This is your brain on autopilot 2.0: The influence of practice on driver workload and engagement during on-road, partially automated driving

McDonnell, AS; Crabtree, KW; Cooper, JM; Strayer, DL · 2024 · publications_jsonl

DOI: 10.1177/00187208231201054

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 how six weeks of practice with Level 2 partially automated driving influences driver workload and engagement. While Level 2 automation handles steering and acceleration, drivers must remain supervisory monitors to handle "edge cases." The authors address a gap in the literature regarding whether familiarity with this technology reduces cognitive load or leads to disengagement over time, a concern known as the automation conundrum. Previous research yielded mixed results, with some studies suggesting drivers disengage while others found no change in physiological arousal. This research specifically tests whether extended practice alters these cognitive states during real-world on-road driving. The study employed a 2 (Level of Automation: Level 0 manual vs. Level 2 partial) × 2 (Interstate: I-15 simple vs. I-80 complex) × 2 (Session: 1 vs. 2) factorial design with 30 participants. Participants drove on two interstate highways with varying complexity. Cognitive workload was measured behaviorally using a Detection Response Task (DRT), where participants responded to vibrotactile stimuli, and neurophysiologically using EEG metrics of frontal theta power (workload) and parietal alpha power (engagement). Session 1 served as the baseline. Participants then took the test vehicle home for a six-week familiarization period, driving in partial automation mode daily. Session 2 repeated the experimental driving conditions to assess changes after practice. Behavioral results indicated that partially automated driving initially placed higher cognitive load on drivers than manual driving, evidenced by slower DRT reaction times. However, six weeks of practice significantly decreased this workload, but only when driving in the simpler environment (I-15). Reaction times improved from Session 1 to Session 2 exclusively under partial automation conditions on the simpler highway; no significant practice effect was observed for manual driving or on the more complex I-80 highway. Conversely, neurophysiological measures showed null effects. EEG data revealed no significant differences in frontal theta power or parietal alpha power between manual and partially automated driving, nor did these neural metrics change significantly between Session 1 and Session 2. The findings suggest that driver workload is influenced by automation level, highway complexity, and practice, but these effects are detectable only at the behavioral level, not the neural level. The lack of change in EEG metrics implies that while drivers may become behaviorally more efficient with practice, their underlying neural engagement and workload states do not shift in the expected directions. This discrepancy highlights the complexity of measuring driver state and suggests that behavioral metrics like DRT may be more sensitive to the immediate cognitive demands of supervisory control than the specific EEG bands tested. The study concludes that practice reduces the behavioral burden of partial automation in simple environments, but does not alter the fundamental neural engagement patterns associated with monitoring automated systems.

Key finding

Level 2 partial automation increased rather than decreased driver workload (slower DRT RTs vs. manual driving), and six weeks of daily practice reduced that workload only on the simpler highway (I-15), not on the more complex I-80. EEG frontal theta and parietal alpha showed null effects of automation level and session, indicating practice and automation effects appeared behaviorally but not neurally.

Methodology

on_road

Sample size: N=30 (12 female, 18 male; M_age=35.73, SD=9.34)

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. Discovered via tag_papers on 2026-05-30 (3 acquisition events logged).

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• hands on hands off engagement
• automation
• situational awareness
• temporal
• mental demand
• automation complacency bias

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: physiological data, behavioral performance data
• Theoretical Contribution: theory or model