1 January 2013 in Frontiers in Neural Circuits
We designed a novel assisted closed-loop optimization protocol to improve the efficiency of brain computer interfaces (BCI) based on steady state visually evoked potentials (SSVEP). In traditional paradigms, the control over the BCI-performance completely depends on the subjects’ ability to learn from the given feedback cues. By contrast, in the proposed protocol both the subject and the machine share information and control over the BCI goal. Generally, the innovative assistance consists in the delivery of online information together with the online adaptation of BCI stimuli properties. In our case, this adaptive optimization process is realized by (i) a closed-loop search for the best set of SSVEP flicker frequencies and (ii) feedback of actual SSVEP magnitudes to both the subject and the machine. These closed-loop interactions between subject and machine are evaluated in real-time by continuous measurement of their efficiencies, which are used as online criteria to adapt the BCI control parameters. The proposed protocol aims to compensate for variability in possibly unknown subjects’ state and trait dimensions. In a study with N = 18 subjects, we found significant evidence that our protocol outperformed classic SSVEP-BCI control paradigms. Evidence is presented that it takes indeed into account interindividual variabilities: e.g. under the new protocol, baseline resting state EEG measures predict subjects’ BCI performances. This paper illustrates the promising potential of assisted closed-loop protocols in BCI systems. Probably their applicability might be expanded to innovative uses, e.g. as possible new diagnostic/therapeutic tools for clinical contexts and as new paradigms for basic research.
24 October 2017 in IEEE Access
Reconstruction of a hand's motion with electroencephalography (EEG) signals is a challenging problem that has not been solved yet. Most related studies rely on a motion tracking system to record a sequence of hand coordinate values paired with biosignals, in order to train a mapping function between them. For amputees, this approach is not possible. There are also only a few studies about how different training techniques may affect the accuracy of a motion reconstruction system. A virtual avatar for presenting different upper limb motions was developed. Subjects were asked to follow the avatar's motion, while the subject's EEG and electromyography (EMG) signals were recorded and paired with avatar's hand trajectory values. This task was performed under three conditions: repeating the motion by memory, repeating the motion while watching it on a screen, and repeating the motion while seeing it in virtual reality (VR). We did not find any significant difference between the three conditions in terms of correlation values. Still, we found that using both EEG and EMG at the same time led to a better result than using only one of them. Additionally, significant differences were found in the EEG activity, suggesting that even if the task (moving the arm) was the same for the three conditions, the brain dynamics were different. Specifically, we found that using the VR resulted in a higher alpha desynchronization during the motion. Finally, our results, when only EEG signals were used, were comparable with other studies that have used a motion tracking system.
13 May 2021 in Journal of Neural Engineering
Objective. In many real-world decision tasks, the information available to the decision maker is incomplete. To account for this uncertainty, we associate a degree of confidence to every decision, representing the likelihood of that decision being correct. In this study, we analyse electroencephalography (EEG) data from 68 participants undertaking eight different perceptual decision-making experiments. Our goals are to investigate (1) whether subject- and task-independent neural correlates of decision confidence exist, and (2) to what degree it is possible to build brain computer interfaces that can estimate confidence on a trial-by-trial basis. The experiments cover a wide range of perceptual tasks, which allowed to separate the task-related, decision-making features from the task-independent ones. Approach. Our systems train artificial neural networks to predict the confidence in each decision from EEG data and response times. We compare the decoding performance with three training approaches: (1) single subject, where both training and testing data were acquired from the same person; (2) multi-subject, where all the data pertained to the same task, but the training and testing data came from different users; and (3) multi-task, where the training and testing data came from different tasks and subjects. Finally, we validated our multi-task approach using data from two additional experiments, in which confidence was not reported. Main results. We found significant differences in the EEG data for different confidence levels in both stimulus-locked and response-locked epochs. All our approaches were able to predict the confidence between 15% and 35% better than the corresponding reference baselines. Significance. Our results suggest that confidence in perceptual decision making tasks could be reconstructed from neural signals even when using transfer learning approaches. These confidence estimates are based on the decision-making process rather than just the confidence-reporting process.
11 July 2022 in Journal of Neural Engineering
Objective. We investigated whether a recently introduced transfer-learning technique based on meta-learning could improve the performance of brain–computer interfaces (BCIs) for decision-confidence prediction with respect to more traditional machine learning methods. Approach. We adapted the meta-learning by biased regularisation algorithm to the problem of predicting decision confidence from electroencephalography (EEG) and electro-oculogram (EOG) data on a decision-by-decision basis in a difficult target discrimination task based on video feeds. The method exploits previous participants’ data to produce a prediction algorithm that is then quickly tuned to new participants. We compared it with with the traditional single-subject training almost universally adopted in BCIs, a state-of-the-art transfer learning technique called domain adversarial neural networks, a transfer-learning adaptation of a zero-training method we used recently for a similar task, and with a simple baseline algorithm. Main results. The meta-learning approach was significantly better than other approaches in most conditions, and much better in situations where limited data from a new participant are available for training/tuning. Meta-learning by biased regularisation allowed our BCI to seamlessly integrate information from past participants with data from a specific user to produce high-performance predictors. Its robustness in the presence of small training sets is a real-plus in BCI applications, as new users need to train the BCI for a much shorter period. Significance. Due to the variability and noise of EEG/EOG data, BCIs need to be normally trained with data from a specific participant. This work shows that even better performance can be obtained using our version of meta-learning by biased regularisation.
17 August 2016 in Frontiers in Robotics and AI
Predicting a hand’s position using only biosignals is a complex problem that has not been completely solved. The only reliable solutions currently available require invasive surgery. The attempts using non-invasive technologies are rare, and usually have led to lower correlation values between the real and the reconstructed position than those required for real-world applications. In this study, we propose a solution for reconstructing the hand’s position in three dimensions using EEG and EMG to detect from the shoulder area. This approach would be valid for most trans-humeral amputees. In order to find the best solution, we tested four different architectures for the system based on artificial neural networks. Our results show that it is possible to reconstruct the hand’s motion trajectory with a correlation value up to 0.809 compared to a typical value in the literature of 0.6. We also demonstrated that both EEG and EMG contribute jointly to the motion reconstruction. Furthermore, we discovered that the system architectures do not change the results radically. In addition, our results suggest that different motions may have different brain activity patterns that could be detected through EEG. Finally, we suggest a method to study non-linear relations in the brain through the EEG signals, which may lead to a more accurate system.