When analysing facial expressions, it is not only the final expression itself, but also its formation that plays an important role when attempting to decipher its meaning. Currently in research there are two techniques for describing the dynamics of facial expression; quantitative and temporal based analysis. Quantitative-based techniques attempt to determine the amplitude of the expression in terms of intensity levels, where the levels correspond to some measure of the extent to which the expression is present on the face. Temporal-based techniques split the expression into three temporal phases (onset-apex-offset). In this paper we focus on the temporal aspects of facial expression formation, describing our research into applying a non-linear manifold extraction technique for modelling these temporal phases. We present initial results of our technique for modelling the temporal aspects of both posed and spontaneous facial expressions.