Guang Ouyang

Residue iteration decomposition (RIDE): A new method to separate ERP components on the basis of latency variability in single trials

Event-related brain potentials (ERPs) are important research tools because they provide insights into mental processing at high temporal resolution. Their usefulness, however, is limited by the need to average over a large number of trials, sacrificing information about the trial-by-trial variability of latencies or amplitudes of specific ERP components. Here we propose a novel method based on an iteration strategy of the residues of averaged ERPs (RIDE) to separate latency-variable component clusters. The separated component clusters can then serve as templates to estimate latencies in single trials with high precision. By applying RIDE to data from a face-priming experiment, we separate priming effects and show that they are robust against latency shifts and within-condition variability. RIDE is useful for a variety of data sets that show different degrees of variability and temporal overlap between ERP components.

A toolbox for residue iteration decomposition (RIDE): A method for the decomposition, reconstruction, and single trial analysis of event related potentials

BACKGROUND Conventionally, event-related brain potentials (ERPs) are obtained by averaging a number of single trials. This can be problematic due to trial-to-trial latency variability. Residue iteration decomposition (RIDE) was developed to decompose ERPs into component clusters with different latency variability and to re-synchronize the separated components into a reconstructed ERP. NEW METHOD RIDE has been continuously upgraded and now converges to a robust version. We describe the principles of RIDE and detailed algorithms of the functional modules of a toolbox. We give recommendations and provide caveats for using RIDE from both methodological and psychological perspectives. RESULTS RIDE was applied to several data samples to demonstrate its ability to decompose and reconstruct latency-variable components of ERPs and to retrieve single trial variability information. Different functionalities of RIDE were shown in appropriate examples. COMPARISON WITH EXISTING METHODS RIDE employs several modules to achieve a robust decomposition of ERP. As main innovations RIDE (1) is able to extract components based on the combination of known event markers and estimated latencies, (2) prevents distortions much more effectively than previous methods based on least-square algorithms, and (3) allows time window confinements to target relevant components associated with sub-processes of interest. CONCLUSIONS RIDE is a convenient method that decomposes ERPs and provides single trial analysis, yielding rich information about sub-components, and that reconstructs ERPs, more closely reflecting the combined activity of single trial ERPs. The outcomes of RIDE provide new dimensions to study brain-behavior relationships based on EEG data.

Updating and validating a new framework for restoring and analyzing latency-variable ERP components from single trials with residue iteration decomposition (RIDE)

Trial-to-trial latency variability pervades cognitive EEG responses and may mix and smear ERP components but is usually ignored in conventional ERP averaging. Existing attempts to decompose temporally overlapping and latency-variable ERP components show major limitations. Here, we propose a theoretical framework and model of ERPs consisting of temporally overlapping components locked to different external events or varying in latency from trial to trial. Based on this model, a new ERP decomposition and reconstruction method was developed: residue iteration decomposition (RIDE). Here, we describe an update of the method and compare it to other decomposition methods in simulated and real datasets. The updated RIDE method solves the divergence problem inherent to previous latency-based decomposition methods. By implementing the model of ERPs as consisting of time-variable and invariable single-trial component clusters, RIDE obtains latency-corrected ERP waveforms and topographies of the components, and yields dynamic information about single trials.

Overcoming limitations of the ERP method with Residue Iteration Decomposition (RIDE): A demonstration in go/no-go experiments

The usefulness of the event-related potential (ERP) method can be compromised by violations of the underlying assumptions, for example, confounding variations of latency and amplitude of ERP components within and between conditions. Here we show how the ERP subtraction method might yield misleading information due to latency variability of ERP components. We propose a solution to this problem by correcting for latency variability using Residue Iteration Decomposition (RIDE), demonstrated with data from representative go/no-go experiments. The overlap of N2 and P3 components in go/no-go data gives rise to spurious topographical localization of the no-go-N2 component. RIDE decomposes N2 and P3 based on their latency variability. The decomposition restored the N2 topography by removing the contamination from latency-variable late components. The RIDE-derived N2 and P3 give a clearer insight about their functional relevance in the go/no-go paradigm.

Separating stimulus-driven and response-related LRP components with Residue Iteration Decomposition (RIDE)

When the lateralized readiness potential (LRP) is recorded in stimulus-response compatibility (SRC) tasks, two processes may overlap in the LRP, stimulus-driven response priming and activation based on response selection rules. These overlapping processes are hard to disentangle with standard analytical tools. Here, we show that Residue Iteration Decomposition (RIDE), based on latency variability, separates the overlapping LRP components from a Simon task into stimulus-driven and response-related components. SRC affected LRP amplitudes only in the stimulus-driven component, whereas LRP onsets were affected only in the response-locked component. Importantly, the compatibility effect in reaction times was more similar to the effect in the onsets of the RIDE-derived response-locked LRP component than in the unseparated LRP. Thus, RIDE-separated LRP components are devoid of distortions inherent to standard LRPs.

Exploiting the intra-subject latency variability from single-trial event-related potentials in the P3 time range: A review and comparative evaluation of methods

HIGHLIGHTSIntra‐subject variability (ISV) is reflected by single trial event‐related potentials (ERPs).We reviewed and compared eight algorithms for measuring ISV on ERP.The methods were applied on both simulated and empirical data.The relations between ISV measured from ERP and ISV from behavior were examined.The latency‐invariant ERP component cluster biases the measurement of ISV. ABSTRACT The intra‐subject variability (ISV) in brain responses during cognitive processing across experimental trials has been recognised as an important facet of neural functionality reflecting an intrinsic neurophysiological characteristic of the brain. In recent decades, ISV in behaviour has been found to be significantly associated with cognitive functioning varying across individuals, development, ages, and pathological conditions. Event‐related potentials (ERPs) measured in single trials are important tools for characterizing ISV at the neural level. However, due to the overlapping spectra of noise and signals, the retrieval of information from single‐trial ERPs related to cognitive processing has been a challenge. We review the major problems that researchers face in the estimation of ISV in single‐trial ERPs. Then, we present an extensive evaluation of several methods of single‐trial latency estimation based on both simulated and real data. The relationships of ISV in ERPs and reaction times are compared between the different single‐trial methods to assess their relative efficiency in predicting task performance from neural signals. The pros and cons of the methods are discussed.

Re-examination of Chinese Semantic Processing and Syntactic Processing: Evidence from Conventional ERPs and Reconstructed ERPs by Residue Iteration Decomposition (RIDE)

A number of studies have explored the time course of Chinese semantic and syntactic processing. However, whether syntactic processing occurs earlier than semantics during Chinese sentence reading is still under debate. To further explore this issue, an event-related potentials (ERPs) experiment was conducted on 21 native Chinese speakers who read individually-presented Chinese simple sentences (NP1+VP+NP2) word-by-word for comprehension and made semantic plausibility judgments. The transitivity of the verbs was manipulated to form three types of stimuli: congruent sentences (CON), sentences with a semantically violated NP2 following a transitive verb (semantic violation, SEM), and sentences with a semantically violated NP2 following an intransitive verb (combined semantic and syntactic violation, SEM+SYN). The ERPs evoked from the target NP2 were analyzed by using the Residue Iteration Decomposition (RIDE) method to reconstruct the ERP waveform blurred by trial-to-trial variability, as well as by using the conventional ERP method based on stimulus-locked averaging. The conventional ERP analysis showed that, compared with the critical words in CON, those in SEM and SEM+SYN elicited an N400–P600 biphasic pattern. The N400 effects in both violation conditions were of similar size and distribution, but the P600 in SEM+SYN was bigger than that in SEM. Compared with the conventional ERP analysis, RIDE analysis revealed a larger N400 effect and an earlier P600 effect (in the time window of 500–800 ms instead of 570–810ms). Overall, the combination of conventional ERP analysis and the RIDE method for compensating for trial-to-trial variability confirmed the non-significant difference between SEM and SEM+SYN in the earlier N400 time window. Converging with previous findings on other Chinese structures, the current study provides further precise evidence that syntactic processing in Chinese does not occur earlier than semantic processing.

Stochastic Oscillation in Self-Organized Critical States of Small Systems: Sensitive Resting State in Neural Systems.

Self-organized critical states (SOCs) and stochastic oscillations (SOs) are simultaneously observed in neural systems, which appears to be theoretically contradictory since SOCs are characterized by scale-free avalanche sizes but oscillations indicate typical scales. Here, we show that SOs can emerge in SOCs of small size systems due to temporal correlation between large avalanches at the finite-size cutoff, resulting from the accumulation-release process in SOCs. In contrast, the critical branching process without accumulation-release dynamics cannot exhibit oscillations. The reconciliation of SOCs and SOs is demonstrated both in the sandpile model and robustly in biologically plausible neuronal networks. The oscillations can be suppressed if external inputs eliminate the prominent slow accumulation process, providing a potential explanation of the widely studied Berger effect or event-related desynchronization in neural response. The features of neural oscillations and suppression are confirmed during task processing in monkey eye-movement experiments. Our results suggest that finite-size, columnar neural circuits may play an important role in generating neural oscillations around the critical states, potentially enabling functional advantages of both SOCs and oscillations for sensitive response to transient stimuli.

Effects on P3 of spreading targets and response prompts apart

When key-press responses to targets have to be withheld until the presentation of response prompts, target-evoked P3 amplitudes are reduced and so is the P3 difference between rare and frequent targets (the oddball effect on P3). Recently we showed that this even applied when go-signals followed targets by 100ms. Here we aimed at replicating this result with more fine-grained temporal resolution in 100ms steps from 0ms to 500ms, and dissecting the P3 complex to stimulus- and response-related portions by applying residue iteration decomposition (RIDE). Frequent and rare target stimuli (in random series) were followed by go signals (and occasional no-go signals), with block-wise fixed stimulus-onset asynchronies (SOAs) from 0ms to 500ms. Target-evoked P3 amplitudes decreased monotonically across SOAs. Part of this decrease might have been due to an overlapping Contingent Negative Variation (CNV) prior to go signals, increasing across SOAs. When CNV was subtracted out by forming rare-frequent difference waveforms, oddball-P3 was largest at SOA 0, smallest at SOA 500, and equally large at SOAs 100-400. According to RIDE, it was P3s response-related part that was increased at SOA0. These results may be interpreted in terms of the stimulus-response-link reactivation hypothesis of P3.

Dissociating the Influence of Affective Word Content and Cognitive Processing Demands on the Late Positive Potential

The late positive potential (LPP) elicited by affective stimuli in the event-related brain potential (ERP) is often assumed to be a member of the P3 family. The present study addresses the relationship of the LPP to the classic P3b in a published data set, using a non-parametric permutation test for topographical comparisons, and residue iteration decomposition to assess the temporal features of the LPP and the P3b by decomposing the ERP into several component clusters according to their latency variability. The experiment orthogonally manipulated arousal and valence of words, which were either read or judged for lexicality. High-arousing and positive valenced words induced a larger LPP than low-arousing and negative valenced words, respectively, and the LDT elicited a larger P3b than reading. The experimental manipulation of arousal, valence, and task yielded main effects without any interactions on ERP amplitude in the LPP/P3b time range. The arousal and valence effects partially differed from the task effect in scalp topography; in addition, whereas the late positive component elicited by affective stimuli, defined as LPP, was stimulus-locked, the late positive component elicited by task demand, defined as P3b, was mainly latency-variable. Therefore LPP and P3b manifest different subcomponents.