Neural Representations Across Multiple Stages in a Go/No-Go Task: Evidence From Multivariate Pattern Analysis.

Area of Science:

• Cognitive Neuroscience
• Computational Psychiatry

Background:

• The Go/No-Go task is a standard paradigm in psychological research.
• Traditional analysis of event-related potentials (ERPs) in this task offers limited insight into temporal dynamics.

Purpose of the Study:

• To investigate the multi-stage temporal dynamics of neural representations during the Go/No-Go task.
• To compare univariate ERP analysis with multivariate pattern analysis (MVPA) for deeper insights.

Main Methods:

• Electroencephalography (EEG) data from 72 participants performing a Go/No-Go task.
• Univariate ERP analysis and MVPA techniques including decoding, temporal generalization, and weight projection.

Main Results:

• Go and No-Go trials were distinguishable by ~100 ms using MVPA.
• Distinct temporal phases of neural representation were identified: early (~300-400 ms), specific (~400-700 ms), and late (~700-1700 ms).
• Neural activity involved occipital, prefrontal, and parietal regions, with trial history influencing early and late processing stages.

Conclusions:

• The Go/No-Go task relies on a dynamic, multi-stage neural code.
• Neural representations evolve over time and are modulated by recent trial history.
• MVPA offers a more detailed understanding of temporal dynamics than traditional ERP analysis.