According to the Scalar Expectancy Theory, humans are equipped with a biological internal clock, possibly modulated by attention and arousal. Both emotions and pain are arousing and can absorb attentional resources, thus causing distortions of temporal perception. The aims of the present single-event fMRI study were to investigate: a) whether observation of facial expressions of pain interferes with time production; and b) the neural network subserving this kind of temporal distortions. Thirty healthy volunteers took part in the study. Subjects were asked to perform a temporal production task and a concurrent gender discrimination task, while viewing faces of unknown people with either pain-related or neutral expressions. Behavioural data showed temporal underestimation (i.e., longer produced intervals) during implicit pain expression processing; this was accompanied by increased activity of right middle temporal gyrus, a region known to be active during the perception of emotional and painful faces. Psycho-Physiological Interaction analyses showed that: 1) the activity of middle temporal gyrus was positively related to that of areas previously reported to play a role in timing: left primary motor cortex, middle cingulate cortex, supplementary motor area, right anterior insula, inferior frontal gyrus, bilateral cerebellum and basal ganglia; 2) the functional connectivity of supplementary motor area with several frontal regions, anterior cingulate cortex and right angular gyrus was correlated to the produced interval during painful expression processing. Our data support the hypothesis that observing emotional expressions distorts subjective time perception through the interaction of the neural network subserving processing of facial expressions with the brain network involved in timing. Within this frame, middle temporal gyrus appears to be the key region of the interplay between the two neural systems.