Understanding and forecasting brain functions is the major challenge of our times. The focus of this endeavor is understanding and forecasting learning events, such as the dynamic adaptation of beams connecting neuronal cards in Edelman’s Theory of Neuronal Group Selection (TNGS). We have conceived, designed and evaluated a new paradigm for constructing and using collective knowledge by Web interactions that we called ViewpointS. By exploiting the similarity with the TNGS we conjecture that it may be metaphorically considered a Collective Brain, especially effective in the case of trans-disciplinary representations. Far from being without doubts, in the paper we present the reasons (and the limits) of our proposal that aims to become a useful integrating tool for future quantitative explorations of individual brain functions as well as of collective wisdom at different degrees of granularity. We are therefore challenging each of the current approaches: the logical one in the semantic Web, the statistical one in mining and deep learning, the social one in recommender systems based on authority and trust; not in each of their own preferred field of operation, rather in their integration weaknesses far from the holistic and dynamic behavior of the human brain.
of Invited Talks Decision Support Systems in Neurodegenerative Diseases Diagnosis, Treatment and Management
There have been significant efforts in the direction of improving accuracy in detecting human action using skeleton joints. Determining actions in a noisy environment is still challenging since the Cartesian coordinate of the skeleton joints provided by depth sense camera depends on camera position and skeleton position. In a few of the human-computer interaction applications, skeleton position and camera position keep changing. The proposed method recommends using relative positional values instead of actual Cartesian coordinate values. Recent advancements in the Convolution Neural Network (CNN) help us achieve higher prediction accuracy using image format input. To represent skeleton joints in image format, we need to represent skeleton information in matrix form with equal height and width. With some depth sense cameras, the number of skeleton joints provided is limited, and we need to depend on relative positional values to have a matrix representation of skeleton joints. We can show near the state-of-the-art performance on MSR 3-Dimensional (3D) data and the new representation of skeleton joints. We have used image shifting instead of interpolation between frames, which helps us have state-of-the-art performance.
Herein, we are proposing a novel and radical pipeline that will facilitate the repurposing of approved drugs in an unprecedented way that will eventually yield invaluable insights and results that will aid the pharma-medical domain to tackle many more pathologies using weaponry that has already been approved, is safe for the public, is very rapid relatively to conventional drug design and requires no further significant investment to be made. The ultimate goal is to develop a novel clinical concept and establish a computer-aided pipeline that will facilitate and rationalize the repurposing of approved drugs, orphan drugs and generics. The end result of the described pipeline is a competitive and reliable software that will be made available for the scientific community.
Dyslexia is a disorder characterized by reading impairment and it is affected by both neurocognitive and developmental factors. The aim of this paper is to present the latest scientific evidence which reveal different neurocognitive interventions that can improve brain function in order for children with reading difficulties and dyslexia to reach long-term reading improvement. As well as to evaluate the level of accomplished results of each neurocognitive intervention. Understanding the difficulties children with dyslexia face on their academic and personal life, it is important to expand research on how to successfully intervene in the process of learning. A variety of interventions such as visual perceptual training, intensive reading training, motor cognition training and several others have been shown to have positive effects on dyslexic children.
Virtual reality, an immersing and interactive information communication technology, is changing the fundamental of medical education by creating more astonishing vivid realistic clinical environment and cases. Here, we proposed a virtual reality medical education platform with intelligent real-time emotion evaluation and help system, named Hypocrates+. By this platform, medical students can gain their medical knowledge and experience through well designed virtual clinical cases and environments with less frustration. Experiments shows that the overall mean frustration before getting the help was 0.53 and the mean frustration after was 0.50; better performance is related to lower frustration. We concluded that Hypocrates+ is a good start for developing a popular virtual medical education platform.
The proliferation of Internet has introduced new technological advances into digital education. One of them is the Massive Open Online Courses (MOOCs). MOOCs are online learning environments offering educational programs to large numbers of geographically dispersed students, free of charge. The rapid development of MOOCs leads to investigate their provided quality of learning, consisting of a combination of factors such as the development life cycle of MOOCs, the quality criteria and the involved members. In view of the above, this paper presents QUMMEL (Quality Model for MOOCs and E-Learning) which is a novel reference model for assessing the quality in e-learning and MOOCs. QUMMEL is a three-dimensional model, being consisted of distinct phases, perspectives and roles. It represents a holistic approach for ensuring quality in either a MOOC or an e-learning environment in terms of pedagogical, technological and strategic perspectives. The evaluation results of applying the QUMMEL in the development of a MOOC are very promising and can offer a fertile ground to foster quality in e-learning.
This study is designed to test the hypothesis of whether writing and typing can be detected as different patterns within the same cognitive task. We designed this pilot study with five frequently-conducted learning-related tasks. We used the four electrode Muse Headset. Sixteen healthy subjects participated in this six-session experiment. In each session, we instructed them to conduct five different one-minute tasks, including reading, copying by writing, copying by typing, answering a question by writing and answering a question by typing. We compared the performance of classifiers of different categories within the same context, including the same users, the same feature extraction approach, and the same training and testing split. Most of the machine learning and deep learning algorithms could correctly classify the five tasks (20% by chance), the best algorithms achieved an accuracy for individual subjects of up to 70% for within session training and 44% for between session training.
In recent years there has been a great increase in research on cognitive enhancement issues of students and adults. For this purpose, different interventions and a variety of strategies have been proposed. According to previous studies, mathematical problem solving is a related intervention, as it has been argued that frequent involvement in such process is positively related to the improvement of one’s cognition. This article presents the design and the first results of an exploratory qualitative research that investigate the effect of mathematical problem solving on the cognitive enhancement of the elderly. The findings of this research confirm the aforementioned belief and implicate that a study accompanied by the appropriate neuroimaging techniques should be conducted.
In this study we propose a new machine learning classification method to distinguish brain activity patterns for healthy subjects. We used ElectroEncephaloGraphic (EEG) data associated with five userdefined mental tasks. We collected a data set using the Muse headband with four EEG electrodes (TP9, AF7, AF8, and TP10). Sixteen healthy subjects participated in this six-session experiment. In each session, we instructed them to conduct five different one-minute tasks, we abbreviated the tasks as think, count, recall, breathe and draw. After dealing with noise and outliers, we first fairly compared the performance of existing classifiers, including linear classifiers, non-linear Bayesian classifiers, nearest-neighbor classifiers, ensemble methods, and deep learning, with the same settings. Among these, Random Forest (RF), and Long Short-Term Memory (LSTM) outperform others. We then introduced a new ensemble classifier called Time Continuity Voting (TCV), combining these top two. The timewise cross-validation results showed that TCV could correctly classify the five tasks (20% by chance) with an accuracy of 70% which is at least 6% higher than the top individual classifiers.
Virtual Patients (VPs) and Affective Learning are promising tools in the research for enhancement of educational efficacy. The purpose of this research is to explore the effect of medical error and the emotions it evokes on learning by using these tools. A sample of four undergraduate medical students took part in the experiment. Each student managed two VPs, while connected to biosignal recording devices (heart rate, skin conductance, brain activity, pupil diameter). Before and after managing the VPs, each student filled in a sheet for each VP, containing one competence self-evaluation question and one knowledge assessment question, so that possible differences in their responses could be spotted. The results showed that: a) medical errors with VPs can probably have slight effects on the affect state as indicated by the biosignals, b) some of the errors made by the students with virtual patients did contribute to learning – for the rest of the errors there were no control questions. This research was unable to establish a correlation between the affect state following an error and the learning outcome.
Brain-based learning is the understanding of the human brain functions and its application in educational environments for meaningful learning. Brain-Based Learning adapts the learning process based on the function of human brain, providing a learner-centered tutoring environment. To this direction, a personalized brain-based quiz game was developed applying the principles of brain-based learning and Marzano Taxonomy for promoting meaningful learning and improving students’ higher order cognitive functions. Thus, the system adapts quiz content based on student knowledge level, emotional state and the learning goal set. Regarding the control grouped pre-test and post-test experiment; the results reveal that this approach has a positive effect on students’ performance, outperforming the traditional e-assessment systems.
Mild cognitive impairment (MCI) affects an increasing number of the elderly. MCI is in many cases the first sign of dementia, which is one of the main reasons for disability in the elderly worldwide. In people with MCI, it is essential to regularly evaluate their cognitive status due to the uncertain course of symptoms, that is not completely achieved with traditional assessment procedures. Technological advances, especially in mobile and portable devices, have made possible the delivery of new forms of cognitive training and new methods for cognitive evaluation of healthy older adults or of individuals with MCI. Easy to use and engaging applications can serve as a new means of providing enjoyable cognitive interventions that offer the required commitment of their users. The ability of long-term, remote and autonomous monitoring of the cognitive course of individuals in combination with new technological measures, new specialized algorithms and built-in sensors in their mobile devices offer new data streams to clinicians and the possibility of timely and personalized intervention. This review presents existing cognitive interventions and evaluation studies in adults with or without cognitive impairment, that utilized mobile devices in this context. Their encouraging results provide the breeding ground for the proposal of developing new interventions on future research, that will take advantage of the potentials and prospects of mobile digital devices technology.
Affective Computing is one of the most active research topics in education. Increased interest in emotion recognition through text channels makes sentiment analysis (i.e., the Natural Language Processing task of determining the valence in texts) a state-of-the-practice tool. Considering the domain-dependent nature of sentiment analysis as well as the heterogeneity of the educational domain, development of robust sentiment classifiers requires an in-depth understanding of the effect of the teaching-learning context on model performance. This work investigates machine learning-based sentiment classification on datasets comprised of student posts in forums, pertaining to two different academic courses. Different dataset configurations were tested, aiming to compare performance: i) between single-course and multi-course classifiers, ii) between in-course and cross-course classification. A sentiment classifier was built for each course, exhibiting a fair performance. However, classification performance dramatically decreased, when the two models were transferred between courses. Additionally, classifiers trained on a mixture of courses underperformed single-course classifiers. Findings suggested that sentiment analysis is a course-dependent task and, as a rule of thumb, less but course-specific information results in more effective models than more but non-specialized information.
Despite the potential benefits that Assistive Technologies (AT) could provide for People with Down Syndrome (PDS) and Intellectual Disabilities (PID), research and implementation of emerging AT for learning has mainly focused on developing adaptive accessible solutions and evaluating cognitive function. Unlike the parallel, but equally important role, of all stakeholders and factors involved in PDS and PID support and learning including medical practitioners, families and professionals, have not received adequate attention. This paper describes an interdisciplinary collaboration and multilevel evaluation focusing both on investigating the potential improvement of PDS memory performance after participating in cognitive training sessions. Cognitive assessment in 20 PDS working memory performance was conducted, whereas questionnaires were distributed to 29 relevant stakeholders evaluating the possible correlation between educational feasibility and usability of the AT introduced. Overall, the results showed that there was a significant improvement on PDS memory performance and significant positive correlation in between different variables of AT educational feasibility and usability.
Artificial Intelligence (AI) is among the top technological trends that are expected to shape the future of teaching and learning. Chatbots, or conversational agents, are software that can interact with a human user turn by turn using natural language, and they are considered as exemplar utilization of AI and Machine Learning (ML) in education. Despite the increased interest around educational applications of chatbots, there is a lack of chatbot integration into formal learning settings. This work introduces the case of a Conversational Virtual Patient (CVP) prototype for training decision-making skills regarding thromboembolism in medical students. In contrast to typical virtual patients which rely on predefined, multiple-choice input, the proposed CVP employs Natural Language Processing and ML techniques to allow students to formulate their own utterances and interact with the virtual patient in natural language. In view of employing participatory design and co-creating open access chatbots for healthcare curricula, this CVP prototype will form the basis for the co-creation sessions, by familiarizing stakeholders with the chatbot technology and enabling the requirement elicitation process. Future steps in further co-developing the CVP prototype are discussed.
Emotion greatly affects learning. Affective states, such as motivation, interest and attention, have been identified to cause changes in brain and body activity. Heart Rate (HR), Electro-Dermal Activity (EDA), and Electroencephalography (EEG) reflect physiological expressions of the human body that change according to emotional changes. In reverse, changes of bio-signal recordings can be linked to emotional changes. Virtual/Mixed Reality (V/MR) applications can be used in medical education to enhance learning. This work is a proof of application study of wearable, bio-sensor based affect detection in a learning processes, that includes the Microsoft HoloLens V/MR platform. Wearable sensors for HR, EDA and EEG signals recordings were used during two educational scenarios run by a medical doctor. The first was a conventional scenario-based Virtual Patient case for the participant's bio-signal baselines canonization. The second was a V/MR exploratory educational neuroanatomy resource. After pre-processing and averaging, the HR and EDA recordings displayed a considerable increase during the V/MR case against the baseline. The alpha rhythm, of the EEG, had a borderline degrease and the theta over beta ration a borderline increase. These results indicate an increased attention/concentration state. They also demonstrate that the usage of bio-sensors assist in the detection the emotional state and could provide real-time, affective learning analytics using V/MR in medical education.
In an attempt to make the therapeutic aspect less aversive, more attractive and engaging, virtual reality, an increasingly popular application in healthcare, offers an interesting alternative to pharmacological treatments. Positive emotions may improve the cognitive abilities of people suffering from cognitive impairment. Virtual reality can provide immersive and efficient relaxation tool. This paper presents an experiment where 19 people with Subjective Cognitive Decline (SCD) were immersed in a virtual environment representing a savannah. The hypothesis is that the environment may help them reducing their frustration by relaxing. Participants’ brain activity was recorded using the Emotiv Epoc headset and the virtual savannah experience lasted 10 min. Results suggest that frustration decreased when participants were surrounded by the virtual savannah and that the positive effects continued afterwards.
Serendipitous discovery, invention or artistic creation are among the most exciting and utmost relevant phenomena strongly related to human learning . At the moment, there are very few measurable criteria helping to understand and foster serendipity. In other papers [ 1 – 3 ] we have presented, discussed and exemplified a new paradigm/model/method/system/environment – called ViewpointS – that represents our efforts to overcome many current existing limitations in generic Information Systems or search engines (e.g.: Google) as well as in other social media (e.g.: recommender systems) offering information retrieval solutions based on the proximity of available resources. We also have also exposed how ViewpointS may facilitate serendipitous discovery in an unprecedented way. In this paper, we wish to further motivate this last conjecture by proposing to explore two main research directions that did not convey sufficient attention by previous researchers (in particular those active in recommender systems): 1. assessing brain states in order to understand and forecast serendipitous human learning events triggered by emotions; 2. enhancing collective wisdom , since Human-Computer Interactions do not occur today between a human and a single machine (or algorithm), but within a community of humans and machines that continuously update “knowledge” beyond the scene. Both directions (assessment of brain states, collective wisdom) are currently on separate ways; we propose to combine them within one unified approach called ViewpointS.
Distance education has been an alternative to traditional education for several decades. However, during the pandemic, distance education was the only solution that could be applied, in order to continue the educational process and at the same time to protect the health of students and their families. This form is described as Emergency Remote Teaching, because although it is based on online teaching practices, it does not fully use the principles of distance education and does not include educational design and operation models developed in the context of distance education. Despite the limited integration of the principles of distance education, efforts are made to enhance learning, develop knowledge and skills, and modify students’ attitudes through the educational process. In this context, the role of educational neuroscience is important and the use of the results of relevant research is crucial in order to enhance learning. After a brief description of the framework of implementation of Emergency Remote Teaching, this article presents difficulties that arise from this implementation and describes ways in which these difficulties can be reduced, according to findings in the field of educational neuroscience research.
