Performance monitoring for brain-computer-interface actions

Neurofeedback through the lens of reinforcement learning

Despite decades of experimental and clinical practice, the neuropsychological mechanisms underlying neurofeedback (NF) training remain obscure. NF is a unique form of reinforcement learning (RL) task, during which participants are provided with rewarding feedback regarding desired changes in neural patterns. However, key RL considerations - including choices during practice, prediction errors, credit-assignment problems, or the exploration-exploitation tradeoff - have infrequently been considered in the context of NF. We offer an RL-based framework for NF, describing different internal states, actions, and rewards in common NF protocols, thus fashioning new proposals for characterizing, predicting, and hastening the course of learning. In this way we hope to advance current understanding of neural regulation via NF, and ultimately to promote its effectiveness, personalization, and clinical utility.

An Investigation of Awareness and Metacognition in Neurofeedback with the Amygdala Electrical Fingerprint

Awareness theory posits that individuals connected to a brain-computer interface can learn to estimate and discriminate their brain states. We used the amygdala Electrical Fingerprint (amyg-EFP) - a functional Magnetic Resonance Imaging-inspired Electroencephalogram surrogate of deep brain activation - to investigate whether participants could accurately estimate their own brain activation. Ten participants completed up to 20 neurofeedback runs and estimated their amygdala-EFP activation (depicted as a thermometer) and confidence in this rating during each trial. We analysed data using multilevel models, predicting the real thermometer position with participant rated position and adjusted for activation during the previous trial. Hypotheses on learning regulation and improvement of estimation were not confirmed. However, participant ratings were significantly associated with the amyg-EFP signal. Higher rating accuracy also predicted higher subjective confidence in the rating. This proof-of-concept study introduces an approach to study awareness with fMRI-informed neurofeedback and provides initial evidence for metacognition in neurofeedback.

How using brain-machine interfaces influences the human sense of agency

Brain-machine interfaces (BMI) allows individuals to control an external device by controlling their own brain activity, without requiring bodily or muscle movements. Performing voluntary movements is associated with the experience of agency ("sense of agency") over those movements and their outcomes. When people voluntarily control a BMI, they should likewise experience a sense of agency. However, using a BMI to act presents several differences compared to normal movements. In particular, BMIs lack sensorimotor feedback, afford lower controllability and are associated with increased cognitive fatigue. Here, we explored how these different factors influence the sense of agency across two studies in which participants learned to control a robotic hand through motor imagery decoded online through electroencephalography. We observed that the lack of sensorimotor information when using a BMI did not appear to influence the sense of agency. We further observed that experiencing lower control over the BMI reduced the sense of agency. Finally, we observed that the better participants controlled the BMI, the greater was the appropriation of the robotic hand, as measured by body-ownership and agency scores. Results are discussed based on existing theories on the sense of agency in light of the importance of BMI technology for patients using prosthetic limbs.

Delineating implicit and explicit processes in neurofeedback learning

Neurofeedback allows humans to self-regulate neural activity in specific brain regions and is considered a promising tool for psychiatric interventions. Recently, methods have been developed to use neurofeedback implicitly, prompting a theoretical debate on the role of awareness in neurofeedback learning. We offer a critical review of the role of awareness in neurofeedback learning, with a special focus on recently developed neurofeedback paradigms. We detail differences in instructions and propose a fine-grained categorization of tasks based on the degree of involvement of explicit and implicit processes. Finally, we review the methods used to measure awareness in neurofeedback and propose new candidate measures. We conclude that explicit processes cannot be eschewed in most current implicit tasks that have explicit goals, and suggest ways in which awareness could be better measured in the future. Investigating awareness during learning will help understand the learning mechanisms underlying neurofeedback learning and will help shape future tasks.

Response-Related Signals Increase Confidence But Not Metacognitive Performance

Confidence judgments are a central tool in metacognition research. In a typical task, participants first perform perceptual (first-order) decisions and then rate their confidence in these decisions. The relationship between confidence and first-order accuracy is taken as a measure of metacognitive performance. Confidence is often assumed to stem from decision-monitoring processes alone, but processes that co-occur with the first-order decision may also play a role in confidence formation.

Confidence judgments are a central tool in metacognition research. In a typical task, participants first perform perceptual (first-order) decisions and then rate their confidence in these decisions. The relationship between confidence and first-order accuracy is taken as a measure of metacognitive performance. Confidence is often assumed to stem from decision-monitoring processes alone, but processes that co-occur with the first-order decision may also play a role in confidence formation. In fact, some recent studies have revealed that directly manipulating motor regions in the brain, or the time of first-order decisions relative to second-order decisions, affects confidence judgments. This finding suggests that confidence could be informed by a readout of reaction times in addition to decision-monitoring processes. To test this possibility, we assessed the contribution of response-related signals to confidence and, in particular, to metacognitive performance (i.e., a measure of the adequacy of these confidence judgments). In human volunteers, we measured the effect of making an overt (vs covert) decision, as well as the effect of pairing an action to the stimulus about which the first-order decision is made. Against our expectations, we found no differences in overall confidence or metacognitive performance when first-order responses were covert as opposed to overt. Further, actions paired to visual stimuli presented led to higher confidence ratings, but did not affect metacognitive performance. These results suggest that confidence ratings do not always incorporate motor information.

Cortical and subcortical mechanisms of brain-machine interfaces

Technical advances in the field of Brain-Machine Interfaces (BMIs) enable users to control a variety of external devices such as robotic arms, wheelchairs, virtual entities and communication systems through the decoding of brain signals in real time. Most BMI systems sample activity from restricted brain regions, typically the motor and premotor cortex, with limited spatial resolution. Despite the growing number of applications, the cortical and subcortical systems involved in BMI control are currently unknown at the whole-brain level. Here, we provide a comprehensive and detailed report of the areas active during on-line BMI control. We recorded functional magnetic resonance imaging (fMRI) data while participants controlled an EEG-based BMI inside the scanner. We identified the regions activated during BMI control and how they overlap with those involved in motor imagery (without any BMI control). In addition, we investigated which regions reflect the subjective sense of controlling a BMI, the sense of agency for BMI-actions. Our data revealed an extended cortical-subcortical network involved in operating a motor-imagery BMI. This includes not only sensorimotor regions but also the posterior parietal cortex, the insula and the lateral occipital cortex. Interestingly, the basal ganglia and the anterior cingulate cortex were involved in the subjective sense of controlling the BMI. These results inform basic neuroscience by showing that the mechanisms of BMI control extend beyond sensorimotor cortices. This knowledge may be useful for the development of BMIs that offer a more natural and embodied feeling of control for the user. Hum Brain Mapp 38:2971-2989, 2017. © 2017 Wiley Periodicals, Inc.