Brain-Based Intelligent Tutoring Systems

我们的目标是受到我们从以前的潜意识学习实验中获得的一些结果的启发。在这些工作中，我们发现脑电波分析可以在一瞬间区分认知推理和直觉推理之间的变化，换句话说，从有意识模式到潜意识模式。这种变化类似于学习者从认知任务切换到游戏时所观察到的变化。他们不是继续寻找解决方案，而是玩弄系统直到找到解决方案，但只是通过反复试验，没有任何认知努力。这导致了知识获取中的一个重要问题，它与大脑状态相对应：参与或脱离。在第一种情况下，学习者有兴趣（动机）去探索可能的解决方案，并可能收敛到最终解决方案。在第二种情况下，学习者对继续完成任务不再感兴趣（这有时解释了游戏的衍生）。我们的实验还强调了两种典型的推理：认知推理和直觉推理。虽然前者通常涉及大脑的意识部分，但后者更多地与潜意识部分联系在一起。有时，解决问题的情况会先后涉及认知和直觉推理，在两种模式之间来回切换。*在这两种情况下，情绪对促进或阻碍学习都起着重要作用。大脑的情绪状态不仅允许信息传递到意识区域，也允许信息传递到潜意识区域，而最后一个区域似乎包含更多的知识，处理信息的速度也更快。正如我们在之前的研究中所看到的，情感部分非常重要，因为它决定了大脑接受或不接受信息。直觉行为更多出现在电子游戏中。例如，在飞行训练或游戏中，由此产生的经验和能力将与即时反应（反射）有关*根据我们之前的工作，我们希望更精确地检测导致投入或脱离投入的情绪条件和大脑状态。这意味着要检测学习发生时的情绪状态（例如学习者是否兴奋）和学习状态（学习目标过高），并测量由此产生的大脑状态（以便更清楚地了解学习的不同步骤和大脑状态的演变）。此外，智能系统可以通过对学习者大脑活动的持续控制，为学习者提供建议，使学习者处于最佳的学习情绪状态。正如我们所看到的，大脑是这个智能系统的中心。智能个人顾问（agent）可以跟踪大脑状态的演变，提供个性化的建议，优化学习。**为了达到这个目标，我们需要探索更多的可能性，通过脑电图信号来检测参与和脱离参与（当学习者无法理解问题并找到解决方案时）的不同步骤。如何准确、实时地检测和监测学习者的情绪状态是当前的主要问题。新兴设备允许学习者和他的环境之间进行直接互动（眼动追踪系统，脑电波与头盔的互动）**我们研究的总体目标是(1)建立一个集成的交互式基于大脑的环境，能够检测认知和直觉学习中参与和脱离的条件，使用视频游戏；(2)建立一个完整的基于大脑的视频游戏学习环境，能够提供最佳的教学策略，以促进学习者的记忆和知识获取。

Our objectives are inspired by some results we have obtained from previous experiments on subconscious learning . In these works we have discovered that brainwaves analysis allowed to distinguish a change, at a moment, between cognitive reasoning and intuitive reasoning, in other words from conscious to subconscious mode. This change is similar to what is observed when learners switch from a cognitive task to gaming. Instead of continuing to find a solution, they play with the system until they encounter a solution but just by trial and error without any cognitive effort. This leads to an important problem in knowledge acquisition which corresponds to a cerebral state: engagement or disengagement. In the first one, the learner is interested (motivated) to explore the possible solutions and could converge to the final solution. In the second one, the learner is no more interested to continue on the task (which sometimes explains the derivation to gaming). Our experiments highlighted also two typical kinds of reasoning: cognitive reasoning and intuitive reasoning. While the first involves generally the conscious part of the brain, the second is more linked to the subconscious part. Sometimes, a problem solving situation will involve successively cognitive and intuitive reasoning, switching back and forth between the two modes. *In both cases emotions play an important role to facilitate or not learning. The emotional conditions of the brain allow for the transmission of information not only to the conscious part but also to subconscious region and this last one seems to contain more knowledge and to process the information at a higher speed. As we have seen in our previous studies, the affective part is extremely important as it prepares the brain to accept or not the information. Intuitive behaviour is more present in video games. For instance, in flight training or games the resulting experience and abilities will concern immediate reactions (reflexes)*Following our previous works, we want to detect more precisely emotional conditions and cerebral states that lead to engagement or disengagement. This implies to detect the emotional condition (for instance a learner is excited or not) and learning condition (the learning objectives are too high) in which learning occurs, and measure the resulting cerebral state (for understanding more clearly the different steps of learning and the evolution of cerebral states). Moreover, a continuous control of learner's brain activity could be used by an intelligent system to give advices to the learner and put him into the best emotional conditions for learning. As we can see the brain is at the center of this intelligent system. An intelligent personal adviser (agent) could follow the evolution of cerebral states and provide personalized advices to optimize learning.**To reach this goal, we need to explore more possibilities to detect the different steps of engagement and disengagement (when a learner cannot understand the problem and find the solution) through EEG signals. The main problem remains to detect and monitor the learner emotional state with precision and in real time. Emerging device allow to have a direct interaction between the learner and his environment (eye tracking systems, brainwaves interaction with helmets)**The general objectives of our research are (1) to build an integrated interactive Brain-based environment able to detect the conditions of engagement and disengagement both in cognitive and intuitive learning,using a video game, and (2) to build a complete Brain-based video game learning environment able to provide the best pedagogical strategies to foster learner's memorization and knowledge acquisition.