We present a method for real-time parallel compression of neural signals, achieving a 400:1 compression ratio while preserving temporal and spike count accuracy. This approach utilizes a monostable multivibrator-based circuit model that compresses neural activity within a 20 ms time frame into a single voltage point, effectively representing the entire signal before digitization. Using patch-clamp recordings from pyramidal cells in layer 4 of the visual cortex in young guinea pigs, this technique compresses a 499.95 ms dataset sampled at 20 kHz into 25 data points. Decompressed signals retained all 10 action potentials (APs) from the original dataset, with a worst-case timing deviation of 4.15 ms. Spike timing accuracy was further validated using the Victor-Purpura distance, resulting in a distance of 2.217 and a normalized value of 0.2217, respectively. By significantly reducing the data size required for transmission, this method demonstrates the potential to support multi-channel microelectrode arrays (MEAs) in wireless systems with limited throughput, enabling scalable and efficient neural recording technologies.