Methodological aspects of using a wearable eye-tracker to support diagnostic clinical evaluation of prolonged disorders of consciousness

Abstract

Objective: To evaluate the feasibility of using a wearable eye-tracker when assessing patients with prolonged disorders of consciousness using the Coma Recovery Scale Revised (CRS-R), focusing on technical challenges.

Design: A methodological investigation with descriptive and analytical elements.

Subjects: Four patients with prolonged disorders of consciousness were recruited from the rehabilitation clinic of a regional rehabilitation unit.

Methods: A selection of subtests in the CRS-R were performed while recording eye movements with a wearable eye-tracker.

Results: No major adverse reactions were observed, suggesting likely patient acceptability. Calibration was not always possible. However, distinct eye movements were discernible from the recorded data even without calibration, and analysis of these produced results with the potential to support clinical assessment.

Conclusion: Eye tracking was feasible during clinical assessment for this patient group. Recording eye movement responses in these easily fatigued patients has the potential to add sensitivity for detection of conscious responses and to complement clinical examination. Further study is merited. Current hardware and software limitations can be overcome with manual data processing and analysis; however, significant developments in automating data processing will be required for broader clinical application.

Keywords: CRS-R; Coma Recovery Scale Revised; diagnostic accuracy; disorders of consciousness; eye tracking.

Fig. 1

The 2 large diameter circles represent the areas of interest (AOIs). Visual fixations are symbolized by numbered circles, where a larger diameter represents a longer fixation duration. In this example most of the patient’s fixations were pointed straight ahead towards the examiner, 3 fixations pointed at the left-hand side AOI (hits) and none pointed at the right-hand side AOI.

Fig. 2

Graph showing 10 s of eye movement recording before the tests started (raw data). No visual or auditory stimuli were presented at this time. The gaze is pointed to the left at a fairly stable vector value of 0.4. There are gaze drifts towards the left-hand side, interrupted by rebound saccades to the right every 1,200–1,400 ms. A linear fit (right eye) resulted in an intercept 0.434, R-square 0.059 with a slope (6.225*10^–6) that was significantly different from zero.

Fig. 3

Graph showing eye movement recording when the patient is asked to look at an object to the left. An increasing vector value is expected if the patient responds appropriately. A sequential and consistent increase in vector value (at 10,500–13,500 ms), followed by a rebound change (at 13,500–14,500 ms), was observed during the exposure to the stimuli.

Fig. 4

No stimulus. A stable fixation just to the left of the centre of the scene camera view. Due to technical reasons, the pre-test recording was, unfortunately, shorter than intended.

Fig. 5

Visual pursuit, in which the mirror is moved from left to right. In the first phase (67,000–69,000 ms) the patient’s visual fixation on the mirror image is established. At approximately 69,200 ms the mirror is starting to move towards the right (increasing pixel-value) and a smooth visual pursuit can be seen up until the endpoint at 73,000 ms.

Fig. 6

Graph of the subsequent gaze direction vector while the mirror is moved sideways toward the patient’s right-hand side. A sequential and consistent decrease in vector value was observed during the exposure to the stimuli. This corresponded to the expected eye movement behaviour when the patient’s eyes move to the right.

Fig. 7

Fixation test. The lamp was moved from centre position and downward. According to the coordinate system of the scene camera an increase in pixel value corresponds to a downward movement. In the plot it can be seen that the initial vertical gaze direction (time stamp 166,000–167,000 ms) is at the centre (420–600 pixels). It then increases value, peaking at almost 1,080 pixels, after which it returns to centre value (time stamp 171,000 ms).