Optimising analysis choices for multivariate decoding: creating pseudotrials using trial averaging and resampling

Scrivener, C. L.; Grootswagers, T.; Woolgar, A.

doi:10.1101/2023.10.04.560678

Cited by 4 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For ERP-based decoding, we band-pass filtered the data to 1-6 Hz to isolate low-frequency evoked potentials as described in [8]. To increase signal to noise, we then averaged pairs of trials into 'pseudo-trials' using a random selection of the available trials [74].…”

Section: Eeg Multivariate Decoding Analysismentioning

confidence: 99%

“…For ERP-based decoding, we band-pass filtered the data to 1-6 Hz to isolate low-frequency evoked potentials as described in 8 . To increase signal to noise, we then averaged pairs of trials into 'pseudo-trials' using a random selection of the available trials 81 . We then trained and tested a linear support vector machine classifier (LibSVM), on the filtered instantaneous voltage across all sensors at each timepoint (the interval between two timepoints is 20 ms), to distinguish the category of the trials using leave-one-pseudo-trial-out cross-validation.…”

Section: Eeg Multivariate Decoding Analysismentioning

confidence: 99%

See 1 more Smart Citation

Parietal alpha stimulation causally enhances attentional information coding in evoked and oscillatory activity

Lu,

Michael,

Scrivener

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

SummarySelective attention is a fundamental cognitive mechanism that allows people to prioritise task-relevant information while ignoring irrelevant information. Previous research has suggested key roles of parietal evoked potentials like N2pc, and parietal oscillatory responses like alpha power, in spatial attention tasks. However, the informational content of these signals is less clear. Spatial attention tasks require not only representation of spatial information (i.e., where to attend) but also task (e.g., what to attend to) and visual feature information (e.g., details of visual stimuli). Whether all these features are represented in both evoked and oscillatory responses, and with what dynamics, is unknown. Moreover, the question of whether parietal evoked potentials and/or alpha powercausallyinfluence the coding of these different task elements is yet unresolved. Here, we used concurrent TMS-EEG to causally manipulate parietal alpha power and evoked potentials and investigate their roles in coding task-relevant information in a selective attention task. First, using EEG-only data, we investigated the decodability and temporal dynamics of evoked potentials and alpha power in coding different aspects of a selective attention task (where to attend,what to attend to, andvisual feature information). We found that evoked potentials coded all three types of task-relevant information with distinct temporal dynamics, and alpha oscillations carried information regarding bothwhere to attendandwhat to attend to. Then, we applied rhythmic-TMS (rTMS) at individual alpha frequency over the right intraparietal sulcus (IPS), to examine the causal roles of evoked potentials and alpha oscillations in selective attention. Compared with the control arrhythmic-TMS, alpha rTMS increased power and inter-trial phase coherence in alpha band and induced more negative N2pc amplitudes. Moreover, alpha rTMS causally and specifically improved multivariate decoding accuracy of the information aboutwhere to attendbut notwhat to attend toorfeature informationduring task performance. This TMS-induced change in decoding information aboutwhere to attend, encoded in both evoked potentials and alpha power, predicted TMS-induced changes in response errors and reaction time. These findings illuminate the dynamics with which the complementary aspects of a selective attention task are encoded in evoked and oscillatory brain activity. Moreover, they reveal a specific and causal role of IPS-controlled evoked and oscillatory activity in carrying behaviour-driving information about where to focus attention.

show abstract

Section: Eeg Multivariate Decoding Analysismentioning

confidence: 99%

Section: Eeg Multivariate Decoding Analysismentioning

confidence: 99%

Parietal alpha stimulation causally enhances attentional information coding in evoked and oscillatory activity

Lu,

Michael,

Scrivener

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Aperiodic and oscillatory systems underpinning human domain-general cognition

Lu,

Dermody,

Duncan

et al. 2024

Commun Biol

View full text Add to dashboard Cite

Domain-general cognitive systems are essential for adaptive human behaviour, supporting various cognitive tasks through flexible neural mechanisms. While fMRI studies link frontoparietal network activation to increasing demands across various tasks, the electrophysiological mechanisms underlying this domain-general response to demand remain unclear. Here, we used MEG/EEG, and separated the aperiodic and oscillatory components of the signals to examine their roles in domain-general cognition across three cognitive tasks using multivariate analysis. We found that both aperiodic (broadband power, slope, and intercept) and oscillatory (theta, alpha, and beta power) components coded task demand and content across all subtasks. Aperiodic broadband power in particular strongly coded task demand, in a manner that generalised across all subtasks. Source estimation suggested that increasing cognitive demand decreased aperiodic broadband power across the brain, with the strongest modulations overlapping with the frontoparietal network. In contrast, oscillatory activity showed more localised patterns of modulation, primarily in frontal or occipital regions. These results provide insights into the electrophysiological underpinnings of human domain-general cognition, highlighting the critical role of aperiodic broadband power.

show abstract

Aperiodic and oscillatory systems underpinning human domain-general cognition

Lu,

Dermody,

Duncan

et al. 2024

Preprint

View full text Add to dashboard Cite

Domain-general cognitive systems are essential for adaptive human behaviour, supporting various cognitive tasks through flexible neural mechanisms. From decades of fMRI studies, we know that a particular network of frontoparietal brain regions plays a role in supporting many different kinds of cognitive activity, with increased activity and information coding in response to increasing task demands. However, it is not clear whether there is a corresponding domain-general electrophysiological response to demand, such as a change in aperiodic or oscillatory power. Here we used irregular-resampling auto-spectral analysis (IRASA) to separate the aperiodic and oscillatory components of MEG/EEG signals, and analysed them with multivariate pattern analysis (MVPA) to investigate their roles in domain-general cognition. We found that both aperiodic (broadband power, slope, and intercept) and oscillatory (theta, alpha, and beta power) components coded both task demand and content across three cognitive tasks. Aperiodic broadband power in particular strongly coded task demand, in a manner that generalised across all subtasks, suggesting that modulation of aperiodic broadband power may reflect a domain-general response to multiple sorts of cognitive demand. Source estimation suggested that increasing cognitive demand decreased aperiodic activity across most of the brain, with the strongest modulations partially overlapping with the frontoparietal multiple-demand network. In contrast, oscillatory activity in the theta, alpha and beta bands showed more localised patterns of modulation, primarily in frontal (beta, theta) or occipital (alpha, theta) regions. The spatial pattern of demand-related modulation was significantly correlated across space in individuals, with positive correlations between theta and beta power, while both were negatively correlated with alpha power. These results provide novel insights into the electrophysiological underpinnings of human domain-general cognition, suggesting roles for both aperiodic and oscillatory systems, with changes in aperiodic broadband power being the clearest domain-general electrophysiological correlate of demanding cognitive activity.

show abstract

Optimising analysis choices for multivariate decoding: creating pseudotrials using trial averaging and resampling

Cited by 4 publications

References 21 publications

Parietal alpha stimulation causally enhances attentional information coding in evoked and oscillatory activity

Parietal alpha stimulation causally enhances attentional information coding in evoked and oscillatory activity

Aperiodic and oscillatory systems underpinning human domain-general cognition

Aperiodic and oscillatory systems underpinning human domain-general cognition

Contact Info

Product

Resources

About