Head Scanning by Drivers With Left Homonymous Visual Field Loss: A Naturalistic Driving Pilot Study

Little is known about compensatory scanning by drivers with homonymous visual field loss (HVFL) in daily driving.

To use naturalistic driving methods to investigate the feasibility of quantifying head scanning behaviors of HVFL drivers in daily driving.

Cross-sectional pilot study with video and GPS data collected from dashboard cameras installed in participants' personal vehicles between October 2023 and June 2024.

Intersections with left-turn maneuvers in daily driving.

Two drivers with HVFL (ages 52 and 63 yr) and two with normal vision (NV; ages 47 and 62 yr), all male, recruited from study site databases.

Head scanning behaviors were quantified from driver-facing videos, including the number and magnitude of leftward and rightward scans. Road-facing videos were manually annotated for environmental features associated with situations in which no leftward scan was made.

HVFL drivers exhibited greater leftward than rightward scan magnitude (∼20° difference), whereas NV drivers showed minimal difference (∼3°). Cross-traffic from the left increased left scan frequency for HVFL drivers but not NV drivers. HVFL drivers never failed to scan to the left at stop-sign intersections, although NV drivers sometimes did, but they did fail to scan left at some traffic light and no-sign intersections.

These results demonstrate the feasibility of using naturalistic driving methods to quantify compensatory scanning as well as failures to scan in daily driving of individuals with HVFL. The methods have potential for application in larger-scale research studies as well as in clinical driving rehabilitation. Plain-Language Summary: Drivers with homonymous visual field loss (HVFL), the loss of half the visual field in both eyes, need to make large head scans (head turns) to see hazards approaching on the nonseeing side. However, little is known about how drivers with HVFL scan in daily driving because most prior studies have been conducted in a driving simulator or during a road test in an unfamiliar vehicle. We used an in-car camera in each participant's own vehicle to record head scanning before left turns at intersections. Drivers with HVFL made larger and more frequent head scans toward their nonseeing side than drivers with normal vision, suggesting compensatory scanning by HVFL drivers; however, both groups sometimes failed to scan in low-demand situations (e.g., when no traffic was present). These findings suggest that our methods can be successfully used to investigate scanning behaviors in daily driving and are feasible for implementation in a study with more participants.