Accounting for stimulus and participant effects in event-related potential analyses to increase the replicability of studies

Is auditory awareness graded or dichotomous: Electrophysiological correlates of consciousness at different depths of stimulus processing

The level-of-processing (LoP) hypothesis postulates that transition from unaware to aware visual stimuli is either graded or dichotomous depending on the depth of stimulus processing. Humans can be progressively aware of the low-level features, such as colors or shapes, while the high-level features, such as semantic category, enter consciousness in an all-or none fashion. Unlike in vision, sounds always unfold in time, which might require mechanisms dissimilar from visual processing. We tested the LoP hypothesis in hearing for the first time by presenting participants with words of different categories, spoken in different pitches near the perceptual threshold. We also assessed whether different electrophysiological correlates of consciousness, the auditory awareness negativity (AAN) and late positivity (LP), were associated with LoP. Our findings indicate that LoP also applies to the auditory modality. AAN is an early correlate of awareness independent of LoP, while LP was modulated by awareness, performance accuracy and the level of processing.

Effects of Inversion and Fixation Location on the Processing of Face and House Stimuli - A Mass Univariate Analysis

Most Event Related Potential studies investigating the time course of visual processing have focused mainly on the N170 component. Stimulus orientation affects the N170 amplitude for faces but not for objects, a finding interpreted as reflecting holistic/configural processing for faces and featural processing for objects. Furthermore, while recent studies suggest where on the face people fixate impacts the N170, fixation location effects have not been investigated in objects. A data-driven mass univariate analysis (all time points and electrodes) was used to investigate the time course of inversion and fixation location effects on the neural processing of faces and houses. Strong and widespread orientation effects were found for both faces and houses, from 100-350ms post-stimulus onset, including P1 and N170 components, and later, a finding arguing against a lack of holistic processing for houses. While no clear fixation effect was found for houses, fixation location strongly impacted face processing early, reflecting retinotopic mapping around the C2 and P1 components, and during the N170-P2 interval. Face inversion effects were also largest for nasion fixation around 120ms. The results support the view that facial feature integration (1) depends on which feature is being fixated and where the other features are situated in the visual field, (2) occurs maximally during the P1-N170 interval when fixation is on the nasion and (3) continues past 200ms, suggesting the N170 peak, where weak effects were found, might be an inflexion point between processes rather than the end of a feature integration into a whole process.

Optimal filters for ERP research I: A general approach for selecting filter settings

Filtering plays an essential role in event-related potential (ERP) research, but filter settings are usually chosen on the basis of historical precedent, lab lore, or informal analyses. This reflects, in part, the lack of a well-reasoned, easily implemented method for identifying the optimal filter settings for a given type of ERP data. To fill this gap, we developed an approach that involves finding the filter settings that maximize the signal-to-noise ratio for a specific amplitude score (or minimizes the noise for a latency score) while minimizing waveform distortion. The signal is estimated by obtaining the amplitude score from the grand average ERP waveform (usually a difference waveform). The noise is estimated using the standardized measurement error of the single-subject scores. Waveform distortion is estimated by passing noise-free simulated data through the filters. This approach allows researchers to determine the most appropriate filter settings for their specific scoring methods, experimental designs, subject populations, recording setups, and scientific questions. We have provided a set of tools in ERPLAB Toolbox to make it easy for researchers to implement this approach with their own data.

lmeEEG: Mass linear mixed-effects modeling of EEG data with crossed random effects

BACKGROUND

Mixed-effects models are the current standard for the analysis of behavioral studies in psycholinguistics and related fields, given their ability to simultaneously model crossed random effects for subjects and items. However, they are hardly applied in neuroimaging and psychophysiology, where the use of mass univariate analyses in combination with permutation testing would be too computationally demanding to be practicable with mixed models.

NEW METHOD

Here, we propose and validate an analytical strategy that enables the use of linear mixed models (LMM) with crossed random intercepts in mass univariate analyses of EEG data (lmeEEG). It avoids the unfeasible computational costs that would arise from massive permutation testing with LMM using a simple solution: removing random-effects contributions from EEG data and performing mass univariate linear analysis and permutations on the obtained marginal EEG.

RESULTS

lmeEEG showed excellent performance properties in terms of power and false positive rate.

COMPARISON WITH EXISTING METHODS

lmeEEG overcomes the computational costs of standard available approaches (our method was indeed more than 300 times faster).

CONCLUSIONS

lmeEEG allows researchers to use mixed models with EEG mass univariate analyses. Thanks to the possibility offered by the method described here, we anticipate that LMM will become increasingly important in neuroscience. Data and codes are available at osf.io/kw87a. The codes and a tutorial are also available at github.com/antovis86/lmeEEG.

Optimal Filters for ERP Research I: A General Approach for Selecting Filter Settings

Filtering plays an essential role in event-related potential (ERP) research, but filter settings are usually chosen on the basis of historical precedent, lab lore, or informal analyses. This reflects, in part, the lack of a well-reasoned, easily implemented method for identifying the optimal filter settings for a given type of ERP data. To fill this gap, we developed an approach that involves finding the filter settings that maximize the signal-to-noise ratio for a specific amplitude score (or minimizes the noise for a latency score) while minimizing waveform distortion. The signal is estimated by obtaining the amplitude score from the grand average ERP waveform (usually a difference waveform). The noise is estimated using the standardized measurement error of the single-subject scores. Waveform distortion is estimated by passing noise-free simulated data through the filters. This approach allows researchers to determine the most appropriate filter settings for their specific scoring methods, experimental designs, subject populations, recording setups, and scientific questions. We have provided a set of tools in ERPLAB Toolbox to make it easy for researchers to implement this approach with their own data.

Instant Effects of Semantic Information on Visual Perception

Does our perception of an object change once we discover what function it serves? We showed human participants (n = 48, 31 females and 17 males) pictures of unfamiliar objects either together with keywords matching their function, leading to semantically informed perception, or together with nonmatching keywords, resulting in uninformed perception. We measured event-related potentials to investigate at which stages in the visual processing hierarchy these two types of object perception differed from one another. We found that semantically informed compared with uninformed perception was associated with larger amplitudes in the N170 component (150-200 ms), reduced amplitudes in the N400 component (400-700 ms), and a late decrease in alpha/beta band power. When the same objects were presented once more without any information, the N400 and event-related power effects persisted, and we also observed enlarged amplitudes in the P1 component (100-150 ms) in response to objects for which semantically informed perception had taken place. Consistent with previous work, this suggests that obtaining semantic information about previously unfamiliar objects alters aspects of their lower-level visual perception (P1 component), higher-level visual perception (N170 component), and semantic processing (N400 component, event-related power). Our study is the first to show that such effects occur instantly after semantic information has been provided for the first time, without requiring extensive learning.SIGNIFICANCE STATEMENT There has been a long-standing debate about whether or not higher-level cognitive capacities, such as semantic knowledge, can influence lower-level perceptual processing in a top-down fashion. Here we could show, for the first time, that information about the function of previously unfamiliar objects immediately influences cortical processing within less than 200 ms. Of note, this influence does not require training or experience with the objects and related semantic information. Therefore, our study is the first to show effects of cognition on perception while ruling out the possibility that prior knowledge merely acts by preactivating or altering stored visual representations. Instead, this knowledge seems to alter perception online, thus providing a compelling case against the impenetrability of perception by cognition.

The early processing of fearful and happy facial expressions is independent of task demands - Support from mass univariate analyses

Most ERP studies on facial expressions of emotion have yielded inconsistent results regarding the time course of emotion effects and their possible modulation by task demands. Most studies have used classical statistical methods with a high likelihood of type I and type II errors, which can be limited with Mass Univariate statistics. FMUT and LIMO are currently the only two available toolboxes for Mass Univariate analysis of ERP data and use different fundamental statistics. Yet, no direct comparison of their output has been performed on the same dataset. Given the current push to transition to robust statistics to increase results replicability, here we compared the output of these toolboxes on data previously analyzed using classic approaches (Itier & Neath-Tavares, 2017). The early (0-352 ms) processing of fearful, happy, and neutral faces was investigated under three tasks in a within-subject design that also controlled gaze fixation location. Both toolboxes revealed main effects of emotion and task but neither yielded an interaction between the two, confirming the early processing of fear and happy expressions is largely independent of task demands. Both toolboxes found virtually no difference between neutral and happy expressions, while fearful (compared to neutral and happy) expressions modulated the N170 and EPN but elicited maximum effects after the N170 peak, around 190 ms. Similarities and differences in the spatial and temporal extent of these effects are discussed in comparison to the published classical analysis and the rest of the ERP literature.

Standardized measurement error: A universal metric of data quality for averaged event-related potentials

Event-related potentials (ERPs) can be very noisy, and yet, there is no widely accepted metric of ERP data quality. Here, we propose a universal measure of data quality for ERP research-the standardized measurement error (SME)-which is a special case of the standard error of measurement. Whereas some existing metrics provide a generic quantification of the noise level, the SME quantifies the data quality (precision) for the specific amplitude or latency value being measured in a given study (e.g., the peak latency of the P3 wave). It can be applied to virtually any value that is derived from averaged ERP waveforms, making it a universal measure of data quality. In addition, the SME quantifies the data quality for each individual participant, making it possible to identify participants with low-quality data and "bad" channels. When appropriately aggregated across individuals, SME values can be used to quantify the combined impact of the single-trial EEG noise and the number of trials being averaged together on the effect size and statistical power in a given experiment. If SME values were regularly included in published articles, researchers could identify the recording and analysis procedures that produce the highest data quality, which could ultimately lead to increased effect sizes and greater replicability across the field.

Physiological correlates of subjective emotional valence and arousal dynamics while viewing films

An exploration of the physiological correlates of subjective emotional states has theoretical and practical significance. Previous studies have reported that subjective valence and arousal correspond to facial electromyography (EMG) and electrodermal activity (EDA), respectively, across stimuli. However, the reported results were inconsistent, no study investigated subjective-physiological concordance across time, and measures of arousal remain controversial. To investigate these issues, while healthy adults (n = 20) viewed emotional films, we assessed overall and continuous ratings of valence and arousal and recorded EMG from the corrugator supercilii and zygomatic major, EDA from the palms and forehead, and nose-tip temperature. The corrugator and zygomatic EMG were negatively and positively associated with valence ratings, respectively, across stimuli and time. EDA (both sites) and nose-tip temperature were positively and negatively associated with arousal ratings, respectively, across stimuli and time. It is concluded that subjective emotional valence and arousal dynamics have specific physiological correlates.