We propose a bio-inspired framework for pattern recognition in the insect brain. Departing from a well-known body of knowledge about the insect brain we investigate what features are required or useful to learn input patterns rapidly and in a stable manner. The underlying plasticity is situated in the insect mushroom bodies and requires a reinforcement signal to associate the stimulus with a proper response. As a proof of con- cept we used our model insect brain to classify the well-known MNIST database of handwritten digits which is a popular benchmark for classi- fiers. We show that the structural organization of the insect brain appears to be suitable for both fast learning of new stimuli and reasonable per- formance in stationary conditions. It, furthermore, is extremely robust to damage to the structures involved in sensory processing. Finally, we also speculate that the level of confidence in a classification decision may be assessed and potentially increased by a dynamical representation of the input information. This additional function of the antennal lobe dy- namics adds to the previous suggestion of temporal decorrelation of input information.