In this study we investigated a question about a possibility to form auditory-visual associations during discrimination of speech sounds (phonemes) when they presented simultaneously with visual stimuli. It was used two types of static images--geometrical figures and images of hand gestures. We judge about formation of auditory-visual association by behavioural disturbance (reaction time, percentage of wrong reactions) after changes in experimental protocol viz. partial substitution of preceiding pairs of stimuli for the new combinations of previous images and sounds. It seems possible to get pronounced auditory-visual association when using images of hand gestures in combinnation with phoneme sounds. On the contrary, employment of images of geometrical figures instead of gestures produced only traces of auditory-visual association. We suggest that it is possible to explain the observed differencies by involvement of additional neuronal populations in auditory-visual assosiation that are preferentially responsive to biological motion.