A fast neuronal signal-sensitive continuous-wave near-infrared imaging system.

We have developed a continuous-wave near-infrared imaging system to measure fast neuronal signals. We used a simultaneous sampling method with a separate high-speed analog-to-digital converter for each input channel, which provides a much larger point sample in a digital lock-in algorithm, higher temporal resolution, and lower crosstalk among detected channels. Without any analog filter, digital lock-in detection with a large point sample suppresses noise excellently, making the system less complex and offering better flexibility. In addition, using a custom-made collimator, more photons can reach the brain tissue due to the smaller divergence angle. Performance analysis shows high detection sensitivity (on the order of 0.1 pW) and high temporal resolution (~50 Hz, 48 channels). Simulation experiments show that intensity changes on the order of 0.01% can be resolved by our instrument when averaging over approximately 500 stimuli. In vivo experiments over the motor cortex show that our instrument can detect fast neuronal signals in the human brain.