Roller bearings are critical components in rotating machinery and are often the focus of condition monitoring. Early fault detection is essential for ensuring safe operation, minimizing downtime, and enhancing profitability. Acoustic Emission (AE) monitoring has emerged as a popular non-destructive testing technique for identifying bearing defects. This paper builds upon previous research by presenting a new method that utilizes spiking neural network (SNN) – the third generation of neural networks – for bearing fault detection. AE signals were collected from bearings with induced faults and transformed into spectrograms using the short-time Fourier transform. These spectrograms were used to train an SNN model, with some data reserved for validation. Results demonstrate that a simple SNN architecture can accurately classify bearing faults across a range of operational speeds, underlining the potential of SNNs for effective, time-frequency-based bearing fault diagnosis.