Eye-Tracking as a Tool to Evaluate Functional Ability in Everyday Tasks in Glaucoma

Table 1

Summary of eye-tracking studies referenced in this work with regard to their participants and eye-tracking devices.

Study	Cohort demographics	Eye-tracker (fps)	Main findings
Burton et al. [48]	53 bilateral glaucoma (mean age ); 40 controls (mean age )	EyeLink 1000 (1000)	Reduction in reading speed for lower contrast text was greater in glaucoma patients than controls.
Smith et al. [49]	14 bilateral glaucoma (median age 69, IQR 64 to 81)	EyeLink 1000 (1000)	Slower performance and more regression when reading with the worse eye, compared to better eye. Differences in performance not related to magnitude of difference in VF mean deviation index between eyes.
Burton et al. [50]	18 advanced bilateral glaucoma (mean age ); 39 controls (mean age )	EyeLink 1000 (500)	Similar reading speeds between groups. Some glaucoma patients read slower than controls, partly explained by differences in eye movement behavior.
Prado Vega et al. [20]	23 glaucoma (mean age ); 12 controls (mean age )	Smart Eye (60)	Glaucoma patients missed more peripherally projected stimuli during driving in a simulator than controls. Glaucoma patients did not use compensatory visual search patterns.
Kübler et al. [21]	6 binocular glaucoma (mean age ); 8 controls (mean age )	Dikablis (25)	Glaucoma patients who passed the driving test in the simulator showed increased number of head and gaze movements toward eccentric regions of the VF in comparison to patients who failed.
Crabb et al. [23]	9 binocular glaucoma (mean age ); 10 controls (mean age )	EyeLink (250)	Patients showed different eye movement characteristics (more saccades) than controls when viewing driving scenes in a hazard perception test.
Kasneci et al. [30]	10 binocular glaucoma (mean age ); 10 controls (mean age )	Dikablis (25)	Patients who passed the on-road driving test focused longer on the central VF and performed more glances towards the area of their VF defect than patients who failed.
Kübler et al. [51]	10 binocular glaucoma (mean age ); 10 controls (mean age )	Dikablis (25)	Patients can be identified based on their visual scanpath while driving above chance levels.
Sippel et al. [38]	10 binocular glaucoma (mean age ); 10 controls (mean age )	Dikablis (25)	Patients who showed good performance during supermarket shopping made more glances towards the VF defect area.
Vargas-Martín and Peli [52]	5 retinitis pigmentosa (mean age ); 3 controls (mean age )	ISCAN (60)	Retinitis pigmentosa patients exhibited narrower scanning strategy than controls.
Ivanov et al. [53]	25 retinitis pigmentosa (mean age )	Tobii Glasses (30)	An exploratory saccadic training improved search performance, as well as mobility performance.
Dive et al. [39]	12 bilateral glaucoma (mean age ); 13 controls (mean age )	iViewX^TM (50)	Glaucoma patients took longer to complete the task, with longer fixations and more eye and head movements, than controls.
Smith et al. [24]	20 bilateral glaucoma (mean age ); 20 controls (mean age )	EyeLink II (500)	Glaucoma patients took longer to find targets in photographs.
Crabb et al. [28]	44 glaucoma (median age 69, IQR 63–77); 32 controls (median age 70, IQR 64–75)	EyeLink 1000 (1000)	Differences in signature scanpath patterns when watching television could separate glaucoma from controls.