Fig. 1 summarizes the experimental methodology that is detailed
in four levels: the classification task, classification paradigm,
ordering strategy and training mode. Three classification
tasks were studied: Discriminating real from artifact red signals
enabling Down syndrome detection (“red” in Fig. 1), discriminating
real from artifact green signals enabling Patau syndrome
detection (“green” in Fig. 1) and accomplishing both tasks simultaneously
(i.e., a four-class classification task) enabling the
detection of both syndromes (“4-class” in Fig. 1). We implemented
two classification paradigms. The first was a monolithic
classifier, i.e., a single fuzzy ARTMAP classifier that associated