Horizontal Random Saccades

About

Horizontal Random Saccade (HRS) Test for Nystagmus. The HRS test presents a target stimulus, a white circular dot on a black background, of 0.2 degrees’ diameter on a horizontal plane. The dot begins in the center of the screen and moves randomly on the horizontal axis, every 400-1500ms, and is always visible. Three 'guaranteed points' are initiated as the first dot located in the center of the screen, then at 1/5, 2/5, 3/5 of the test duration of 60 seconds. Guaranteed points will appear for a duration of two seconds. Guaranteed points are at visual degrees from central are right +20, left -20, left –23 and central gaze at zero degrees. The eccentric points of gaze are used to calculate a nystagmus metric.

Report Metrics

Nystagmus (Horizontal Random Saccades): participants are required to track a small dot that moves randomly on the horizontal axis, for 60 seconds. Nystagmus, a voluntary eye movement, is captured when the dot stops in the central gaze position (0 degrees) and eccentric gaze positions (20 and 23 degrees).

1.[0, 20, 23] Degrees Velocity Fast Phase (Mean): the mean velocity of eye movements in deg/sec that are within 3 cm of the stimuli and moving in the direction from the center of the screen and passing the stimuli to one edge of the screen.

2.[0, 20, 23] Degrees Velocity Fast Phase (SD): the standard deviation of the velocity of eye movements that are within 3 cm of the stimuli and moving in the direction from the center of the screen and passing the stimuli to one edge of the screen.

3.[0, 20, 23] Degrees Velocity Slow Phase (Mean): the mean velocity of eye movements in deg/sec that are within 3 cm of the stimuli and moving in the direction from the edge of the screen and passing the stimuli to the center of the screen.

4.[0, 20, 23] Degrees Velocity Slow Phase (SD): the standard deviation of the velocity of eye movements that are within 3 cm of the stimuli and moving in the direction from the edge of the screen and passing the stimuli to the center of the screen.

5.[0, 20, 23] Degrees Duration Fast Phase (Mean): the mean duration of the eye movements in milliseconds that are within 3 cm of the stimuli and moving in the direction from the center of the screen and passing the stimuli to one edge of the screen.

6.[0, 20, 23] Degrees Duration Fast Phase (SD): the standard deviation of the duration of the eye movements that are within 3 cm of the stimuli and moving in the direction from the center of the screen and passing the stimuli to one edge of the screen.

7.[0, 20, 23] Degrees Duration Slow Phase (Mean): the mean duration of the eye movements in milliseconds that are within 3 cm of the stimuli and moving in the direction from the edge of the screen and passing the stimuli to the center of the screen.

8.[0, 20, 23] Degrees Duration Slow Phase (SD): the standard deviation of the duration of the eye movements that are within 3 cm of the stimuli and moving in the direction from the edge of the screen and passing the stimuli to the center of the screen.

Test-Retest Reliability Analysis

Download HRS Reliability and Validity Article 

The 62 eye tracking variables from trials 1 and 2 were analyzed using R (statistical package) reliability procedure.  Tables 6-11 present the means and standard deviations for trials 1 and 2, the Cronbach’s Alpha correlations between the Trial 1 and Trial 2 and associated the test-retest reliability decisions.  Table 12 presents the average Cronbach’s Alpha across all variables for each test. All eye tracking variables demonstrated Acceptable (.7) to Excellent (.9) test-retest reliability.  No variables were found to be Unacceptable (<.6).

Table 6.  Test-retest Reliability of Horizontal Random Saccades Digitized Eye Tracking Variables

Variable	Trial 1 Mean	Trial 1    SD	Trial 2  Mean	Trial 2    SD	CA	Decision
0 degrees velocity fast phase	0.96	0.39	0.89	0.15	0.8	Acceptable
0 degrees velocity slow phase	1.39	0.98	1.10	0.96	0.9	Acceptable
0 degrees duration fast phase	179.87	47.92	188	56.37	0.9	Acceptable
0 degrees duration slow phase	140.35	80.34	148.23	76.59	0.9	Acceptable
20 degrees velocity right fast phase	1.867	1.20	1.762	1.11	0.8	Acceptable
20 degrees velocity right slow phase	5.201	2.89	5.413	2.97	0.8	Acceptable
20 degrees duration right fast phase	403.78	145.88	434.65	153.26	0.9	Acceptable
20 degrees duration right slow phase	498.99	87.27	502.23	91.46	0.9	Acceptable
20 degrees velocity left fast phase	2.38	1.49	2.39	1.50	0.9	Acceptable
20 degrees velocity left slow phase	6.00	2.54	6.41	2.57	0.7	Acceptable
20 degrees duration left fast phase	354.23	120.29	362.90	138.48	0.9	Acceptable
20 degrees duration left slow phase	501.43	180.76	484.79	187.54	0.8	Acceptable
23 degrees velocity left fast phase	2.73	1.01	2.32	1.71	0.8	Acceptable
23 degrees velocity left slow phase	6.32	2.67	6.12	4.50	0.9	Acceptable
23 degrees duration left fast phase	386.68	180.55	397.57	172.14	0.9	Acceptable
23 degrees duration left slow phase	467.89	147.87	465.87	167.90	0.9	Acceptable

 

HRS Test Regression

The HRS logistic regression examined the predictive validity of the HRS test. The sixteen eye tracking sub variables from the HRS test were entered as the predictor/ independent variables and the clinical evaluation score (normal functioning coded as 0; abnormal functioning coded as 1) for horizontal random saccade behavior was entered as the predicted/ dependent variable.  The full model of sixteen predictor variables significantly predicted clinically evaluated HRS status (χ² = 51.730, df = 16, n = 52, p < .0005). The model accounted for between 78% to 89% of the variance (Nagelkerke R² = .890) in status classification with overall 92.3% of individuals correctly predicted to their known status.  Table 13 gives coefficients, odds ratios, and probability values (for Wald statistics) for each of the predictor variables.  For the HRS test, none of the individual sub variables were statistically significant predictors on their own.  That is, all of the Wald statistics were nonsignificant (p’s>.05). Table 14 presents the degree of agreement between the known status, and the logistic regression function predicted status classification. Based on these results, the following were found for identifying the normal status at 50.7% prevalence: sensitivity = .91, specificity = .94.

HRS Classification Table

Observed			Predicted
			HRS		% Correct
			Normal	Not Normal
HRS model	HRS	Normal	19	2	90.5
		Not Normal	2	29	93.5
	Overall Percentage				92.3

 

Cluster Analysis

The model-based clustering using EM algorithm analysis created five distinct age group: 5-8, 9-16, 17-28, 29-52, and 53-62. Further, we conducted stability testing to establish that the data sample used for cluster analysis that is representative of the entire population. The stability testing involved sub-sampling 10 individuals from the experimental population for each age group.  These sub-samples were then compared against the entire population norm to assess cluster solution. The comparison of the sample norms and the population norms showed the cluster solution was appropriate in numbers and quality (Calinski-Harabasz Index = 16.61 with average inter-cluster distance = 56.73).  The descriptive statistics for all variables derived from the six RightEye Vision Tests for the 5 clusters are shown tables 25-30.

Descriptive Statistics Horizontal Random Saccades Clustered by Age
	5 - 8		9 - 16		17 - 28		29 - 52		53 - 62
Test	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
0 Degrees velocity fast	0.89	0.88	0.92	0.61	0.95	0.39	0.97	0.59	0.97	0.62
0 Degrees velocity slow phase	1.20	0.99	1.26	0.87	1.40	0.64	1.41	0.73	1.39	0.97
0 Degrees duration fast phase	167.37	68.71	173.98	56.32	180.34	45.32	186.49	53.67	191.43	78.56
0 Degrees duration slow phase	133.78	90.32	137.68	70.62	141.23	50.43	145.32	61.21	152.78	67.89