Predictors of real-time fMRI neurofeedback performance and improvement-A machine learning mega-analysis

Real-time fMRI neurofeedback is an increasingly popular neuroimaging technique that allows an individual to gain control over his/her own brain signals, which can lead to improvements in behavior in healthy participants as well as to improvements of clinical symptoms in patient populations. However, a considerably large ratio of participants undergoing neurofeedback training do not learn to control their own brain signals and, consequently, do not benefit from neurofeedback interventions, which limits clinical efficacy of neurofeedback interventions. As neurofeedback success varies between studies and participants, it is important to identify factors that might influence neurofeedback success. Here, for the first time, we employed a big data machine learning approach to investigate the influence of 20 different design-specific (e.g. activity vs. connectivity feedback), region of interest specific (e.g. cortical vs. subcortical) and subject-specific factors (e.g. age) on neurofeedback performance and improvement in 608 participants from 28 independent experiments. With a classification accuracy of 60% (considerably different from chance level), we identified two factors that significantly influenced neurofeedback performance: Both the inclusion of a pre-training no-feedback run before neurofeedback training and neurofeedback training of patients as compared to healthy participants were associated with better neurofeedback performance. The positive effect of pre-training no-feedback runs on neurofeedback performance might be due to the familiarization of participants with the neurofeedback setup and the mental imagery task before neurofeedback training runs. Better performance of patients as compared to healthy participants might be driven by higher motivation of patients, higher ranges for the regulation of dysfunctional brain signals, or a more extensive piloting of clinical experimental paradigms. Due to the large heterogeneity of our dataset, these findings likely generalize across neurofeedback studies, thus providing guidance for designing more efficient neurofeedback studies specifically for improving clinical neurofeedback-based interventions. To facilitate the development of data-driven recommendations for specific design details and subpopulations the field would benefit from stronger engagement in open science research practices and data sharing.

Thymio

Fanny Riedo

Technology is now an important part of our lives. We often see robots cited as the future of education, and reports of their imminent entrance in schools. New projects create buzz in the media and online, but when we look at the actual situation, very few robots are currently used in education, and most of the time, the platform used is the Lego Mindstorms. Why so little diversity? What do robot actually bring to the learning experience? How can we design good educational robots? Hopes are that they bring additional motivation to pupils. Since the use of robots is fun, the learning is supposed to become easier. Robot projects and activities are also expected to foster thinking skills, collaboration, and creative spirit. Finally, there is a need to educate people on technology for two reasons. The first is to break the "black box" image they have of technology, and the second is to encourage them into technical careers. Thanks to the Swiss National Centre of Competence in Research Robotics (NCCR Robotics), we could develop some innovative concepts in educational robotics, and implement one such pedagogical tool. We designed a small wheeled robot with many sensors, and LEDs making its internal state apparent to the user. A simple, white look makes it a neutral base for creating one’s own application, for all age and gender groups. Different user interfaces allow to make it accessible to everybody: • Pre-programmed behaviours that demonstrate its different possibilities • A Visual Programming Language (VPL), without text and based on event-action pairs • The Aseba script language (text-based), with a comprehensive development environment to accompany and inform the user The resulting platform, Thymio II, is completely open-source and open-hardware. It was mass-produced and commercialised at a low cost. This gave the opportunity to evaluate the public’s response to it. We could assess that the robot design is well received and appreciated by different age and gender groups. It seems particularly popular with girls. We analysed the expectations of the different age categories and proposed activities that fitted their specific needs. We could also validate that users of Thymio II learn notions of programming, understand essential concepts such as what sensors are, what is the relationship between the robot, the computer, and the programming environment. With the VPL, they could quickly grasp the meaning of events and event-action pairs. We realised that in spite of the interest it generated, the robot was not used much at home or in schools. We think that there is a need for more guidance and that parallels should be drawn with e-learning for the use at home. In schools, we observed that teachers who use robots are pioneers, who invest time and sometimes money into new technologies out of personal interest. The others do not feel strongly against robotics but are probably discouraged by the lack of institutional injunction, appropriate training, budget, and ready-to-use pedagogical materials. At the end of this work, we conclude by giving a set of guidelines, based on our experience, for the design of educational robots. This project demonstrated very promising results and we believe that it can be a first step toward renewing teaching habits.

EPFL2015

Thymio

Fanny Riedo

EPFL2015