 mice

Nicotine treatment has known to produce an inverse relationship between body weight and food intake in rodents. Present study determined the effect of repeated treatment with nicotine and withdrawal in control and obese mice, on: (1) body weight, caloric intake and energy expenditure; (2) hypothalamic neuropeptides mRNA expression; and (3) serum leptin. 21-week-old C57BL/6 mice (n = 65) received nicotine (3.0 mg/kg/day; 2 weeks) and saline (1 ml/kg/day; 2 weeks) subcutaneously. Animals were given either a normal-fat (10% kcal from fat, NF) or a high-fat diet (45% kcal from fat, HF) from the 12th week to 25th week. While, nicotine treatment for 14 days induced an increase in hypothalamic agouti-related protein, cocaine- and amphetamine- regulated transcript, pro-opiomelanocortin mRNA expressions, nicotine also produced a reducing effect in body weight gain and leptin concentration in NF mice. High-fat diet induced obese mice showed a blunted hypothalamic and leptin response to nicotine. Remarkable weight loss in obese mice was mediated not just by decreasing caloric intake, but also by increasing total energy expenditure (EE). During nicotine withdrawal period, weight gain occurred in NF and HF groups, which was ascribed to a decrease in EE rather than changes in caloric intake. Hypothalamic AgRP might play a role for maintaining energy balance under the nicotine-induced negative energy status.

Half-Select Free and Bit-Line Sharing 9T SRAM for Reliable Supply Voltage Scaling

This paper presents a half-select free 9T SRAM to facilitate reliable SRAM operation in the near-threshold voltage region. In the proposed SRAM, the half-select disturbance, which results in instable operations in 6T SRAM cell, can be completely eliminated by adopting cross-access selection of row and column word-lines. To minimize the area overhead of the half-select free 9T SRAM cell, a bit-line and access transistors between the adjacent cells are shared using a symmetric shared node that connects two cells. In addition, a selective pre-charge scheme considering the preferably isolated unselected cells has also been proposed to reduce the dynamic power consumption. The simulation results with the most probable failure point method show that the proposed 9T SRAM cell has a minimum operating voltage (VMIN) of 0.45 V among the half-select free SRAM cells. The test chip with 65-nm CMOS technology shows that the proposed 9T SRAM is fully operated at 0.35 V and 25 °C condition. Under the supply voltages between 0.35 and 1.1 V, the 4-kb SRAM macro is operated between 640 kHz and 560 MHz, respectively. The proposed 9T SRAM shows the best voltage scalability without any assist circuit while maintaining small macro area and fast operation frequency.

Bit-width reduction and customized register for low cost convolutional neural network accelerator

This paper presents a low area and energy efficient hardware accelerator for the deep convolutional neural networks (CNNs). Based on the multiply-accumulate (MAC) based architecture, three design techniques are proposed to reduce the hardware cost of the convolutional computations. First, to reduce the computational bit-width of convolutions, an adaptive bit-width reduction scheme is proposed based on differential input method. The bit-width reduction approach can reduce the 37 % of operation bit-width with almost ignorable CNN accuracy degradation. Second, it has been found that adapting bi-directional filtering window in CNN accelerator can considerably reduce the energy for data movement with much smaller number of memory accesses. To expedite the bi-directional filtering operations, we also propose a bidirectional first-input-first-output (bi-FIFO). With SRAM bit-cell layout manner, the proposed bi-FIFO facilitates fast data re-distribution with area and energy efficiency. To verify the effectiveness of the proposed techniques, the AlexNet accelerator has been designed. The numerical results show that the proposed adaptive bit-width reduction scheme achieves 25.9% and 47.3% of area and energy savings, respectively. The bi-FIFO based accelerator also achieves 33 % improved processing time.