Task difficulty modulates the effect of mind wandering on phase dynamics

The study investigates how mind wandering—attention shifting from an ongoing task to task-unrelated thoughts—modulates phase dynamics of neural activity during sensory and motor processing, and whether task difficulty moderates these effects. Prior work shows mind wandering reduces ERP amplitudes in sensory and cognitive components (including P300), and can impair motor control, suggesting a shared mechanism for sensory and motor attenuation. Neural entrainment aligns intrinsic neural oscillations with external events; intertrial phase coherence (ITPC) quantifies the consistency of oscillatory phase across trials and serves as an index of temporal precision and segregation. The authors hypothesize that mind wandering, reflecting internally oriented cognition, reduces temporal precision (ITPC) of externally driven neural responses, and that task difficulty—by modulating executive resource demands—determines the extent to which mind wandering affects ITPC. Based on the resource-control account, easier tasks (lower external demands) should be more susceptible to mind wandering’s impact than difficult tasks. The study tests these hypotheses across four EEG experiments spanning visual input (semantic and color tasks) and motor output (key-release tasks) with manipulated difficulty.

The authors ground their work in the perceptual decoupling hypothesis, which posits that mind wandering decouples cognition from external stimuli and tasks. ERP studies report reduced early sensory and late cognitive processing during mind wandering, with P300 decreases indicating resource reallocation to internal thoughts. Mind wandering also disrupts motor control, consistent with shared mechanisms across sensory and motor domains. Neural entrainment and phase alignment provide a framework for how external events synchronize brain activity; ITPC captures the precision of this entrainment. Prior studies tie theta-band dynamics to language-memory interfaces, and midfrontal theta to cognitive control and choice integration. Movement-related delta-theta phase locking in contralateral motor regions supports motor execution. Task difficulty has been shown to modulate mind wandering’s effects on motor control and thought content. The Baseline model proposes reciprocal interdependence between internal and external cognition with shared neural substrates, suggesting phase dynamics could index shared resource allocation.

Design: Four EEG experiments with thought-probe methodology assessed on-task vs. off-task states in blocks. Experiments 1 and 2 were sensory (SART-like) tasks using the same Chinese word stimuli but different task demands; Experiments 3 and 4 were motor key-release tasks manipulating movement difficulty. Task difficulty was operationalized by behavioral demands indexed primarily by reaction times.

• Experiment 1 (semantic classification/SART): Participants responded to nonanimal words and withheld for animal words. Semantic difficulty was manipulated via high-frequency (HF, easier) versus low-frequency (LF, harder) words. EEG focus: left occipital electrodes (P7, PO7, PO5). ROI: theta (4–7 Hz), 100–300 ms post-stimulus.
• Experiment 2 (color classification): Participants judged word color (respond to colored words; withhold to white). Despite focusing on color, semantics could be automatically activated, potentially inducing interference. EEG focus: occipital electrodes bilaterally (left: P7/PO7/PO5; right: P8/PO8/PO6) for visual processing; additionally, midfrontal electrodes (FC1, FCZ, F1) to probe cognitive control. ROI: theta (4–7 Hz), 100–300 ms post-stimulus.
• Experiment 3 (motor key-release with handedness): Right-handed participants executed key release with right index (RI, easier) and left index (LI, harder) fingers. EEG focus: contralateral fronto-central (RI: F1/F3/FC1/FC3; LI: F2/FZ/FC2/FCZ) and centro-parietal (RI: CP3; LI: CP4) regions. ROI: delta-theta (2–7 Hz), −100 to 100 ms around movement onset.
• Experiment 4 (more difficult two-finger motor): Right-hand two-finger releases, RIM (index+middle) and RIR (index+ring), both more difficult than RI. EEG focus: left-hemisphere fronto-central (F1/F3/FC1/FC3) and centro-parietal (CP3, P3). ROI: delta-theta (2–7 Hz), −100 to 100 ms around movement onset. Thought probes: At the end of each block, participants reported whether they had been on-task or off-task. The six trials preceding each probe were categorized accordingly (e.g., HF-On vs. HF-Off). EEG acquisition and preprocessing: 64-channel EEG; rereferenced, bandpass filtered; epochs extracted (Experiments 1–2: 1500 ms, Experiments 3–4: 2200 ms); manual artifact rejection; ocular/motion artifact correction via two-step ICA (EEGLAB/EPOS). ITPC computed following established methods; electrodes, time windows, and frequency bands selected based on prior knowledge and data-driven topographies. Behavioral measures: Reaction time (RT) mean and variability (ICV = SD/mean). Commission errors reported in SI for SART. Statistics: Two-factor repeated-measures ANOVAs tested effects of Probe (On vs. Off) and Difficulty factor (e.g., HF vs. LF; RI vs. LI; RIM vs. RIR). Significant interactions followed by simple effects and noninferiority t-tests (margin 0.01) to assess practical significance. Spearman correlations evaluated associations between ITPC and RT mean/ICV across conditions and ROIs. Data availability: OSF (https://osf.io/pj527/).

• Mind wandering prevalence: ~50% across experiments; no significant differences between difficulty conditions in Experiments 1–3; in Experiment 4, RIM (40.38%, SE 1.16%) < RIR (47.97%, SE 1.54%).
• Behavioral variability: Mind wandering increased RT variability (ICV) across all experiments; also slowed RTs in Experiments 2–4. Experiment 1 (semantic classification, left occipital P7/PO7/PO5, theta 4–7 Hz, 100–300 ms):
• ANOVA: main effect of Probe, F(1,28)=14.368, P<0.001, η²=0.339; Probe×Word frequency interaction, F(1,28)=4.271, P=0.048, η²=0.132; no main effect of Word frequency, F(1,28)=1.268, P=0.270.
• Simple effects: HF-Off < HF-On (P<0.001, η²=0.359); LF-Off vs. LF-On n.s. (P=0.657).
• Noninferiority t-tests (margin 0.01): HF significant, t(28)=2.88, P=0.008, 95% CI [0.018, 0.056]; LF not significant, t(28)=0.51, P=0.62. Experiment 2 (color classification):
• Occipital theta ITPC (100–300 ms, 4–7 Hz): Left (P7/PO7/PO5): main Probe effect F(1,28)=6.442, P=0.017, η²=0.187; marginal Word frequency F=4.155, P=0.051; no interaction. Right (P8/PO8/PO6): main Probe effect F(1,28)=8.206, P=0.008, η²=0.227; others n.s. Mind wandering reduced visual ITPC bilaterally in this easy sensory task.
• Midfrontal theta (FC1/FCZ/F1): main Probe effect F(1,28)=8.171, P=0.008, η²=0.226; Probe×Word frequency interaction F(1,28)=4.211, P=0.049, η²=0.131; no main Word frequency effect. Simple effects: HF-Off < HF-On (P=0.007, η²≈0.296); LF-Off vs. LF-On n.s. (P=0.626). Noninferiority: HF significant t(28)=2.92, P=0.006, 95% CI [0.029, 0.117]; LF not significant t(28)=1.20, P=0.240. Experiment 3 (motor key-release with handedness; delta-theta 2–7 Hz, −100 to 100 ms):
• Fronto-central (contralateral): main Probe marginal F(1,27)=4.022, P=0.055; main Hand F(1,27)=4.835, P=0.037; Probe×Hand interaction F(1,27)=4.833, P=0.037. Simple effects: RI-Off < RI-On (P=0.034, ηp²=0.268); LI-Off vs. LI-On n.s. (P=0.709). Noninferiority: RI significant t(27)=2.14, P=0.042, 95% CI [0.011, 0.045]; LI not significant t(27)=−0.61, P=0.544.
• Centro-parietal (contralateral CP3/CP4): main Probe F(1,27)=20.044, P<0.001; no Hand effect, no interaction. Mind wandering reduced ITPC for both RI and LI at CP. Experiment 4 (two-finger, more difficult, right-hand; delta-theta 2–7 Hz, −100 to 100 ms):
• Fronto-central (F1/F3/FC1/FC3): no main effects (Probe F=0.292, P=0.593; Type F=0.882, P=0.356) and no interaction.
• Centro-parietal (CP3, P3): main Probe F(1,28)=5.295, P=0.029; no Type effect; no interaction. Mind wandering reduced ITPC at CP for both RIM and RIR. Correlations (ITPC vs. behavior):
• Sensory tasks (Experiments 1–2, left occipital P7/PO7/PO5, 4–7 Hz, 100–300 ms): Negative correlations with RT mean and ICV. • Exp1 overall: RT mean r=−0.352, P<0.001; ICV r=−0.474, P<0.001. Condition-wise significant in HF-On/Off; weaker/non-significant in LF. • Exp2 overall: RT mean r=−0.292, P=0.002; ICV r=−0.455, P<0.001. Strongest in HF-On and LF-On.
• Motor tasks (centro-parietal, 2–7 Hz, −100 to 100 ms): Exp3 RT mean r=−0.208, P=0.028; Exp4 r=−0.314, P<0.001; no consistent correlations with ICV. Overall: Mind wandering reduces ITPC in easier sensory and motor tasks (fronto-central motor sites), with difficult tasks showing preserved ITPC; centro-parietal motor ITPC reductions occur regardless of difficulty. Higher ITPC relates to faster and less variable performance.

Findings demonstrate that the impact of mind wandering on phase coherence is contingent on task difficulty: easy tasks show reduced ITPC during off-task states, whereas difficult tasks generally maintain high ITPC. This pattern holds across sensory input and motor output processes, supporting a shared mechanism. ITPC indexes temporal precision and segregation of neural responses relative to external events; reductions during mind wandering imply temporal imprecision in neural entrainment, consistent with longer-timescale internal processes dominating during off-task states. In sensory tasks, reduced theta ITPC at occipital sites and midfrontal control-related theta (only in easier conditions) indicates that mind wandering compromises both early visual alignment and control-related integration when demands are low. In motor tasks, fronto-central contralateral regions (linked to motor control) show difficulty-dependent ITPC reductions, while centro-parietal contralateral regions exhibit reductions regardless of difficulty, suggesting functional dissociations within the motor network regarding susceptibility to internal-external balance. These results align with the resource-control account: easy tasks engage fewer executive resources, permitting greater internally oriented processing that “dilutes” phase alignment to external events; difficult tasks constrain mind wandering, preserving temporal precision. The Baseline model’s reciprocal interdependence of internal and external cognition is supported, with phase coherence/dynamics serving as a neural substrate for shared cognitive-executive resources.

Across four EEG experiments, mind wandering decreased phase coherence (ITPC) in easy sensory and motor tasks but not in difficult ones, indicating that task difficulty modulates the influence of internal cognition on externally driven neural dynamics. The effects were robust in visual theta-band ITPC and motor delta-theta ITPC, with a notable distinction between fronto-central (difficulty-dependent) and centro-parietal (difficulty-independent) motor regions. Behavioral-neural correlations further showed that higher ITPC is associated with faster and less variable responses, underscoring ITPC as an index of temporal precision. These findings support a reciprocal balance between internal and external cognition, consistent with the Baseline model, and suggest that phase dynamics are a shared neural substrate mediating resource allocation. Future research should: (1) explore extremely challenging tasks to determine thresholds where mind wandering ceases to influence ITPC, (2) dissociate intentional vs. unintentional mind wandering and examine how their rates vary with task difficulty, and (3) probe how the temporal structure and complexity of mind-wandering content changes with task demands.

Task difficulty manipulations were relative (no extremely difficult conditions), preventing identification of an absolute threshold where mind wandering no longer affects ITPC. The results thus cannot determine whether, at very high difficulty, mind wandering would also reduce phase coherence due to inability to attend to external events. Thought probes employed binary on-task/off-task reports, limiting differentiation among types of mind wandering (e.g., intentional vs. unintentional), which may vary with task difficulty. Additionally, some key behavioral outcomes (e.g., commission errors) suggest motor decoupling rather than mind wandering per se in SART errors, highlighting interpretive nuances. Finally, selection of electrodes and frequency bands was guided by prior knowledge and ROI/topography, which, while principled, may overlook effects in other regions or bands.