The study of brain dynamics has been approached from different mathematical strategies with the aim to obtain more information of the bioelectrical signals coming from an electroencephalogram (EEG). Although several of these tools try to conciliate the fact of using linear approaches to study a non-linear phenomenon, during the last years a set of complementary and alternative approaches has been used to mine deeper into the nature of the brain dynamics and its correlates with experience, thoughts and actions. One of these approaches comes from chaos theory and fractal geometry and considers a model where brain activity, as a time series of an electrical potential variation (EEG) recorded from the scalp, can be assumed as a dynamical state that moves in the range of 1/f α (fractal) noise. In this model it is possible to see brain dynamics as a functional state who moves from more chaotic or unpredictable dynamics (Brown or Gaussian uncorrelated noise), to quasi-chaotic or statistical noise (fractal or self-similar noise). The amount of order into the chaotic EEG background must be an indicator of the organization of brain procedural resources dealing with different circumstances. At the same time, it must reflect the degree of inter-individual differences and intra-individual variability in the way each different brain works. To test this hypothesis we implemented a study where EEG activity was recorded during the performing of a simple visual intelligent test (Raven test, abbreviated version of 15 questions) in a set of 10 adults to study their similarities and differences in the rendering of the test (estimated range of IQ) and in the process of solving the easy and the difficult part of the test. We estimated the Hurst exponent and the fractal dimension of the time series for each of the 14 EEG channels (Emotiv-Epoc® BCI headset) and searched for correlations and consistencies in the values of H and the difficulty level of the cognitive task. We found that order tend to emerge from a chaotic background when brain focuses on problem solving, rising the degree of predictability, self-similarity and persistent behavior of the EEG signal

The cognitive approach of learning agrees in that behavior of the individual correlates with certain complex and dynamic mental processing operations modulated by certain internal mechanisms learned or improved during life. Applied to university learning, the demand towards students no longer focuses only in the results of the process but, ideally, in the development of individual styles based on cognitive preferences and learning strategies. Several tests such as Kolb and VAK or Hermann dominances allow distinguishing some learning styles such as: Divergent, Assimilator, Convergent and Accommodator; Visual; Kinesthetic; and Auditory. Meanwhile, by using electrophysiological tools like an electroencephalographic (EEG) recording it is now possible to measure individual differences in the varying electrical activity present in the cerebral cortex while a student face a cognitive problem or performs a test of intelligence. In this study, we tested 20 students from their early years of engineering to be classified in three systems of learning styles. The students were then subjected to a test of intelligence (Raven test, abbreviated version of 15 questions) with increasing levels of complexity. The time length to solve the test taken for students previously classified in four main styles yielded by the Kolb test, were analyzed. Then we grouped them by correspondences with the visual, auditory and kinesthetic predominance yielded by the VAK test. The response time was measured and the absolute frequency of response time for each question was calculated. During the Raven test execution an electroencephalogram with Emotiv Epoc Brain-Computer Interface of 14 channels was recorded. After subtracting baseline and eliminating common artifacts with the help of EEGLAB toolbox under Matlab, we obtained clean signals for theta, alpha, beta and gamma bands. Using the set of learning styles classificatory tools and the electrophysiological analysis we started to look for variability and consistencies that support or rise new questions about adequate usefulness of common used psycho-cognitive and behavioral learning styles tests. In our sample we found relatively low discriminative resolution from all learning styles test applied, but a promising research field to study electrical brain activity phenomenology associated with learning and solving problem strategies of the brain

The objective of this study is to design a decision-making tool to carry out a sequence of intermediate processing analytic steps in order to render a graphic visualization of significant correlations between pairs of EEG channels. We submitted a set of students to an abbreviated version of a visual intelligent test while we recorded EEG activity covering frontal, temporal, parietal and occipital areas of the brain cortex through the scalp. The graphs generated show correlations between specific ranges of R Spearman values and pairs of electrodes relevant for the particular study needed. We used this tool to make a descriptive classification of subjects according to the inter-channels correlation maps generated by our decision-making tool. We called correlation “dominances” to the areas defined for a high concentration of significant correlations. This term has been used previously as a way to classify styles of people's thinking-acting in relation to four descriptive referential brain areas. We contrasted our correlation dominances results with this psychological-questionnaire tool in search of consistencies. To design the brain connectivity module, we used Rapid Application Development (RAD) methodology. The tool has been developed in Matlab to analyze signal data ranges delta, theta, alpha, beta, and gamma; all these obtained from the EEG cleaned off artifacts, pre-processed and separated into bands using EEGLAB. Pearson correlation is utilized to detect synchronic connectivity dominances (correlations) through the brain areas. The results are presented in terms of cross-correlation maps, or correlation matrices, between the 14 channels of EEG signal for different EEG frequency bands

This paper presents a conceptual model for an Intelligent System built to support the scheduling for an underground mine in order to supervise its operation. The system is composed by a Simulation Model linked to a Knowledge Based System designed by means of hierarchical, colored and temporal Petri Nets. Simulation Model allows simulating the operation of the production, reduction and transport levels in the mine. Knowledge Based System is activated by events produced in daily operations and yields the results of registered events and the actions taken to solve the problem, generating operation rules. The proposed model allows different types of mine operations and scenarios providing data for decision-making. The system helps to evaluate different policies for programming the activities in the mine thus seeking to enlarge the equipment productivity. The model also allows the feasibility assessment of the Daily Master Plan based on the input data of the simulation model.