The interaction of brain signals and the Internet of Things (IoT) is an emerging field. The application of electroencephalographic (EEG) signals is mainly divided into two categories: 1) signal analysis for diagnosis and 2) brain-computer interfaces (BCI) that can control devices through brain signals. Some of the applications of EEG are early detection of seizures, wheelchair control, and smart homes. EEG processing is a complex task that requires expertise in fields such as neuroscience and computer science. There are frameworks to facilitate EEG processing, such as MNE-Python. Distributed and parallel processing is essential when dealing with large-scale EEG data or when scalable models are needed. We have reviewed platforms like Webrain and frameworks based on existing Big Data processing tools like Spark to understand the capabilities and solutions proposed to solve these challenges. Based on the findings, we have proposed and implemented a platform based on cloud technology called MindGate. While supporting the main features of previous works, our proposed platform is developed based on four objectives. These goals include 1) storage and processing of EEG signals, 2) real-time and online access, 3) providing benchmarks based on pipeline components and aggregated data, and 4) specialized social network. The implementation of this platform has provided the possibility of using the applications of EEG signals and the details of the analysis process in addition to researchers and people interested in this field. Developing and sharing processing pipeline components between users in this system is possible. The proposed platform allows cloud resources to store and eva‎luate the developed artificial intelligence models online and streaming. Considering the sensitivity of EEG signals to the individual differences of humans, to generalize artificial intelligence models in the face of this problem, and to solve the lack of data in the discussion of training models, the proposed platform systematically benchmarks models automatically and periodically according to aggregated data. Also, to facilitate communication and ongoing collaboration among researchers in this field, we have developed a specialized integrated social network that allows for sharing the results of the processing conducted on the platform.

زينب زالي

فرزانه شايق بروجني

مسعودرضا هاشمي , امير اخوان بي تقصير