Electro- and magneto-encephalography are functional neuroimaging modalities characterised by their ability to quantify dynamic spatiotemporal activity within the brain. However, the visualisation techniques used to illustrate these effects are currently limited to single- or multi-channel time series plots, topographic scalp maps and orthographic cross-sections of the spatiotemporal data structure. Whilst these methods each have their own strength and weaknesses, they are only able to show a subset of the data and are suboptimal at articulating one or both of the space-time components.Here, we propose Porthole and Stormcloud, a set of data visualisation tools which can automatically generate context appropriate graphics for both print and screen with the following graphical capabilities: Animated two-dimensional scalp maps with dynamic timeline annotation and optional user interaction; Three-dimensional construction of discrete clusters within sparse spatiotemporal volumes, rendered with 'cloud-like' appearance and augmented by cross-sectional scalp maps indicating local maxima. These publicly available tools were designed specifically for visualisation of M/EEG spatiotemporal statistical parametric maps, however, we also demonstrate alternate use cases of posterior probability maps and weight maps produced by machine learning classifiers. In principle, the methods employed here are transferrable to visualisation of any spatiotemporal image.
Keywords: EEG; MEG; Machine learning; Space-time cube; Statistical parametric mapping; Visualisation.