Whole-bundle nerve recording is an easy technique to gauge peripheral nerve activities. However, this conventional technique fails to detail individual fiber activities. Aiming for a signal resolution at the single-fiber level, we established a relatively novel experimental model so-called ‘oligofiber recordings’. In vitro splanchnic sympathetic nerve-thoracic spinal cord preparations were obtained from Sprague-Dawley neonatal rats. Whole-bundle splanchnic nerves were incubated in a glass micropipette containing 0.5% collagenase for 90 min. The dissociated nerve fascicles were then brought into a small caliber micropipette for electrical signal recordings. Oligofiber activities that displayed several distinct spike potential waveforms were often achieved. Automation of spike sorting was based on spike waveform features using a series of custom-made LabVIEW programs incorporated with MATLAB scripts. Spike data were automatically grouped by k-means clustering algorithms followed by verification of their waveform homogeneity by principal component analysis (PCA). Dissimilar waveforms with exceeding Hotelling’s T2 distances from the cluster centroids were retrieved by a subtraction algorithm (SA), which partially resolved overlapped spikes. Both T2-selected and SA-retrieved spikes were combined as unit activities. To evaluate if unit activities truly originated from single fibers, we examined the probability distribution of interspike intervals (ISIs) and determined if a change of waveform features was a function of their preceding ISIs. Using the oligofiber recording techniques, we could simultaneously examine, on average, ~3 single fiber activities per experiment. With some modifications, these techniques should be applicable to any peripheral nerve recordings.