Trial-level sequence modeling reveals hidden dynamics of dual-task interference
DOI: 10.1371/journal.pcbi.1014302
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates the neural dynamics of dual-task interference, specifically addressing whether the same cognitive operations underlie multitasking regardless of stimulus timing. Traditional theories, such as the central bottleneck and capacity sharing models, assume that cognitive processes are merely delayed or shortened during interference but have not directly tested if the underlying operations remain identical. The authors argue that previous reliance on trial-averaged data obscures trial-by-trial variability and hidden dynamics. To resolve this, they employed a novel methodological approach combining Hidden Multivariate Pattern (HMP) analysis with deep spatiotemporal sequence modeling of single-trial electroencephalography (EEG) data. The researchers analyzed an open dataset from a psychological refractory period (PRP) paradigm involving 24 participants performing a visual flanker task (Task 1) followed by an auditory pitch discrimination task (Task 2). Stimulus-onset asynchrony (SOA) was manipulated into Long (1200 ms) and Short (300 ms) conditions. First, HMP analysis was applied to the Long SOA condition to identify three distinct cognitive operations for each task: Encoding, Central, and Response. These operations served as ground truth labels to train a deep spatiotemporal sequence model based on the Mamba state space architecture. This model was then used to decode single-trial EEG data in both Long and Short conditions, allowing for the identification of cognitive operations and their temporal sequences on a trial-by-trial basis. The results demonstrated that the same Encoding, Central, and Response operations identified in the Long SOA condition were present in the Short SOA condition, confirming shared cognitive processes across interference levels. Embedding analysis showed high similarity between operations across conditions, with negligible cosine distance differences for the same operations compared to other operations. Crucially, the study revealed that the sequence of these operations varied significantly in the Short SOA condition. While the Long condition predominantly featured a serial sequence (Task 1 completes before Task 2 begins), the Short condition exhibited six distinct sequences, including five where Task 2 operations initiated before Task 1 completion. These specific sequences predicted behavioral performance; linear mixed-effects models showed that the sequence type significantly influenced Task 1 reaction times and accuracy. For instance, sequences where Task 2 interference occurred earlier in the Task 1 processing stream resulted in different performance outcomes compared to those where interference occurred later. These findings challenge static accounts of dual-task interference, such as the rigid central bottleneck theory, by demonstrating that multitasking involves dynamic, variable sequences of cognitive operations. The study establishes that individual strategies and trial-by-trial variability in the timing of cognitive operations directly affect performance. By validating trial-level sequence modeling as a powerful tool, the authors provide a framework for understanding the hidden neural dynamics of multitasking, suggesting that interference is not a uniform delay but a complex interaction of flexible cognitive sequences. This approach offers a more nuanced view of executive control, highlighting the importance of single-trial analysis over traditional averaging methods in cognitive neuroscience.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | Crossref | — | — | 1 | 2026-06-11 |
| archive | success | canonical_url | — | — | 1 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-25 |
| clean | success | clean | — | — | 1 | 2026-06-11 |
| chunk | success | chunk | — | — | 1 | 2026-06-11 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-11 |
| promote | success | — | — | — | 1 | 2026-06-11 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-25 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: behavioral performance data