Gunnar Waterstraat

Are high-frequency (600 Hz) oscillations in human somatosensory evoked potentials due to phase-resetting phenomena?

OBJECTIVE Median nerve somatosensory evoked potentials (SEP) contain a brief oscillatory wavelet burst at about 600 Hz (σ-burst) superimposed on the initial cortical component (N20). While invasive single-cell recordings suggested that this burst is generated by increased neuronal spiking activity in area 3b, recent non-invasive scalp recordings could not reveal concomitant single-trial added-activity, suggesting that the SEP burst might instead be generated by phase-reset of ongoing high-frequency EEG. Here, a statistical model and exemplary data are presented reconciling these seemingly contradictory results. METHODS A statistical model defined the conditions required to detect added-activity in a set of single-trial SEP. Its predictions were tested by analyzing human single-trial scalp SEP recorded with custom-made low-noise amplifiers. RESULTS The noise level in previous studies did not allow to detect single-trial added-activity in the period concomitant with the trial-averaged σ-burst. In contrast, optimized low-noise recordings do reveal added-activity in a set of single-trials. CONCLUSIONS The experimental noise level is the decisive factor determining the detectability of added-activity in single-trials. A low-noise experiment provided direct evidence that the SEP σ-burst is at least partly generated by added-activity matching earlier invasive single-cell recordings. SIGNIFICANCE Quantitative criteria are provided for the feasibility of single-trial detectability of band-limited added-activity.

Non-invasive single-trial EEG detection of evoked human neocortical population spikes

QUESTION Human high-frequency (>400 Hz) components of somatosensory evoked potentials (hf-SEPs), which can be recorded non-invasively at the scalp, are generated by cortical population spikes, as inferred from microelectrode recordings in non-human primates. It is a critical limitation to broader neurophysiological study of hf-SEPs in that hundreds of responses have to be averaged to detect hf-SEPs reliably. Here, we establish a framework for detecting human hf-SEPs non-invasively in single trials. METHODS Spatio-temporal features were extracted from band-pass filtered (400-900 Hz) hf-SEPs by bilinear Common Spatio-Temporal Patterns (bCSTP) and then classified by a weighted Extreme Learning Machine (w-ELM). The effect of varying signal-to-noise ratio (SNR), number of trials, and degree of w-ELM re-weighting was characterized using surrogate data. For practical demonstration of the algorithm, median nerve hf-SEPs were recorded inside a shielded room in four subjects, spanning the hf-SEP signal-to-noise ratio characteristic for a larger population, utilizing a custom-built 29-channel low-noise EEG amplifier. RESULTS Using surrogate data, the SNR proved to be pivotal to detect hf-SEPs in single trials efficiently, with the trade-off between sensitivity and specificity of the algorithm being obtained by the w-ELM re-weighting parameter. In practice, human hf-SEPs were detected non-invasively in single trials with a sensitivity of up to 99% and a specificity of up to 97% in two subjects, even without any recourse to knowledge of stimulus timing. Matching with the results of the surrogate data analysis, these rates dropped to 62-79% sensitivity and 18-31% specificity in two subjects with lower SNR. CONCLUSIONS Otherwise buried in background noise, human high-frequency EEG components can be extracted from low-noise recordings. Specifically, refined supervised filter optimization and classification enables the reliable detection of single-trial hf-SEPs, representing non-invasive correlates of cortical population spikes. SIGNIFICANCE While low-frequency EEG reflects summed postsynaptic potentials, and thereby neuronal input, we suggest that high-frequency EEG (>400 Hz) can provide non-invasive access to the unaveraged output of neuronal computation, i.e., single-trial population spike activity evoked in the responsive neuronal ensemble.

Cortical somatosensory evoked high-frequency (600 Hz) oscillations predict absence of severe hypoxic encephalopathy after resuscitation

OBJECTIVE Following cardiac arrest (CA), hypoxic encephalopathy (HE) frequently occurs and hence reliable neuroprognostication is crucial to decide on the extent of intensive care. Several investigations predict severe HE leading to persistent unresponsive wakefulness or death, with high specificity. Only few studies attempted to predict absence of severe HE. Cortical somatosensory evoked high-frequency (600Hz) oscillation (HFO) bursts indicate the presence of highly synchronized spiking activity in the primary somatosensory cortex. Since global neuronal damage characterizes severe HE preserved cortical HFOs may early exclude severe HE. METHODS We determined amplitudes of early and late HFO bursts in 302 comatose CA patients after median nerve somatosensory evoked potential (SSEPs) and clinical outcome upon intensive care unit discharge using the cerebral performance category (CPC) scale. RESULTS We detected significant early HFO bursts in 146 patients and late HFO bursts in 95 patients. Only one of 27 unresponsive wakefulness patients had a late HFO burst amplitude above 70nV and all seventeen patients who died despite higher amplitudes died from non-neurological causes. CONCLUSIONS High-frequency SSEP components can reliably be studied in comatose CA patients using standard equipment. SIGNIFICANCE Late HFO burst amplitudes above 70nV largely exclude severe HE incompatible with regaining consciousness.

Distinction between added-energy and phase-resetting mechanisms in non-invasively detected somatosensory evoked responses

Non-invasively recorded averaged event-related potentials (ERP) represent a convenient opportunity to investigate human brain perceptive and cognitive processes. Nevertheless, generative ERP mechanisms are still debated. Two previous approaches have been contested in the past: the added-energy model in which the response raises independently from the ongoing background activity, and the phase-reset model, based on stimulus-driven synchronization of oscillatory ongoing activity. Many criteria for the distinction of these two models have been proposed, but there is no definitive methodology to disentangle them, owing also to the limited information at the single trial level. Here, we propose a new approach combining low-noise EEG technology and multivariate decomposition techniques. We present theoretical analyses based on simulated data and identify in high-frequency somatosensory evoked responses an optimal target for the distinction between the two mechanisms.

On optimal spatial filtering for the detection of phase coupling in multivariate neural recordings

Introduction: Neuronal oscillations synchronize processing in the brain over large spatiotemporal scales and thereby facilitate integration of individual functional modules. Up to now, the relation between the phases of neuronal oscillations and behavior or perception has mainly been analyzed in sensor space of multivariate EEG/MEG recordings. However, sensor‐space analysis distorts the topographies of the underlying neuronal sources and suffers from low signal‐to‐noise ratio. Instead, we propose an optimized source reconstruction approach (Phase Coupling Optimization, PCO). Methods: PCO maximizes the ‘mean vector length’, calculated from the phases of recovered neuronal sources and a target variable of interest (e.g., experimental performance). As pre‐processing, the signal‐to‐noise ratio in the search‐space is maximized by spatio‐spectral decomposition. PCO was benchmarked against several competing algorithms and sensor‐space analysis using realistic forward model simulations. As a practical example, thirteen 96‐channel EEG measurements during a simple reaction time task were analyzed. After time‐frequency decomposition, PCO was applied to the EEG to examine the relation between the phases of pre‐stimulus EEG activity and reaction times. Results: In simulations, PCO outperformed other spatial optimization approaches and sensor‐space analysis. Scalp topographies of the underlying source patterns and the relation between the phases of the source activity and the target variable could be reconstructed accurately even for very low SNRs (−10 dB). In a simple reaction time experiment, the phases of pre‐stimulus delta waves (<0.1 Hz) with widely distributed fronto‐parietal source topographies were found predictive of the reaction times. Discussion and conclusions: From multivariate recordings, PCO can reconstruct neuronal sources that are phase‐coupled to a target variable using a data‐driven optimization approach. Its superiority has been shown in simulations and in the analysis of a simple reaction time experiment. From this data, we hypothesize that the phase entrainment of slow delta waves (<1 Hz) facilitates sensorimotor integration in the brain and that this mechanism underlies the faster processing of anticipated stimuli. We further propose that the examined slow delta waves, observed to be phase‐coupled to reaction times, correspond to the compound potentials typically observed in paradigms of stimulus anticipation and motor preparation.

P69. High-frequency (>500 Hz) recordings from deep brain macroelectrodes reveal local interactions of subcortical multi-unit activity

Introduction While low-frequency local field potentials (LFPs) are generated mainly by postsynaptic potentials, high-frequency (hf-) LFP-activity (>500 Hz) is thought to be a correlate of multi-unit spiking activity. Here, we explore hf-recordings from electrodes implanted for deep brain stimulation for the presence of local interactions of multi-unit activity. Methods LFPs were recorded postoperatively after implantation of deep brain stimulation electrodes (each featuring 4 linearly arranged contacts) in 2 patients (#1: globus pallidus internus for multisegmental dystonia; #2: subthalamic nucleus for Parkinson’s disease) using a dedicated low-noise signal amplifier at a sampling rate of 10 kHz. Imaginary part of coherency was calculated between electrodes as a measure of true brain interactions without interference from volume conduction effects. To rule out competing technical and biological mechanisms, the analysis was repeated on recordings from a dummy head and, resp., on surface-EMG recordings. Results The power spectral density of the recordings significantly deviated from a 1/f-characteristic exhibiting a plateau at 150–2.000 Hz. Moreover, at frequencies >500 Hz, imaginary part of coherency indicated the presence of ‘true’ interactions between neighboring pairs of bipolar chain derivations. These signal characteristics were neither present in the dummy-head recordings nor in the surface-EMG recordings. Discussion and significance LFP recordings from two different subcortical sites and in two different diseases indicated the detectability of locally interacting multi-unit activity using macroelectrodes. Major competing mechanisms could be ruled out in control measurements. Further recordings will address the general consistency of these findings and explore potential correlations with clinical data. Acknowledgement This work was supported by KFO 247 (DFG grant Ni 1308/1-2 ).

P20. Disentangling the effect of pre-stimulus oscillatory power on visual detection using spatio-spectral decomposition and heterogeneous choice models

Introduction Analyzing the influence of pre-stimulus EEG/MEG-oscillations on stimulus detection is typically performed by calculating the oscillatory power in each individual EEG/MEG sensor and a single frequency band and correlating it to the parameters of standard Signal Detection Theory (SDT). In the visual domain, this classical approach revealed an effect of widely distributed alpha activity on the perception bias (i.e., how conservative or liberal subjects decide) but not on the overall sensitivity (i.e., the ability to discriminate between stimulus presence and absence). However, sensor-space EEG/MEG data is spatially non-specific and highly susceptible to far-field activity and noise sources. Moreover, classical equal-variances SDT analysis might lead to spurious results. Here, we extract sources of alpha and theta activity using spatio-spectral decomposition (SSD) and model their collective effect on visual detection effectivity using heterogeneous choice models. Methods 64-multichannel EEG of 30 subjects was recorded during a visual yes/no detection task with 60% stimulus present trials and 40% stimulus absent trials. The stimuli consisted of low-contrast Gabor patches displayed at 10 degrees of visual angle in the left or right hemifield. The two strongest alpha and theta sources common for all subjects were identified by SSD applied at group-level. The rank-normalized alpha and theta powers at stimulus presentation (t = 0 ms) were used as regressors in a heterogeneous choice model. Results The heterogeneous choice model replicated the finding that oscillatory alpha power influences the perceptual bias. Additionally, strong occipital alpha power was found to increase the variability in the detection model. Furthermore, the power of a bipolar theta source, tentatively attributable to a deeper origin, was found to be negatively correlated to the overall sensitivity of the visual detection task. Discussion and significance SSD effectively extracted collectively strongest sources of alpha and theta activity across subjects, hereby increasing the spatial specificity of the data and the signal-to-noise ratio. The heterogeneous choice model revealed that alpha oscillations not only influence perceptual bias but also the variability in the model and that strong regional theta oscillations are correlated with less effective visual detection, potentially due to drowsiness. However, further studies are needed to characterize the neurophysiological underpinnings of these observations.

O14: Non-invasive single-trial EEG detection of human neocortical population spikes

G. Waterstraat1,2, M. Burghoff2,3, T. Fedele1,2,3, H.J. Scheer3, G. Curio1,2,4 1Charite-University Medicine Berlin, Neurophysics Group, Department of Neurology, Campus Benjamin Franklin, Berlin, Germany; 2Bernstein Focus: Neurotechnology Berlin, Berlin, Germany; 3Physikalisch-Technische Bundesanstalt, Institute Berlin, Berlin, Germany; 4Bernstein Center for Computational Neuroscience Berlin, Berlin, Germany