One hundred and eighteen patients with liver cirrhosis were recruited in the outpatient’s clinics of Hospital Clínico and Hospital Arnau de Vilanova, Valencia, Spain. The diagnosis of liver cirrhosis was based on clinical, biochemical, and ultrasonographic data. Exclusion criteria were overt HE or history of overt HE, recent (< 6 months) alcohol intake, infection, recent (< 6 weeks) antibiotic use or gastrointestinal bleeding, use of drugs affecting cognitive function, hepatocellular carcinoma, or neurological or psychiatric disorder. Thirty-five healthy volunteers were included in the study after discarding liver disease by clinical, analytical, and serologic analysis.
All participants were included after signing a written informed consent. Study protocols were approved by the Scientific and Ethical Committees (No. 2017/291) of Hospitals Clínico and Arnau de Vilanova, Valencia, Spain, and were in accordance with the ethical guidelines of the Helsinki Declaration.
The procedure was as follows: (1) recruitment of patients and controls, collecting written informed consent and analytical data; (2) Neuropsychological assessment and (3) eye movements evaluation. Eye movements were measured less than 1 week after neuropsychological measures.
All participants performed the Psychometric Hepatic Encephalopathy Score (PHES) battery, used for diagnosis of MHE. PHES includes five subtests: Digit Symbol test (DST), number connection test A and B (NCT-A and NCT-B), Serial Dotting test (SD), and Line Tracing test (LTT)18,19. The score obtained from each subtest was adjusted for age and educational level using Spanish normality tables (http://www.redeh.org/TEST_phes.htm). Patients were classified as MHE when their score was ≤ − 4 points18.
In order to evaluate selective attention, psychomotor speed, cognitive flexibility and inhibitory mental control, all participants performed the Stroop test, in a colour-word version, performing the congruent, neutral and incongruent tasks, as described in7. The number of items correctly named was adjusted for age according to Spanish normality tables.
Analysis of eye movements
Eye movements were assessed with an OSCANN desk100 equipment from AURA Innovative Robotics (see Fig. S1). OSCANN desk100 is a novel gaze-tracker40 designed for clinical practice use. It is based on VOG technology, and its infrared camera captures images at 100 frames per second. The measurements are made over the dominant eye of the subject (see Supporting information for technical characteristics and software description).
A total of 10 tests were performed in the following order: visually guided saccades test, memory guided saccades test, anti-saccades test, smooth pursuit eye movements, and fixation test (Fig. S2). All tests included a horizontal and a vertical version, with the exception of the smooth pursuit test, which included an additional horizontal version in which the stimulus changes its velocity; and the fixation test. Performing the 10 tests to obtain these 177 parameters takes from 15 to 20 min per subject approximately. Recalibration of the recording camera was performed every 2 or 3 tests to ensure the accuracy of the measurement, and to allow the patient to take a brief break if needed.
The stimulus was a dot 2 cm diameter green dot over a black background. All horizontal versions are performed across the central horizontal axis of the screen, and all vertical versions are performed across the central vertical axis of the screen. The visual field is ± 20° in horizontal and ± 12° in vertical.
In the visually guided saccade tests the stimulus appears in the centre of the screen during 1500 ms and then jumps to a random location on one side of the screen, where it remains the same amount of time (see Fig. S2a). Then, the stimulus appears back in the centre of the screen, and the sequence is repeated. The subject must follow the stimulus. Each version of the test lasts 36 s. Variables measured include response latency when the stimulus moves to a side of the screen (latency), response latency when the stimulus moves back to the centre of the screen (return latency), gain (ratio between stimulus and gaze amplitude), accuracy (hypermetria, identified as positive error, or hypometria, identified as negative error), number of blinks, peak of velocity and anticipated saccades (saccades performed within 80 ms since the stimulus moved) (see Fig. S3a).
In the memory guided saccades tests, the subjects are inquired to remember the position of the stimulus (see Fig. S2b). In this case the stimulus firstly appears in the centre of the screen for 1500 ms, then moves to a random location on one side of the screen and returns back to the centre of the screen, similarly to the visually guided saccade tests. In this initial phase the subject must perform a visually guided saccade towards the places where the stimulus is located. At this point, the stimulus disappears from the screen for another 1500 ms, and the subject must perform a memory guided saccade towards the last side of the screen where the stimulus moved. Each version of this test takes 72 s. The variables analyzed include the same as in the visually guided saccades, as well as the number of correct memory saccades performed.
In the anti-saccades tests the stimulus moves, similarly to the visually guided saccades tests, to a random side of the screen and then back to its centre, repeatedly. However, in these tests the subject must direct their gaze towards the position of the screen opposite of the stimulus, and only look at the stimulus when it is in the centre of the screen (Fig. S2c). Each version of this test takes 36 s. Variables analyzed include latency, velocity and duration of reflexive saccades (saccades automatically performed towards the stimulus instead of away from it), anti-saccade latency and velocity, and accuracy; and number of anticipated, corrected and successful anti-saccades (see Fig. S3b, and Supporting Information).
In the smooth pursuit tests the stimulus moves across the screen with a constant period of 8 s (see Fig. S2d). The subject must follow the stimulus all time. The test lasts 32 s. In this case, a third version of the test was performed, where the stimulus moves horizontally with the same frequency, but its velocity increases and decreases at a constant rate. Latency, number of blinks, catch-up and back-up saccades (saccadic movements performed to go back to the position of the stimulus or keep up with its movement, respectively), square wave jerks (a specific kind of fixation instability), pursuit time, gain, error pursuit, and velocity error are measured in these tests.
In the Fixation test the subject is inquired to look at the immobile stimulus, showed in the centre of the screen (Fig. S2e). This test lasts 20 s. This test measures variables like the number of blinks, the bivariate contour ellipse area (BCEA) and the amplitude, duration, frequency and/or velocity of various kinds of fixation instabilities, which include saccades, micro-saccades, drifts and square-wave jerks.
For many variables, where the final result is averaged from the numerous eye movements performed during the test, a second variable is also generated, which provides information about the standard deviation of the original variable from which it is derived (i.e., latency and standard deviation of latency). A detailed definition of all variables measured in all tests is shown in the Supporting Information.
All values are given as mean ± SEM unless otherwise stated. Results were analyzed using one of three options: one-way ANOVA followed by post-hoc Tukey’s multiple comparison test for variables both parametric and homoscedastic, Welch’s ANOVA followed by Games-Howell’s multiple comparison test for variables parametric but not homoscedastic, and Kruskal–Wallis’ test followed by Dunn’s test for non-parametric variables. Due to the number of variables analyzed in this study, P values obtained were corrected using the false discovery rate (FDR) method41. FDR values < 0.05 were considered significant. When analysing the correlation between different variables, Spearman’s correlation test was performed. For all statistical analyses, data were processed and analyzed using the software R version 4.1.142.
Receiver operating characteristic (ROC) curves were performed using SPSS software (version 24.0; SPSS, Inc., Chicago, IL, USA), and two-sided P values < 0.05 were considered significant.
Compliance with ethical standards
This study was performed in line with the principles of the Declaration of Helsinki. Study protocols were approved by the Scientific and Ethical Committees (No. 2017/291) of Hospitals Clinico and Arnau de Vilanova, Valencia, Spain. All participants were included after written informed consent.