A 10 nV/rt Hz noise level 32-channel neural impedance sensing ASIC is presented for the application of local activation imaging in nerve section. It is increasingly known that the monitoring and control of nerve signals can improve physical and mental health. Major nerves, such as the vagus nerve and the sciatic nerve, consist of a bundle of fascicles. Therefore, to accurately control a particular application without any side effects, we need to know exactly which fascicle was activated. The only way to find locally activated fascicle is to use electrical impedance tomography (EIT). The ASIC to be introduced is designed for neural EIT applications. A neural impedance sensing ASIC was implemented using CMOS 180-nm process technology. The integrated input referred noise was calculated to be 0.46 μVrms (noise floor 10.3 nVrms/rt Hz) in the measured noise spectrum. At an input of 80 mV, the squared correlation coefficient for linear regression was 0.99998. The amplification gain uniformity of 32 channels was in the range of + 0.23% and - 0.29%. Using the resistor phantom, the simplest model of nerve, it was verified that a single readout channel could detect a signal-to- noise ratio of 75.6 dB or more. Through the reservoir phantom, real-time EIT images were reconstructed at a rate of 8.3 frames per second. The developed ASIC has been applied to in vivo experiments with rat sciatic nerves, and signal processing is currently underway to obtain activated nerve cross-sectional images. The developed ASIC was also applied to in-vivo experiments with rat sciatic nerves, and signal processing is currently underway to obtain locally activated nerve cross-sectional images.