Response behavior of specimens of the German cockroach (Blattella germanica) exposed to the Ca²⁺ signal inducing chemicals was characterized through implementation of two computational methods: the Fourier transform analysis and artificial neural networks. Ionomycin, thapsigargin, and their solvent (dimethyl sulfoxide) were topically applied to male German cockroaches, and the movement tracks were continuously observed through the image processing system under semi-natural conditions for 4–5 days. The specimens treated with the chemicals revealed different movement patterns: 1) shaky advancement and entanglement of the movement tracks with the ionomycin treatments; 2) continuous, circular movements with the thapsigargin treatments; and 3) shaky turning movements with the dimethyl sulfoxide treatments. The movement tracks in time domain were further analyzed with the two-dimensional fast Fourier transform (2-D FFT). The coefficients of the 2-D FFT efficiently revealed characteristics that resided in the two-dimensional data of the movement tracks in the frequency domain. Subsequently the magnitudes of the coefficients were trained by self-organizing map (SOM) through unsupervised learning. Classification of the different movement patterns was possible with the trained network. The combined use of the 2-D FFT and the SOM could be an alternative tool to automatically monitor behavioral changes in specimens exposed to stimulating chemicals.