It has been reported [1] that a neural network can be implemented to identify whether students have integrated into their lexicon schemata related concepts by the contents of a school course. Specifically, a neural net is trained to discriminate between successful and unsuccessful students´ semantic priming latencies of schemata related words obtained by a semantic priming study at the beginning and end of a course. This neural network discrimination capacity is based on the idea that once a student has integrated new knowledge in long-term memory then a semantic priming effect is obtained from schemata related words (single word schemata priming; e.g., 2]. The current paper constitutes the first of three documents providing a more in-depth analysis to this approach for cognitive assessment of learning. For instance, the mental representation technique used to obtain natural semantic networks from students and teachers (as opposed to idiosyncratic or artificial semantic nets) to study computer simulated schemata behaviour is put under academic scrutiny. Here, it is argued that statistical properties regarding the kind of semantic net (small world structure, scale free degree) as well as the implicit distributed schema through semantic connectedness among concepts relate to emergent connectionist schemas underlying schemata priming that can be identified by a neural net.