The main objective of this deliverable is to describe the theoretical background of the approach, which the adaptation and personalization mechanisms within the MaTHiSiS ecosystem will adopt. This document describes the mechanism for establishing correspondence between learner behaviour, tracked during the learning experience, and competence over learned concepts (SLAs), with regard to maintain the learner in an optimal affect state, thereby maximising the rate of learning. A theoretical framework for correlating affective and cognitive learner states to SLA competence weights is devised, in order to support the adaptation mechanisms within MaTHiSiS. It is further enhanced taking into account learning styles and parameters encapsulated in the learner’s educational material consumption history and learning progress, in order to provide the theoretical foundation for long-term personalisation. Acknowledging the role of collaborative learning and advantages of social flow, the MaTHiSiS platform also caters for adaptation in multi-learner scenarios.