Search

Objective. Speech decoding, one of the most intriguing brain-computer interface applications, opens up plentiful opportunities from rehabilitation of patients to direct and seamless communication between human species. Typical solutions rely on invasive recordings with a large number of distributed electrodes implanted through craniotomy. Here we explored the possibility of creating speech prosthesis in a minimally invasive setting with a small number of spatially segregated intracranial electrodes. Approach. We collected one hour of data (from two sessions) in two patients implanted with invasive electrodes. We then used only the contacts that pertained to a single stereotactic electroencephalographic (sEEG) shaft or an electrocorticographic (ECoG) stripe to decode neural activity into 26 words and one silence class. We employed a compact convolutional network-based architecture whose spatial and temporal filter weights allow for a physiologically plausible interpretation. Main results. We achieved on average 55% accuracy using only six channels of data recorded with a single minimally invasive sEEG electrode in the first patient and 70% accuracy using only eight channels of data recorded for a single ECoG strip in the second patient in classifying 26+1 overtly pronounced words. Our compact architecture did not require the use of pre-engineered features, learned fast and resulted in a stable, interpretable and physiologically meaningful decision rule successfully operating over a contiguous dataset collected during a different time interval than that used for training. Spatial characteristics of the pivotal neuronal populations corroborate with active and passive speech mapping results and exhibit the inverse space-frequency relationship characteristic of neural activity. Compared to other architectures our compact solution performed on par or better than those recently featured in neural speech decoding literature. Significance. We showcase the possibility of building a speech prosthesis with a small number of electrodes and based on a compact feature engineering free decoder derived from a small amount of training data.

Meditation is a special state of consciousness associated with distinct physiological and brain correlates. Multiple investigations of these correlates reported controversial results and do not lead to full understanding of the underlying processes. However., most studies have proven beneficial effects of regular meditation. Here we investigated the dynamics of multiple neurophysiological parameters: respiration., photoplethysmography., electroencephalography (EEG)., during a controlled staged meditation session. We measured the physiological and brain changes at rest and during guided Taoist meditation in both experienced meditators (N=13) and subjects without any meditation experience (N=15). All subjects followed the same set of audio instructions. We assessed brain activity and autonomic functions at different meditation stages. Indicators of autonomic nervous system activity that differed significantly between the experienced meditators and the novices were considered as indices of interest. A representational similarity analysis was employed to analyze the coupling of these autonomic-function markers with the EEG features. Consistent with previous studies., the experienced meditators demonstrated higher levels of autonomic and central nervous system interaction in the meditative state compared with novice subjects.

The final aim of our research is to develop a braincomputer interface (BCI) for olfaction. Our research program relies on modern olfactory displays and advanced processing of electroencephalography (EEG) and respiratory data in order to develop methods for robust olfactory BCI systems. Here we present the initial results from 17 subjects of our ongoing study. Applying k-nearest neighbors algorithm (k-NN) classification methods we achieve up to 79.9% accuracy for within subject classification of EEG signals for odor pairs. We propose some methods for further improvement of classification algorithm. EEG classification for different olfactory stimuli is the first step in this research, followed by the development of odor-imagery BCIs and odor-based neurofeedback.

Meditation is a consciousness state associated with specific physiological and neural correlates. Numerous investigations of these correlates reported controversial results which prevented a consistent depiction of the underlying neurophysiological processes. Here we investigated the dynamics of multiple neurophysiological indicators during a staged meditation session. We measured the physiological changes at rest and during the guided Taoist meditation in experienced meditators and naive subjects. We recorded EEG, respiration, galvanic skin response, and photoplethysmography. All subjects followed the same instructions split into 16 stages. In the experienced meditators group we identified two subgroups with different physiological markers dynamics. One subgroup showed several signs of general relaxation evident from the changes in heart rate variability, respiratory rate, and EEG rhythmic activity. The other subgroup exhibited mind concentration patterns primarily noticeable in the EEG recordings while no autonomic responses occurred. The duration and type of previous meditation experience or any baseline indicators we measured did not explain the segregation of the meditators into these two groups. These results suggest that two distinct meditation strategies could be used by experienced meditators, which partly explains the inconsistent results reported in the earlier studies evaluating meditation effects. Our findings are also relevant to the development of the high-end biofeedback systems.

Objective. Epilepsy is a widely spread neurological disease, whose treatment often requires resection of the pathological cortical tissue. Interictal spike analysis observed in the non-invasively collected EEG or MEG data offers an attractive way to localize epileptogenic cortical structures for surgery planning purposes. Interictal spike detection in lengthy multichannel data is a daunting task that is still often performed manually. This frequently limits such an analysis to a small portion of the data which renders the appropriate risks of missing the potentially epileptogenic region. While a plethora of automatic spike detection techniques have been developed each with its own assumptions and limitations, none of them is ideal and the best results are achieved when the output of several automatic spike detectors are combined. This is especially true in the low signal-to-noise ratio conditions. To this end we propose a novel biomimetic approach for automatic spike detection based on a constrained mixed spline machinery that we dub as fast parametric curve matching (FPCM). Approach. Using the peak-wave shape parametrization, the constrained parametric morphological model is constructed and convolved with the observed multichannel data to efficiently determine mixed spline parameters corresponding to each time-point in the dataset. Then the logical predicates that directly map to verbalized text-book like descriptions of the expected interictal event morphology allow us to accomplish the spike detection task. Main results. The results of simulations mimicking typical low SNR scenarios show the robustness and high receiver operating characteristic AUC values of the FPCM method as compared to the spike detection performed using more conventional approaches such as wavelet decomposition, template matching or simple amplitude thresholding. Applied to the real MEG and EEG data from the human patients and to rat ECoG data, the FPCM technique demonstrates reliable detection of the interictal events and localization of epileptogenic zones concordant with independent conclusions made by the epileptologist. Significance. Since the FPCM is computationally light, tolerant to high amplitude artifacts and flexible to accommodate verbalized descriptions of an arbitrary target morphology, it is likely to complement the existing arsenal of means for analysis of noisy interictal datasets.

Objective. Brain–computer interfaces (BCIs) decode information from neural activity and send it to external devices. The use of Deep Learning approaches for decoding allows for automatic feature engineering within the specific decoding task. Physiologically plausible interpretation of the network parameters ensures the robustness of the learned decision rules and opens the exciting opportunity for automatic knowledge discovery. Approach. We describe a compact convolutional network-based architecture for adaptive decoding of electrocorticographic (ECoG) data into finger kinematics. We also propose a novel theoretically justified approach to interpreting the spatial and temporal weights in the architectures that combine adaptation in both space and time. The obtained spatial and frequency patterns characterizing the neuronal populations pivotal to the specific decoding task can then be interpreted by fitting appropriate spatial and dynamical models. Main results. We first tested our solution using realistic Monte-Carlo simulations. Then, when applied to the ECoG data from Berlin BCI competition IV dataset, our architecture performed comparably to the competition winners without requiring explicit feature engineering. Using the proposed approach to the network weights interpretation we could unravel the spatial and the spectral patterns of the neuronal processes underlying the successful decoding of finger kinematics from an ECoG dataset. Finally we have also applied the entire pipeline to the analysis of a 32-channel EEG motor-imagery dataset and observed physiologically plausible patterns specific to the task. Significance. We described a compact and interpretable CNN architecture derived from the basic principles and encompassing the knowledge in the field of neural electrophysiology. For the first time in the context of such multibranch architectures with factorized spatial and temporal processing we presented theoretically justified weights interpretation rules. We verified our recipes using simulations and real data and demonstrated that the proposed solution offers a good decoder and a tool for investigating motor control neural mechanisms.

Objective. Feedback latency was shown to be a critical parameter in a range of applications that imply learning. The therapeutic effects of neurofeedback (NFB) remain controversial. We hypothesized that often encountered unreliable results of NFB intervention could be associated with large feedback latency values that are often uncontrolled and may preclude the efficient learning. Approach. We engaged our subjects into a parietal alpha power unpregulating paradigm facilitated by visual NFB based on the individually extracted envelope of the alpha-rhythm at P4 electrode. NFB was displayed either as soon as electroencephalographic (EEG) envelope was processed, or with an extra 250 or 500 ms delay. The feedback training consisted of 15 two-minute long blocks interleaved with 15 s pauses. We have also recorded 2 min long baselines immediately before and after the training. Main results. The time course of NFB-induced changes in the alpha rhythm power clearly depended on NFB latency, as shown with the adaptive Neyman test. NFB had a strong effect on the alpha-spindle incidence rate, but not on their duration or amplitude. The sustained changes in alpha activity measured after the completion of NFB training were negatively correlated to latency, with the maximum change for the shortest tested latency and no change for the longest. Significance. Here we for the first time show that visual NFB of parietal EEG alpha-activity is efficient only when delivered to human subjects at short latency, which guarantees that NFB arrives when an alpha spindle is still ongoing. Such a considerable effect of NFB latency on the alpha-activity temporal structure could explain some of the previous inconsistent results, where latency was neither controlled nor documented. Clinical practitioners and manufacturers of NFB equipment should add latency to their specifications while enabling latency monitoring and supporting short-latency operations.

Closed-loop Neuroscience is based on the experimental approach where the ongoing brain activity is recorded, processed, and passed back to the brain as sensory feedback or direct stimulation of neural circuits. The artificial closed loops constructed with this approach expand the traditional stimulus-response experimentation. As such, closed-loop Neuroscience provides insights on the function of loops existing in the brain and the ways the flow of neural information could be modified to treat neurological conditions.Neural oscillations, or brain rhythms, are a class of neural activities that have been extensively studied and also utilized in brain rhythm-contingent (BRC) paradigms that incorporate closed loops. In these implementations, instantaneous power and phase of neural oscillations form the signal that is fed back to the brain.Here we addressed the problem of feedback delay in BRC paradigms. In many BRC systems, it is critical to keep the delay short. Long delays could render the intended modification of neural activity impossible because the stimulus is delivered after the targeted neural pattern has already completed. Yet, the processing time needed to extract oscillatory components from the broad-band neural signals can significantly exceed the period of oscillations, which puts a demand for algorithms that could minimize the delay.We used EEG data collected in human subjects to systematically investigate the performance of a range of signal processing methods in the context of minimizing delay in BRC systems. We proposed a family of techniques based on the least-squares filter design – a transparent and simple approach, as it required a single parameter to adjust the accuracy versus latency trade-off. Our algorithm performed on par or better than the state-of the art techniques currently used for the estimation of rhythm envelope and phase in closed-loop EEG paradigms.

Objective The rapidly developing paradigm of closed-loop neuroscience has extensively employed brain rhythms as the signal forming real-time neurofeedback, triggering brain stimulation, or governing stimulus selection. However, the efficacy of brain rhythm contingent paradigms suffers from significant delays related to the process of extraction of oscillatory parameters from broad-band neural signals with conventional methods. To this end, real-time algorithms are needed that would shorten the delay while maintaining an acceptable speed-accuracy trade-off. Approach Here we evaluated a family of techniques based on the application of the least-squares complex-valued filter (LSCF) design to real-time quantification of brain rhythms. These techniques allow for explicit optimization of the speed-accuracy trade-off when quantifying oscillatory patterns. We used EEG data collected from 10 human participants to systematically compare LSCF approach to the other commonly used algorithms. Each method being evaluated was optimized by scanning through the grid of its hyperparameters using independent data samples. Main results When applied to the task of estimating oscillatory envelope and phase, the LSCF techniques outperformed in speed and accuracy both conventional Fourier transform and rectification based methods as well as more advanced techniques such as those that exploit autoregressive extrapolation of narrow-band filtered signals. When operating at zero latency, the weighted LSCF approach yielded 75% accuracy when detecting alpha-activity episodes, as defined by the amplitude crossing of the 95th-percentile threshold. Significance The LSCF approaches are easily applicable to low-delay quantification of brain rhythms. As such, these methods are useful in a variety of neurofeedback, brain-computer-interface and other experimental paradigms that require rapid monitoring of brain rhythms.

Neurofeedback (NFB) is a real-time paradigm, where subjects monitor their own brain activity presented to them via one of the sensory modalities: visual, auditory or tactile. NFB has been proposed as an approach to treat neurological conditions and augment brain functions. In many applications, especially in the automatic learning scenario it is important to decrease NFB latency, so that appropriate brain mechanisms can be efficiently engaged. To this end, we propose a novel algorithm that significantly reduces feedback signal presentation in the electroencephalographic (EEG) NFB paradigm. The algorithm is based on the least squares optimization of the finite impulse response (FIR) filter weights and analytic signal reconstruction. In this approach, the trade-off between NFB latency and the accuracy of EEG envelope estimation can be achieved depending on the application needs. Moreover, the algorithm allows to implement predictive NFB by setting latency to negative values while maintaining acceptable envelope estimation accuracy. As such, our algorithm offers significant improvements in cases where subjects need to detect neural events as soon as possible and even in advance.

Increasing evidence suggests that neuronal communication is a defining property of functionally specialized brain networks and that it is implemented through synchronization between population activities of distinct brain areas. The detection of long-range coupling in electroencephalography (EEG) and magnetoencephalography (MEG) data using conventional metrics (such as coherence or phase-locking value) is by definition contaminated by spatial leakage. Methods such as imaginary coherence, phase-lag index or orthogonalized amplitude correlations tackle spatial leakage by ignoring zero-phase interactions. Although useful, these metrics will by construction lead to false negatives in cases where true zero-phase coupling exists in the data and will underestimate interactions with phase lags in the vicinity of zero. Yet, empirically observed neuronal synchrony in invasive recordings indicates that it is not uncommon to find zero or close-to-zero phase lag between the activity profiles of coupled neuronal assemblies. Here, we introduce a novel method that allows us to mitigate the undesired spatial leakage effects and detect zero and near zero phase interactions. To this end, we propose a projection operation that operates on sensor-space cross-spectrum and suppresses the spatial leakage contribution but retains the true zero-phase interaction component. We then solve the network estimation task as a source estimation problem defined in the product space of interacting source topographies. We show how this framework provides reliable interaction detection for all phase-lag values and we thus refer to the method as Phase Shift Invariant Imaging of Coherent Sources (PSIICOS). Realistic simulations demonstrate that PSIICOS has better detector characteristics than existing interaction metrics. Finally, we illustrate the performance of PSIICOS by applying it to real MEG dataset recorded during a standard mental rotation task. Taken together, using analytical derivations, data simulations and real brain data, this study presents a novel source-space MEG/EEG connectivity method that overcomes previous limitations and for the first time allows for the estimation of true zero-phase coupling via non-invasive electrophysiological recordings.

Cortical speech processing is dependent on the mutual interdependence of two distinctive processing streams supporting sound-to-meaning (i.e., ventral stream) and sound-to-articulation (i.e., dorsal stream) mapping. Here, we compared the strengths of intracranial functional connectivity between two main hubs of the dorsal stream, namely the left auditory-related cortex (ARC) and Broca’s region, in a sample of simultaneous interpreters (SIs) and multilingual control subjects while the participants performed a mixed and unmixed auditory semantic decision task. Under normal listening conditions such kind of tasks are known to initiate a spread of activation along the ventral stream. However, due to extensive and specific training, here we predicted that SIs will more strongly recruit the dorsal pathway in order to pre-activate the speech codes of the corresponding translation. In line with this reasoning, EEG results demonstrate increased left-hemispheric theta phase synchronization in SLI compared to multilingual control participants during early task-related processing stages. In addition, within the SI group functional connectivity strength in the left dorsal pathway was positively related to the cumulative number of training hours across lifespan, and inversely correlated with the age of training commencement. Hence, we propose that the alignment of neuronal oscillations between brain regions involved in “hearing” and “speaking” results from an intertwining of training, sensitive period, and predisposition.

In this paper we describe a novel data driven spatial filtering technique that can be applied to the ERP analysis in order to find statistically significant hidden differential activations in the EEG data. The technique is based on the known morphological characteristics of the response. Underlying optimization problem is formulated as a generalized Rayleigh quotient maximization problem. We supply our tech-nique with a relevant randomization-based statistical test to assess the significance of the discovered phenomenon. Furthermore, we describe an application of the proposed method to the EEG data acquired in the study devoted to the analysis of the auditory neuroplasticity. We show how the mismatch negativity component, a tiny and short-lasting negative response that hallmarks the novel stimuli activating primary error-detection mechanisms, can be detected after filtration.

Meditation is increasingly showing beneficial effects for psychiatric disorders. However, learning to meditate is not straightforward as there are no easily discernible outward signs of performance and thus no direct feedback is possible. As meditation has been found to correlate with posterior cingulate cortex (PCC) activity, we tested whether source-space EEG neurofeedback from the PCC followed the subjective experience of effortless awareness (a major component of meditation), and whether participants could volitionally control the signal.Sixteen novice meditators and sixteen experienced meditators participated in the study. Novice meditators were briefly trained to perform a basic meditation practice to induce the subjective experience of effortless awareness in a progressively more challenging neurofeedback test-battery. Experienced meditators performed a self-selected meditation practice to induce this state in the same test-battery. Neurofeedback was provided based on gamma-band (40-57Hz) PCC activity extracted using a beamformer algorithm. Associations between PCC activity and the subjective experience of effortless awareness were assessed by verbal probes.Both groups reported that decreased PCC activity corresponded with effortless awareness (P0.0025 for each group), with high median confidence ratings (novices: 8 on a 0-10 Likert scale; experienced: 9). Both groups showed high moment-to-moment median correspondence ratings between PCC activity and subjective experience of effortless awareness (novices: 8, experienced: 9). Both groups were able to volitionally control the PCC signal in the direction associated with effortless awareness by practicing effortless awareness meditation (novices: median % of time=77.97, P=0.001; experienced: 89.83, P0.0005).These findings support the feasibility of using EEG neurofeedback to link an objective measure of brain activity with the subjective experience of effortless awareness, and suggest potential utility of this paradigm as a tool for meditation training.

Humans often change their beliefs or behavior due to the behavior or opinions of others. This study explored with the use of human event related potentials (ERPs) whether social conformity is based on a general performance monitoring mechanism. We tested the hypothesis that conflicts with a normative group opinion evoke a feedback related negativity (FRN) often associated with performance monitoring and subsequent adjustment of behavior. The experimental results show that individual judgments of facial attractiveness were adjusted in line with a normative group opinion. A mismatch between individual and group opinions triggered a frontocentral negative deflection with the maximum at 200 ms similar to FRN. Overall a conflict with a normative group opinion triggered a cascade of neuronal responses: from an earlier FRN response reflecting a conflict with the normative opinion to a later ERP component (peaking at 380 ms) reflecting a conforming behavioral adjustment. These results add to the growing literature on neuronal mechanisms of social influence by disentangling the conflict monitoring signal in response to the perceived violation of social norms and the neural signal of a conforming behavioral adjustment. © The Author (2012). Published by Oxford University Press.

Connectivity measures are (typically bivariate) statistical measures that may be used to estimate interactions between brain regions from electrophysiological data. We review both formal and informal descriptions of a range of such measures, suitable for the analysis of human brain electrophysiological data, principally electro- and magnetoencephalography. Methods are described in the space-time, space-frequency, and space-time-frequency domains. Signal processing and information theoretic measures are considered, and linear and nonlinear methods are distinguished. A novel set of cross-time-frequency measures is introduced, including a cross-time-frequency phase synchronization measure.

We utilized the event-related potential (ERP) technique to study neural activity associated with different levels of working memory (WM) load during simultaneous interpretation (SI) of continuous prose. The amplitude of N1 and P1 components elicited by task-irrelevant tone probes was significantly modulated as a function of WM load but not the direction of interpretation. Furthermore, the latency of the P1 increased significantly with WM load. The WM load effect on N1 latency, however, did not reach significance. Larger negativity under lower WM loads suggests that more attention is available to process the source message, providing the first electrophysiological evidence in support of the Efforts Model of SI. Relationships between the direction of interpretation and median WM load are also discussed.

The characterization of spatial network dynamics is desirable for a better understanding of seizure physiology. The goal of this work is to develop a computational method for identifying transient spatial patterns from intracranial electroencephalographic (iEEG) data.Starting with bivariate synchrony measures, such as phase correlation, a two-step clustering procedure is used to identify statistically significant spatial network patterns, whose temporal evolution can be inferred. We refer to this as the composite synchrony profile (CSP) method.The CSP method was verified with simulated data and evaluated using ictal and interictal recordings from three patients with intractable epilepsy. Application of the CSP method to these clinical iEEG datasets revealed a set of distinct CSPs with topographies consistent with medial temporal/limbic and superior parietal/medial frontal networks thought to be involved in the seizure generation process.By combining relatively straightforward multivariate signal processing techniques, such as phase synchrony, with clustering and statistical hypothesis testing, the methods we describe may prove useful for network definition and identification.The network patterns we observe using the CSP method cannot be inferred from direct visual inspection of the raw time series data, nor are they apparent in voltage-based topographic map sequences.