Data-driven approach to the estimation of connectivity and time delays in the coupling of interacting neuronal subsystems

Ordinal methods for a characterization of evolving functional brain networks

Ordinal time series analysis is based on the idea to map time series to ordinal patterns, i.e., order relations between the values of a time series and not the values themselves, as introduced in 2002 by C. Bandt and B. Pompe. Despite a resulting loss of information, this approach captures meaningful information about the temporal structure of the underlying system dynamics as well as about properties of interactions between coupled systems. This-together with its conceptual simplicity and robustness against measurement noise-makes ordinal time series analysis well suited to improve characterization of the still poorly understood spatiotemporal dynamics of the human brain. This minireview briefly summarizes the state-of-the-art of uni- and bivariate ordinal time-series-analysis techniques together with applications in the neurosciences. It will highlight current limitations to stimulate further developments, which would be necessary to advance characterization of evolving functional brain networks.

Effect of functional and effective brain connectivity in identifying vowels from articulation imagery procedures

Articulation imagery, a form of mental imagery, refers to the activity of imagining or speaking to oneself mentally without an articulation movement. It is an effective domain of research in speech impaired neural disorders, as speech imagination has high similarity to real voice communication. This work employs electroencephalography (EEG) signals acquired from articulation and articulation imagery in identifying the vowel being imagined during different tasks. EEG signals from chosen electrodes are decomposed using the empirical mode decomposition (EMD) method into a series of intrinsic mode functions. Brain connectivity estimators and entropy measures have been computed to analyze the functional cooperation and causal dependence between different cortical regions as well as the regularity in the signals. Using machine learning techniques such as multiclass support vector machine (MSVM) and random forest (RF), the vowels have been classified. Three different training and testing protocols (Articulation-AR, Articulation imagery-AI and Articulation vs Articulation imagery-AR vs AI) were employed for identifying the vowel being imagined of articulating. An overall classification accuracy of 80% was obtained for articulation imagery protocol which was found to be higher than the other two protocols. Also, MSVM techniques outperformed the RF technique in terms of the classification accuracy. The effect of brain connectivity estimators and machine learning techniques seems to be reliable in identifying the vowel from the subjects' thought and thereby assisting the people with speech impairment.

The Human Organism as an Integrated Interaction Network: Recent Conceptual and Methodological Challenges

The field of Network Physiology aims to advance our understanding of how physiological systems and sub-systems interact to generate a variety of behaviors and distinct physiological states, to optimize the organism's functioning, and to maintain health. Within this framework, which considers the human organism as an integrated network, vertices are associated with organs while edges represent time-varying interactions between vertices. Likewise, vertices may represent networks on smaller spatial scales leading to a complex mixture of interacting homogeneous and inhomogeneous networks of networks. Lacking adequate analytic tools and a theoretical framework to probe interactions within and among diverse physiological systems, current approaches focus on inferring properties of time-varying interactions-namely strength, direction, and functional form-from time-locked recordings of physiological observables. To this end, a variety of bivariate or, in general, multivariate time-series-analysis techniques, which are derived from diverse mathematical and physical concepts, are employed and the resulting time-dependent networks can then be further characterized with methods from network theory. Despite the many promising new developments, there are still problems that evade from a satisfactory solution. Here we address several important challenges that could aid in finding new perspectives and inspire the development of theoretic and analytical concepts to deal with these challenges and in studying the complex interactions between physiological systems.

Time-frequency Techniques in Biomedical Signal Analysis

Objectives: This review outlines the method -ological fundamentals of the most frequently used non-parametric time-frequency analysis techniques in biomedicine and their main properties, as well as providing decision aids concerning their applications.

Methods: The short-term Fourier transform (STFT), the Gabor transform (GT), the S-transform (ST), the continuous Morlet wavelet transform (CMWT), and the Hilbert transform (HT) are introduced as linear transforms by using a unified concept of the time-frequency representation which is based on a standardized analytic signal. The Wigner-Ville dis -tribution (WVD) serves as an example of the ‘quadratic transforms’ class. The combination of WVD and GT with the matching pursuit (MP) decomposition and that of the HT with the empirical mode decomposition (EMD) are explained; these belong to the class of signal-adaptive approaches.

Results: Similarities between linear transforms are demonstrated and differences with regard to the time-frequency resolution and interference (cross) terms are presented in detail. By means of simulated signals the effects of different time-frequency resolutions of the GT, CMWT, and WVD as well as the resolution-related properties of the inter -ference (cross) terms are shown. The method-inherent drawbacks and their consequences for the application of the time-frequency techniques are demonstrated by instantaneous amplitude, frequency and phase measures and related time-frequency representations (spectrogram, scalogram, time-frequency distribution, phase-locking maps) of measured magnetoencephalographic (MEG) signals.

Conclusions: The appropriate selection of a method and its parameter settings will ensure readability of the time-frequency representations and reliability of results. When the time-frequency characteristics of a signal strongly correspond with the time-frequency resolution of the analysis then a method may be considered ‘optimal’. The MP-based signal-adaptive approaches are preferred as these provide an appropriate time-frequency resolution for all frequencies while simultaneously reducing interference (cross) terms.

The Dominant-Subthalamic Nucleus Phenomenon in Bilateral Deep Brain Stimulation for Parkinson's Disease: Evidence from a Gait Analysis Study

Background

It has been suggested that parkinsonian [Parkinson’s disease (PD)] patients might have a “dominant” (DOM) subthalamic nucleus (STN), whose unilateral electrical stimulation [deep brain stimulation (DBS)] could lead to an improvement in PD symptoms similar to bilateral STN-DBS.

Objectives

Since disability in PD patients is often related to gait problems, in this study, we wanted to investigate in a group of patients bilaterally implanted for STN-DBS: (1) if it was possible to identify a subgroup of subjects with a dominant STN; (2) in the case, if the unilateral stimulation of the dominant STN was capable to improve gait abnormalities, as assessed by instrumented multifactorial gait analysis, similarly to what observed with bilateral stimulation.

Methods

We studied 10 PD patients with bilateral STN-DBS. A clinical evaluation and a kinematic, kinetic, and electromyographic (EMG) analysis of overground walking were performed—off medication—in four conditions: without stimulation, with bilateral stimulation, with unilateral right or left STN-DBS. Through a hierarchical agglomerative cluster analysis based on motor Unified Parkinson’s Disease Rating Scale scores, it was possible to separate patients into two groups, based on the presence (six patients, DOM group) or absence (four patients, NDOM group) of a dominant STN.

Results

In the DOM group, both bilateral and unilateral stimulation of the dominant STN significantly increased gait speed, stride length, range of motion of lower limb joints, and peaks of moment and power at the ankle joint; moreover, the EMG activation pattern of distal leg muscles was improved. The unilateral stimulation of the non-dominant STN did not produce any significant effect. In the NDOM group, only bilateral stimulation determined a significant improvement of gait parameters.

Conclusion

In the DOM group, the effect of unilateral stimulation of the dominant STN determined an improvement of gait parameters similar to bilateral stimulation. The pre-surgical identification of these patients, if possible, could allow to reduce the surgical risks and side effects of DBS adopting a unilateral approach.

Pre-processing and transfer entropy measures in motor neurons controlling limb movements

Directed information transfer measures are increasingly being employed in modeling neural system behavior due to their model-free approach, applicability to nonlinear and stochastic signals, and the potential to integrate repetitions of an experiment. Intracellular physiological recordings of graded synaptic potentials provide a number of additional challenges compared to spike signals due to non-stationary behaviour generated through extrinsic processes. We therefore propose a method to overcome this difficulty by using a preprocessing step based on Singular Spectrum Analysis (SSA) to remove nonlinear trends and discontinuities. We apply the method to intracellular recordings of synaptic responses of identified motor neurons evoked by stimulation of a proprioceptor that monitors limb position in leg of the desert locust. We then apply normalized delayed transfer entropy measures to neural responses evoked by displacements of the proprioceptor, the femoral chordotonal organ, that contains sensory neurones that monitor movements about the femoral-tibial joint. We then determine the consistency of responses within an individual recording of an identified motor neuron in a single animal, between repetitions of the same experiment in an identified motor neurons in the same animal and in repetitions of the same experiment from the same identified motor neuron in different animals. We found that delayed transfer entropy measures were consistent for a given identified neuron within and between animals and that they predict neural connectivity for the fast extensor tibiae motor neuron.

Maladaptive neural synchrony in tinnitus: origin and restoration

Tinnitus is the conscious perception of sound heard in the absence of physical sound sources external or internal to the body, reflected in aberrant neural synchrony of spontaneous or resting-state brain activity. Neural synchrony is generated by the nearly simultaneous firing of individual neurons, of the synchronization of membrane-potential changes in local neural groups as reflected in the local field potentials, resulting in the presence of oscillatory brain waves in the EEG. Noise-induced hearing loss, often resulting in tinnitus, causes a reorganization of the tonotopic map in auditory cortex and increased spontaneous firing rates and neural synchrony. Spontaneous brain rhythms rely on neural synchrony. Abnormal neural synchrony in tinnitus appears to be confined to specific frequency bands of brain rhythms. Increases in delta-band activity are generated by deafferented/deprived neuronal networks resulting from hearing loss. Coordinated reset (CR) stimulation was developed in order to specifically counteract such abnormal neuronal synchrony by desynchronization. The goal of acoustic CR neuromodulation is to desynchronize tinnitus-related abnormal delta-band oscillations. CR neuromodulation does not require permanent stimulus delivery in order to achieve long-lasting desynchronization or even a full-blown anti-kindling but may have cumulative effects, i.e., the effect of different CR epochs separated by pauses may accumulate. Unlike other approaches, acoustic CR neuromodulation does not intend to reduce tinnitus-related neuronal activity by employing lateral inhibition. The potential efficacy of acoustic CR modulation was shown in a clinical proof of concept trial, where effects achieved in 12 weeks of treatment delivered 4–6 h/day persisted through a preplanned 4-week therapy pause and showed sustained long-term effects after 10 months of therapy, leading to 75% responders.

Detecting position dependent tremor with the Empirical mode decomposition

Background

Primary bowing tremor (PBT) occurs in violinists in the right bowing-arm and is a highly nonlinear and non-stationary signal. However, Fourier-transform based methods (FFT) make the a priori assumption of linearity and stationarity. We present an interesting case of a violinist with PBT and apply a novel method for nonlinear and non-stationary signals for tremor analysis: the empirical mode decomposition (EMD). We compare the results of FFT and EMD analyses.

Methods

Tremor was measured and quantified in a 50-year-old professional violinist with an accelerometer. Data were analyzed using the EMD, the Hilbert transform, the Hilbert spectrum and the marginal Hilbert spectrum. Findings are compared to the FFT-spectrum and FFT-spectrogram.

Results

We could show that the EMD yields intrinsic mode functions, which represent the tremor and IMFs, which are associated with voluntary movement. The instantaneous frequency and amplitude are obtained. In contrast the low time frequency resolution and the artifacts of voluntary movements are seen in the FFT results.

Conclusions

PBT may present itself as a highly non-stationary and nonlinear phenomenon, which can be accurately analyzed with the EMD, since it gives the instantaneous amplitude and frequency and can identify voluntary from involuntary (tremor) movement.

Electronic supplementary material

The online version of this article (doi:10.1186/s40734-014-0014-z) contains supplementary material, which is available to authorized users.

Intrinsic multi-scale analysis: a multi-variate empirical mode decomposition framework

A novel multi-scale approach for quantifying both inter- and intra-component dependence of a complex system is introduced. This is achieved using empirical mode decomposition (EMD), which, unlike conventional scale-estimation methods, obtains a set of scales reflecting the underlying oscillations at the intrinsic scale level. This enables the data-driven operation of several standard data-association measures (intrinsic correlation, intrinsic sample entropy (SE), intrinsic phase synchrony) and, at the same time, preserves the physical meaning of the analysis. The utility of multi-variate extensions of EMD is highlighted, both in terms of robust scale alignment between system components, a pre-requisite for inter-component measures, and in the estimation of feature relevance. We also illuminate that the properties of EMD scales can be used to decouple amplitude and phase information, a necessary step in order to accurately quantify signal dynamics through correlation and SE analysis which are otherwise not possible. Finally, the proposed multi-scale framework is applied to detect directionality, and higher order features such as coupling and regularity, in both synthetic and biological systems.

Task specific inter-hemispheric coupling in human subthalamic nuclei

Cortical networks and quantitative measures of connectivity are integral to the study of brain function. Despite lack of direct connections between left and right subthalamic nuclei (STN), there are apparent physiological connections. During clinical examination of patients with Parkinson’s disease (PD), this connectivity is exploited to enhance signs of PD, yet our understanding of this connectivity is limited. We hypothesized that movement leads to synchronization of neural oscillations in bilateral STN, and we implemented phase coherence, a measure of phase-locking between cortical sites in a narrow frequency band, to demonstrate this synchronization. We analyzed task specific phase synchronization and causality between left and right STN local field potentials (LFPs) recorded from both hemispheres simultaneously during a cued movement task in four subjects with PD who underwent Deep Brain Stimulation (DBS) surgery. We used a data driven approach to determine inter-hemispheric channel pairs and frequencies with a task specific increase in phase locking.We found significant phase locking between hemispheres in alpha frequency (8–12 Hz) in all subjects concurrent with movement of either hand. In all subjects, phase synchronization increased over baseline upon or prior to hand movement onset and lasted until the motion ceased. Left and right hand movement showed similar patterns. Granger causality (GC) at the phase-locking frequencies between synchronized electrodes revealed a unidirectional causality from right to left STN regardless of which side was moved.Phase synchronization across hemispheres between basal ganglia supports existence of a bilateral network having lateralized regions of specialization for motor processing. Our results suggest this bilateral network is activated by a unilateral motor program. Understanding phase synchronization in natural brain functions is critical to development of future DBS systems that augment goal directed behavioral function.

Interhemispheric functional interactions between the subthalamic nuclei of patients with Parkinson's disease

Parkinson's disease (PD) is characterized by widespread neural interactions in cortico-basal-ganglia networks primarily in beta oscillations (approx. 10-30 Hz), as suggested by previous findings of levodopa-modulated interhemispheric coherence between the bilateral subthalamic nuclei (STN) in local field potential recordings (LFPs). However, due to confounding effects of volume conduction the existence of 'genuine' interhemispheric subcortical coherence remains an open question. To address this issue we utilized the imaginary part of coherency (iCOH) which, in contrast to the standard coherence, is not susceptible to volume conduction. LFPs were recorded from eight patients with PD during wakeful rest before and after levodopa administration. We demonstrated genuine coherence between the bilateral STN in both 10-20 and 21-30 Hz oscillations, as revealed by a non-zero iCOH. Crucially, increased iCOH in 10-20 Hz oscillations positively correlated with the worsening of motor symptoms in the OFF medication condition across patients, which was not the case for standard coherence. Furthermore, across patients iCOH was increased after levodopa administration in 21-30 Hz oscillations. These results suggest a functional distinction between low and high beta oscillations in STN-LFP in line with previous studies. Furthermore, the observed functional coupling between the bilateral STN might contribute to the understanding of bilateral effects of unilateral deep brain stimulation. In conclusion, the present results imply a significant contribution of time-delayed neural interactions to interhemispheric coherence, and the clinical relevance of long-distance neural interactions between bilateral STN for motor symptoms in PD.

Movement induced tremor in musicians and non-musicians reflects adaptive brain plasticity

Evidence exists that motor dexterity is associated with a higher tremor amplitude of physiological tremor. Likewise, lower frequencies are associated with motor control. So far only case reports of a higher amplitude of physiological tremor in musicians exist. Moreover, no study has investigated lower frequencies during a finger movement task in musicians who can be regarded as a model of motor expertise. We developed a model and derived three hypotheses which we investigated in this study: (1) Tremor amplitude is higher in the range of physiological tremor and (2) higher for frequency ranges of dystonic tremor in musicians compared to non-musicians; (3) there is no difference in tremor amplitude at frequencies below 4 Hz. We measured tremor during a finger flexion-extension movement in 19 musicians (age 26.5 ± 8.2 years) and 24 age matched non-musicians (age 26.5 ± 8.7). By using empirical mode decomposition in combination with a Hilbert transform we obtained the instantaneous frequency and amplitude, allowing to compare tremor amplitudes throughout the movement at various frequency ranges. We found a significantly higher tremor amplitude in musicians for physiological tremor and a tendency toward a higher amplitude during most of the movement in the frequency range of 4-8 Hz, which, however, was not significant. No difference was found in the frequency range below 4 Hz for the flexion and for almost the entire extension movement. Our results corroborate findings that the 8-12 Hz oscillatory activity plays a role in motor dexterity. However, our results do not allow for the conclusion that tremor at the frequency range of 4-8 Hz is related to either plasticity induced changes that are beneficial for motor skill development nor to maladaptive changes as, e.g., focal dystonia.

Analysis of dystonic tremor in musicians using empirical mode decomposition

OBJECTIVE

Test the hypotheses that tremor amplitude in musicians with task-specific dystonia is higher at the affected finger (dystonic tremor, DT) or the adjacent finger (tremor associated with dystonia, TAD) than (1) in matched fingers of healthy musicians and non-musicians and (2) within patients in the unaffected and non-adjacent fingers of the affected side within patients.

METHODS

We measured 21 patients, 21 healthy musicians and 24 non-musicians. Participants exerted a flexion-extension movement. Instantaneous frequency and amplitude values were obtained with empirical mode decomposition and a Hilbert-transform, allowing to compare tremor amplitudes throughout the movement at various frequency ranges.

RESULTS

We did not find a significant difference in tremor amplitude between patients and controls for either DT or TAD. Neither differed tremor amplitude in the within-patient comparisons.

CONCLUSION

Both hypotheses were rejected and apparently neither DT nor TAD occur in musician's dystonia of the fingers.

SIGNIFICANCE

This is the first study assessing DT and TAD in musician's dystonia. Our finding suggests that even though MD is an excellent model for malplasticity due to excessive practice, it does not seem to provide a good model for DT. Rather it seems that musician's dystonia may manifest itself either as dystonic cramping without tremor or as task-specific tremor without overt dystonic cramping.

Time-variant partial directed coherence for analysing connectivity: a methodological study

For the past decade, the detection and quantification of interactions within and between physiological networks has become a priority-in-common between the fields of biomedicine and computer science. Prominent examples are the interaction analysis of brain networks and of the cardiovascular-respiratory system. The aim of the study is to show how and to what extent results from time-variant partial directed coherence analysis are influenced by some basic estimator and data parameters. The impacts of the Kalman filter settings, the order of the autoregressive (AR) model, signal-to-noise ratios, filter procedures and volume conduction were investigated. These systematic investigations are based on data derived from simulated connectivity networks and were performed using a Kalman filter approach for the estimation of the time-variant multivariate AR model. Additionally, the influence of electrooculogram artefact rejection on the significance and dynamics of interactions in 29 channel electroencephalography recordings, derived from a photic driving experiment, is demonstrated. For artefact rejection, independent component analysis was used. The study provides rules to correctly apply particular methods that will aid users to achieve more reliable interpretations of the results.

Impact of acoustic coordinated reset neuromodulation on effective connectivity in a neural network of phantom sound

Chronic subjective tinnitus is an auditory phantom phenomenon characterized by abnormal neuronal synchrony in the central auditory system. As recently shown in a proof of concept clinical trial, acoustic coordinated reset (CR) neuromodulation causes a significant relief of tinnitus symptoms combined with a significant decrease of pathological oscillatory activity in a network comprising auditory and non-auditory brain areas. The objective of the present study was to analyze whether CR therapy caused an alteration of the effective connectivity in a tinnitus related network of localized EEG brain sources. To determine which connections matter, in a first step, we considered a larger network of brain sources previously associated with tinnitus. To that network we applied a data-driven approach, combining empirical mode decomposition and partial directed coherence analysis, in patients with bilateral tinnitus before and after 12 weeks of CR therapy as well as in healthy controls. To increase the signal-to-noise ratio, we focused on the good responders, classified by a reliable-change-index (RCI). Prior to CR therapy and compared to the healthy controls, the good responders showed a significantly increased connectivity between the left primary cortex auditory cortex and the posterior cingulate cortex in the gamma and delta bands together with a significantly decreased effective connectivity between the right primary auditory cortex and the dorsolateral prefrontal cortex in the alpha band. Intriguingly, after 12 weeks of CR therapy most of the pathological interactions were gone, so that the connectivity patterns of good responders and healthy controls became statistically indistinguishable. In addition, we used dynamic causal modeling (DCM) to examine the types of interactions which were altered by CR therapy. Our DCM results show that CR therapy specifically counteracted the imbalance of excitation and inhibition. CR significantly weakened the excitatory connection between posterior cingulate cortex and primary auditory cortex and significantly strengthened inhibitory connections between auditory cortices and the dorsolateral prefrontal cortex. The overall impact of CR therapy on the entire tinnitus-related network showed up as a qualitative transformation of its spectral response, in terms of a drastic change of the shape of its averaged transfer function. Based on our findings we hypothesize that CR therapy restores a silence based cognitive auditory comparator function of the posterior cingulate cortex.

Empirical and theoretical aspects of generation and transfer of information in a neuromagnetic source network

Variability in source dynamics across the sources in an activated network may be indicative of how the information is processed within a network. Information-theoretic tools allow one not only to characterize local brain dynamics but also to describe interactions between distributed brain activity. This study follows such a framework and explores the relations between signal variability and asymmetry in mutual interdependencies in a data-driven pipeline of non-linear analysis of neuromagnetic sources reconstructed from human magnetoencephalographic (MEG) data collected as a reaction to a face recognition task. Asymmetry in non-linear interdependencies in the network was analyzed using transfer entropy, which quantifies predictive information transfer between the sources. Variability of the source activity was estimated using multi-scale entropy, quantifying the rate of which information is generated. The empirical results are supported by an analysis of synthetic data based on the dynamics of coupled systems with time delay in coupling. We found that the amount of information transferred from one source to another was correlated with the difference in variability between the dynamics of these two sources, with the directionality of net information transfer depending on the time scale at which the sample entropy was computed. The results based on synthetic data suggest that both time delay and strength of coupling can contribute to the relations between variability of brain signals and information transfer between them. Our findings support the previous attempts to characterize functional organization of the activated brain, based on a combination of non-linear dynamics and temporal features of brain connectivity, such as time delay.

Frontal hyperconnectivity related to discounting and reversal learning in cocaine subjects

BACKGROUND

Functional neuroimaging studies suggest that chronic cocaine use is associated with frontal lobe abnormalities. Functional connectivity (FC) alterations of cocaine-dependent individuals (CD), however, are not yet clear. This is the first study to our knowledge that examines resting FC of anterior cingulate cortex (ACC) in CD. Because ACC is known to integrate inputs from different brain regions to regulate behavior, we hypothesized that CD will have connectivity abnormalities in ACC networks. In addition, we hypothesized that abnormalities would be associated with poor performance in delayed discounting and reversal learning tasks.

METHODS

Resting functional magnetic resonance imaging data were collected to look for FC differences between 27 CD (5 women, age: M = 39.73, SD = 6.14 years) and 24 control subjects (5 women, age: M = 39.76, SD = 7.09 years). Participants were assessed with delayed discounting and reversal learning tasks. With seed-based FC measures, we examined FC in CD and control subjects within five ACC connectivity networks with seeds in subgenual, caudal, dorsal, rostral, and perigenual ACC.

RESULTS

The CD showed increased FC within the perigenual ACC network in left middle frontal gyrus, ACC, and middle temporal gyrus when compared with control subjects. The FC abnormalities were significantly positively correlated with task performance in delayed discounting and reversal learning tasks in CD.

CONCLUSIONS

The present study shows that participants with chronic cocaine-dependency have hyperconnectivity within an ACC network known to be involved in social processing and "mentalizing." In addition, FC abnormalities found in CD were associated with difficulties with delay rewards and slower adaptive learning.

Time-frequency spectral analysis of TMS-evoked EEG oscillations by means of Hilbert-Huang transform

A single pulse of Transcranial Magnetic Stimulation (TMS) generates electroencephalogram (EEG) oscillations that are thought to reflect intrinsic properties of the stimulated cortical area and its fast interactions with other cortical areas. Thus, a tool to decompose TMS-evoked oscillations in the time-frequency domain on a millisecond timescale and on a broadband frequency range may help to understand information transfer across cortical oscillators. Some recent studies have employed algorithms based on the Wavelet Transform (WT) to study TMS-evoked EEG oscillations in healthy and pathological conditions. However, these methods do not allow to describe TMS-evoked EEG oscillations with high resolution in time and frequency domains simultaneously. Here, we first develop an algorithm based on Hilbert-Huang Transform (HHT) to compute statistically significant time-frequency spectra of TMS-evoked EEG oscillations on a single trial basis. Then, we compared the performances of the HHT-based algorithm with the WT-based one by applying both of them to a set of simulated signals. Finally, we applied both algorithms to real TMS-evoked potentials recorded in healthy or schizophrenic subjects. We found that the HHT-based algorithm outperforms the WT-based one in detecting the time onset of TMS-evoked oscillations in the classical EEG bands. These results suggest that the HHT-based algorithm may be used to study the communication between different cortical oscillators on a fine time scale.