Objective . Despite the vast research aimed at improving the performance of steady-state visually evoked potential (SSVEP)-based brain–computer interfaces (BCIs), several limitations exist that restrict the use of such applications for long-term users in the real-world. One of the main challenges has been to reduce training time while maintaining good BCI performance. In view of this challenge, this survey identifies and compares the different training requirements of feature extraction methods for SSVEP-based BCIs. Approach . This paper reviews the various state-of-the-art SSVEP feature extraction methods that have been developed and are most widely used in the literature. Main results . The main contributions compared to existing reviews are the following: (i) a detailed summary, including a brief mathematical description of each feature extraction algorithm, providing a guide to the basic concepts of the state-of-the-art techniques for SSVEP-based BCIs found in literature; (ii) a categorisation of the training requirements of SSVEP-based methods into three categories, defined as training-free methods, subject-specific and subject-independent training methods; (iii) a comparative review of the training requirements of SSVEP feature extraction methods, providing a reference for future work on SSVEP-based BCIs. Significance . This review highlights the strengths and weaknesses of the three categories of SSVEP training methods. Training-free systems are more practical but their performance is limited due to inter-subject variability resulting from the complex EEG activity. Feature extraction methods that incorporate some training data address this issue and in fact have outperformed training-free methods: subject-specific BCIs are tuned to the individual yielding the best performance at the cost of long, tiring training sessions making these methods unsuitable for everyday use; subject-independent BCIs that make use of training data from various subjects offer a good trade-off between training effort and performance, making these BCIs better suited for practical use.

When a recurrent neural network (RNN) language model is used for caption generation, the image information can be fed to the neural network either by directly incorporating it in the RNN – conditioning the language model by ‘injecting’ image features – or in a layer following the RNN – conditioning the language model by ‘merging’ image features. While both options are attested in the literature, there is as yet no systematic comparison between the two. In this paper, we empirically show that it is not especially detrimental to performance whether one architecture is used or another. The merge architecture does have practical advantages, as conditioning by merging allows the RNN’s hidden state vector to shrink in size by up to four times. Our results suggest that the visual and linguistic modalities for caption generation need not be jointly encoded by the RNN as that yields large, memory-intensive models with few tangible advantages in performance; rather, the multimodal integration should be delayed to a subsequent stage.

Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) provide a novel approach for controlling external devices. BCI technologies can be important enabling technologies for people with severe mobility impairment. Endogenous paradigms, which depend on user-generated commands and do not need external stimuli, can provide intuitive control of external devices. This paper discusses BCIs to control various physical devices such as exoskeletons, wheelchairs, mobile robots, and robotic arms. These technologies must be able to navigate complex environments or execute fine motor movements. Brain control of these devices presents an intricate research problem that merges signal processing and classification techniques with control theory. In particular, obtaining strong classification performance for endogenous BCIs is challenging, and EEG decoder output signals can be unstable. These issues present myriad research questions that are discussed in this review paper. This review covers papers published until the end of 2021 that presented BCI-controlled dynamic devices. It discusses the devices controlled, EEG paradigms, shared control, stabilization of the EEG signal, traditional machine learning and deep learning techniques, and user experience. The paper concludes with a discussion of open questions and avenues for future work.

Objective.Quantitative neurophysiological signal parameters are of value in predicting motor recovery after stroke. The novel role of EEG-derived brain symmetry index for motor function prognostication in the subacute phase after stroke is explored.Methods. Ten male stroke patients and ten matched healthy controls were recruited. Motor function was first assessed clinically using the MRC score, its derivative Motricity Index, and the Fugl–Meyer assessment score. EEG was subsequently recorded first with subjects at rest and then during hand grasping motions, triggered by visual cues. Brain symmetry index (BSI) was used to identify the differences in EEG-quantified interhemispheric cortical power asymmetry observable in healthy versus cortical and subcortical stroke patients. Subsequently, any correlation between BSI and motor function was explored.Results. BSI was found to be significantly higher in stroke subjects compared to healthy controls (p=0.023). The difference in BSI was more pronounced in the cortical stroke subgroup (p=0.016). BSI showed only a mild general decrease on repeated monthly recording. Notably, a statistically significant correlation was observed between early BSI and Fugl–Meyer score later in recovery (p<0.050).Conclusions. Brain symmetry index is increased in the subacute poststroke phase and correlates with motor function 1-2 months after stroke.

One of the most important stages in a brain-computer interface (BCI) system is that of extracting features that can reliably discriminate data recorded during different user states. A popular technique used for feature extraction in BCIs is the common spatial patterns (CSP) method, which provides a set of spatial filters that optimally discriminate between two classes of data in the least-squares sense. The method also yields a set of spatial patterns that are associated with the most relevant activity for distinguishing between the two classes. The high recognition rates that have been achieved with the method have led to its widespread adoption in the field. Here, a variant of the CSP method that considers EEG data in its complex form is described. By explicitly considering the amplitude and phase information in the data, the analytic CSP (ACSP) technique can provide a more comprehensive picture of the underlying activity, resulting in improved classification accuracies and more informative spatial patterns than the conventional CSP method. In this paper, we elaborate on the theoretical aspects of the ACSP algorithm and demonstrate the advantages of the method through a number of simulations and through tests on EEG data.

Aim. To evaluate the potential of thermography as an assessment tool for the detection of foot complications by understanding the variations in temperature that occur in type 2 diabetes mellitus (DM). Methods. Participants were categorized according to a medical examination, ankle brachial index, doppler waveform analysis, and 10-gram monofilament testing into five groups: healthy adult, DM with no complications, DM with peripheral neuropathy, DM with neuroischaemia, and DM with peripheral arterial disease (PAD) groups. Thermographic imaging of the toes and forefeet was performed. Results. 43 neuroischaemic feet, 41 neuropathic feet, 58 PAD feet, 21 DM feet without complications, and 126 healthy feet were analyzed. The temperatures of the feet and toes were significantly higher in the complications group when compared to the healthy adult and DM healthy groups. The higher the temperatures of the foot in DM, the higher the probability that it is affected by neuropathy, neuroischaemia, or PAD. Conclusions. Significant differences in mean temperatures exist between participants who were healthy and those with DM with no known complications when compared to participants with neuroischaemia, neuropathy, or PAD. As foot temperature rises, so does the probability of the presence of complications of neuropathy, neuroischaemia, or peripheral arterial disease.