No abstract

No abstract

The accurate detection of motor imagery (MI) from electroencephalography (EEG) is a fundamental, as well as challenging, task to provide reliable control of robotic devices to support people suffering from neuro-motor impairments, e.g., in brain-computer interface (BCI) applications. Recently, deep learning approaches have been able to extract subjectindependent features from EEG, to cope with its poor SNR and high intra-subject and cross-subject variability. In this paper, we first present a review of the most recent studies using deep learning for MI classification, with particular attention to their crosssubject performance. Second, we propose DynamicNet, a Pythonbased tool for quick and flexible implementations of deep learning models based on convolutional neural networks. We showcase the potentiality of DynamicNet by implementing EEGNet, a wellestablished architecture for effective EEG classification. Finally, we compare its performance with the filter bank common spatial pattern (FBCSP) in a 4-class MI task (data from a public dataset). To infer cross-subject classification performance, we applied three different cross-validation schemes. From our results, we show that EEGNet implemented with DynamicNet outperforms FBCSP by about 25%, with a statistically significant difference when cross-subject validation schemes are applied. We conclude that deep learning approaches might be particularly helpful to provide higher cross-subject classification performance in multiclass MI classification scenarios. In the future, it is expected to improve DynamicNet to implement new architectures to further investigate cross-subject classification of MI tasks in real-world scenarios.

We are facing a transition towards interconnection of computing systems, people, and things, where boundaries are disappearing and new challenges are emerging. This trend also applies to smart living environments, which are becoming a cyberphysical ecosystem of devices and individuals. Generally, meta-descriptors such as age of information are exploited to obtain efficient content representation and semantic characterization, with the advantage of better data handling. However, the strong relevance of living support in the involved applications imposes to rethink of this approach whenever it is important to factor the human-in-the-loop. In this paper, we discuss how the investigations related to age of information, in particular aimed at statistical descriptions and/or network operation modeling, can be influenced in such scenarios, for what concerns overarching machine learning for data classification and its impact on the sensing frequency, as well as the presence of data correlation that allows for a parsimonious handling of the updates.