The presence (or telepresence) feeling and virtual reality (VR) sickness affect the task execution in teleoperation. Most teleoperation works have assessed these concepts using objective (physiological signals) and subjective (questionnaires) measurements. However, these works did not include social tasks. To the best of our knowledge, there was no physiological approach in teleoperation of social tasks. We measured presence and VR sickness in a simulation of teleoperated social tasks by questionnaires and analyzed the correlation between their scores and multimodal biomarkers. The results showed some different correlations from the findings of non-teleoperation studies. These correlations were between presence and neural biomarkers in the frontal-central and central regions (for the beta and delta bands) and between VR sickness and brain biomarkers in the occipital region (for the alpha and beta bands) and the mean temperature. This work revealed significant correlations to support some biomarkers as predictors of the trend of presence and VR sickness in simulated teleoperated social tasks. These biomarkers might also be valid to predict the trend of telepresence and motion sickness in teleoperated social tasks in a remote environment.

Brain-computer interface (BCI) requires calibration phase to learn user-specific decoder that translates brain signals into desired commands. Calibration phase can be lengthy and exhausting for motor imagery-based (MI) BCI, and can potentially reduce the effectiveness of BCI system. Transfer learning (TL) approaches have been implemented in the field of BCI to tackle this problem. Transfer learning based on domain adaptation reduces domain discrepancy between two domains. Transfer component analysis (TCA) is one domain adaptation technique that maps features of both domains into a new space while simultaneously reducing their domain discrepancy. Recently, Euclidean alignment (EA) is used as transfer learning technique in BCI preprocessing step to align electroencephalography (EEG) trials between subjects. This paper proposed a combination of both EA and TCA (EA-TCA) as a TL approach for inter-subject transfer learning. This paper also proposed TCA method that can reuse existing projection matrix to transform new target data (TCA-W). The efficacy of EA to this method is also observed (EA-TCA-W). Results indicate that EA-TCA and EA-TCA-W outperform classification accuracy of those without EA by 3.32% and 6.50%, respectively. This concludes that EA can improve performance of both conventional TCA and proposed TCA-W.

Motor imagery (MI)-based brain-computer interface (BCI) is a system that allows users to control computer devices by imaging body part movements or MI tasks. In BCI applications, the classification of MI using electroencephalogram (EEG) is challenging because EEG is highly susceptible to noise and artifacts. The latency and length of MI period also vary between subjects and sessions; however, many conventional applications tend to empirically define time windows for feature extraction. This can lead to lower MI-BCI performance. This paper proposes two mutual information-based time window adaptation (MT) algorithms; sliding window MT (SWMT) and genetic algorithm MT (GAMT). Both algorithms used optimized reference signals and mutual information analysis to constantly adjust the time window starting point and length. Reference signals were optimized based on mutual information analysis and performance evaluation. Feature extraction and classification algorithms were finally applied to evaluate SWMT and GAMT performance. The results indicate that SWMT and GAMT were able to improve the conventional approach by increasing the classification accuracy by 6.00% and 6.37%, respectively.

Advanced Brain-Computer Interface (BCI) paradigms aim to solve some problems as BCI illiteracy and unfamiliarity of the subjects to be able to control their elicited motor imagery (MI) successfully, hence improving training time and performance of BCI systems. This work evaluates the effect and performance of an Implicit BCI supported by the Gaze Monitoring (IBCI-GM) paradigm for virtual rehabilitation therapy of patients suffering from partial or total paralysis of their upper limbs; this paradigm also was compared with alternative forms of advanced BCI methods such as Virtual Reality-based BCI (VR-BCI) with a head-mounted display (HMD) and a computer screen (CS). Eight subjects participated in the experiments; four subjects tested the VR-BCI with a CS, and the rest of them tested both BCI advanced methods (IBCI-GM and VR-BCI with an HMD). The subjects were asked to control a virtual arm through MI of flexion and extension movements. The VR-BCI HMD was the approached best method; however, IBCI-GM had significant results and was more practical for users, but it depends on the ability to perform eye movements to be applied by patients. Therefore, these methods should be tested with more subjects to have definitive results.

A fuzzy genetic algorithm to optimize spatial filter selection can improve the performance of P300-based brain computer interfaces (BCI); genetic algorithm searches an optimal configuration supported by a fuzzy inference system, it would reduce the error calculated during a 4 fold crossvalidation. The performance is measured through the accuracy and the bit rate, 4 methods based on fuzzy logic and Bayesian linear discriminant analysis are considered for the performance comparison. This proposed method has obtained significant results for healthy persons and post stroke patients, accuracies above 90% and bit rates greater than 8 bits/min for the most of cases evaluated in a P300-based BCI using the Hoffman approach.

Visual feedback in a brain computer interface (BCI) influences significantly in its performance; but when this BCI will be applied in a rehabilitation therapy for post stroke patients, it will be a determining factor to enhance the neuroplasticity process. Many studies have demonstrated the efficiency of virtual reality (VR) in a BCI, motor cortex increases its activation levels due to be an immersive environment for the subject; specifically, a BCI based on motor imagery (MI) with VR feedback has positive effects in patients. This work proposes to apply a BCI with VR to support an upper limb rehabilitation therapy for post stroke patients, but it has been tested with eighteen healthy subjects, they performed MI tasks of flexion and extension of their arms, then they could see a virtual arm in 3D performing the same requested movement. Comparison of power spectral density (PSD) estimation is done during MI tasks in training and online test sessions, and feedback in online test sessions; and it is possible to observe the brain activity in alpha and beta bands remained during online sessions, topographic map shows activated premotor and motor cortex areas, it is significant evidence for the application of this system to motor disable patients.

In this work, we present the use of Kernel Extreme Learning Machine (Kernel ELM) on electroencephalography EEG brain signals in order to classify the P300 wave during the subject development an oddball paradigm. Also, we propose a selection criteria in order to improve the classification accuracy. In this study, the brain signals of healthy and disabled subjects which suffered a stroke were recorded, analyzed and classified. The results reported that the best classification accuracy and average bitrate were 100% using target by block evaluation and 18.38 bits per minute, respectively . These results are compared to various machine learning algorithms so that our results outperformed them.

This work presents a brain computer interface (BCI) framework for upper limb rehabilitation of post stroke patients, combining BCI and virtual reality (VR) technology; a VR feedback is shown to the participants to achieve a greater activation of certain brain regions involved with the performing of upper limb motor task. This system uses an adaptive neuro-fuzzy inference system (ANFIS) classifier to discriminate between a motor task and rest condition, the first one classifies between extension and rest conditions; and the second one classifies between flexion and rest conditions. In the training stage, eight healthy subjects participated in the sessions, the best accuracies are 99.3% and 88.9%, as a result of cross-validation. Meanwhile, the best accuracy in online test is 89%. The methodology here presented can be straightforwardly employed as a rehabilitation system for brain repair in individuals with neurological diseases or brain injury.

Adaptive neuro fuzzy Inference systems (ANFIS) has been applied in brain computer interfaces (BcI) in different ways such as mapping of P300 or fusing information from EEG channels and it has reached high classification accuracy. This work proposes a combination of ANFIS classifiers by voting for a single-trial detection of a P300 wave in a BCI, using four channels; five healthy subjects and three post-stroke patients have participated in this study, each participant performs 4 BCI sessions, crossvalidation is applied to evaluate the classifier performance. The results of average accuracy were greater than 75% for all subjects, similar results were gotten for healthy subjects and post-stroke patients, but the better classifiers for each subject have achieved accuracies greater than 80%.

This work presents the development of a brain computer interface as an alternative communication channel to be used in Robotics. It encompasses the implementation of an electroencephalograph (EEG), as well as the development of all computational methods and necessary techniques to identify mental activities. The developed brain computer interface (BCI) is applied to activate the movements of a 120lb mobile robot, associating four different mental activities to robot commands. The interface is based on EEG signal analyses, which extract features that can be classified as specific mental activities. First, a signal preprocessing is performed from the EEG data, filtering noise, using a spatial filter to increase the scalp signal resolution, and extracting relevant features. Then, different classifier models are proposed, evaluated and compared. At last, two implementations of the developed classifiers are proposed to improve the rate of successful commands to the mobile robot. In one of the implementations, a 91% average hit rate is obtained, with only 1.25% wrong commands after 400 attempts to control the mobile robot.