Software filter for detecting spikes superimposed on a fluctuating baseline.

We describe and evaluate a software procedure for detecting and discriminating action potentials superimposed on a fluctuating DC baseline. Using an algorithm implemented in Fortran and Assembly language, spikes are detected by comparing the low-pass filtered first derivative of the signal with a preset threshold constant. This detection algorithm efficiently removes the DC baseline fluctuations usually encountered in sensory receptor recordings. It can function in real-time depending on the specific hardware configuration and program implementation. This approach enables the user to acquire direct DC-amplified signals, minimizing filter distortions of the action potential waveform.