Independent Component Analysis (ICA) is a very common instantiation of the Blind Source Separation (BSS) problem [1]. In the context of biomedical signal analysis, ICA is generally applied to multi-channel recordings of physiological phenomena to de-noise and extract meaningful information underlying the recordings. The novel Spatio-Temporal ICA (ST-ICA) framework [2], uses both spatial and temporal information derived from multi-channel time-series to extract underlying sources. In contrast, the standard implementation of the ICA algorithm generally uses only limited spatial information to inform the separation process. When applied to multiple underlying brain sources, for example, STICA can extract both spectrally and spatially overlapping sources. When applied to the analysis of biomedical images, ICA is usually applied as a means of feature extraction and/ or image classification, de-mixing statistically independent sources from each other. STICA extends this capability by incorporating both spectral, as well as spatial, cues in the data. The biggest problem by far in the application of any ICA method to biomedical signals results from the fact that ICA generally returns as many independent components (ICs) as measurement channels, for a given set of measurements. For the most part these ICs can be at best as independent as possible – i.e. there is usually resulting dependency between the ICs. Also, different ICs are representative of different noise or signal sources underlying the signal measurements, and usually domain knowledge is required to identify the components (or groups of components) as relevant sources. For the most part this means that ICA can never really be totally automated as the source selection process is highly subjective. Here we debate the use of STICA to decompose a series of functional medical images into their underlying constituents (ICs) and see to what extent domain knowledge plays a part in selecting/ identifying extracted components correctly based on joint spectral and/or spatial cues present in the extracted data.