Emerging Techniques and Future Development Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

Detecting Eye Movement Abnormalities from Concussion Jun Maruta a  · Jamshid Ghajar a, b  

a

 

b

Brain Trauma Foundation, and Department of Neurosurgery, Stanford University School of Medicine, Stanford, Calif., USA  

 

Abstract An attention-based biomarker may be useful for concussion screening. A key role of attention is to generate time-based expectancies of specific sensory information, and it is postulated that postconcussion cognitive impairments and symptoms may stem from a primary deficit in this predictive timing mechanism. There is a close relationship between gaze and attention, but in addressing predictive timing, there is a need for an appropriate testing paradigm and methods to quantify oculomotor anomalies. We have utilized a continuous predictive visual tracking paradigm because human visual tracking requires predicting the temporal course of a stimulus and dynamically synchronizing the required action with the stimulus. We have shown that concussion patients often show disrupted gaze-target synchronization characterized by large gaze position error variability and overall phase advancement. Various attention components interact with visual tracking, and thus there is a possibility that different neurological and physiological conditions produce identifiable visual tracking characteristics. Analyzing neuromotor functions, specifically oculomotor synchronization, can © 2014 S. Karger AG, Basel provide a fast, accurate, and reliable assessment of cognitive functions.

Detecting eye movement abnormalities may provide a fast, accurate, and reliable way to screen for concussion. Eyes move for different reasons and purposes. For example, eyes may move to stabilize or optimize vision when there is a relative movement of the visual world, e.g. during head movements. This class of eye movements is phylogenetically old and largely under reflexive control. Another important goal of eye movement is to foveate, i.e. to orient the eyes to capture the image of a specific object within the fovea. Not all animals have a fovea, but it is highly developed in humans, allowing for highly specific image analysis by the brain. On the other hand, the central allocation of this specialized function entails sharp vision to be limited only within a small selected part of the captured image, and redirecting this locus of sharp vision requires

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Eye Movement and Eye Movement Recording

Rotation

Extraocular muscle Motoneuron + input

Muscle contraction

a

Essence of video-oculography Image acquisition Image analysis Gaze estimation

b

movement of the eye, i.e. foveation. Afoveate animals, such as rabbits, do not move the eyes around to examine objects [1]. Although visual acuity is determined by foveal vision, the relationship between the ability to see clearly and to orient the gaze accurately tends to be overlooked. However, the relationship is most easily described in terms of attention, as inherently suggested by selective allocation of resources for information processing. The neural basis of the overlap between gaze and attention networks has been demonstrated using functional magnetic resonance imaging [2]. Eye movements are rotation controlled by a set of muscles attached to the eyeball and the skull (fig. 1a). Eye movements can be described in terms of angles of rotation in specific coordinate planes or axes. There are different classes of eye movement, which include visual fixation, saccade, and smooth pursuit. Visual fixation holds the image of a stationary object on the fovea when the head is stationary. It is not a nonmovement, but is under active neural control. A saccade causes a rapid change in the eye orientation and can bring the image of the object of interest onto the fovea. Smooth pursuit aids continuous stabilization of the image of a moving target on the fovea. There are other

Detecting Eye Movement Abnormalities from Concussion Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

227

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Fig. 1. Eye movement measurement. a Specific objects are brought into the line of sight through eye rotation. b Video-oculography reconstructs angles of eye rotation from two-dimensional images. The crosshairs indicate centers of the pupil and the reflection of an infrared light source.

classes of movements including vergence and ocular reflexes, as well as those resulting from a failure to hold a steady gaze, such as tremor, drift, and spontaneous nystagmus. Visual fixation, smooth pursuit, saccades, and vergence are elements of eye movement that can support foveal vision and thus can be purposive. However, different classes of eye movements can co-occur or interact with one another. Quantification of eye movements can provide valuable clues related to the working of the brain. One way to measure eye movement is to analyze sequential images of the eye (fig. 1b). Recent technological advances in image processing and miniaturization have produced noninvasive video-oculography units capable of recording eye positions with high spatial and temporal resolutions, and appropriate for clinical settings. Three-dimensional rotation of the eye is captured as changes in the locations of specific landmarks within two-dimensional images, from which the angle of eye rotation that physically took place is estimated. The pupil is usually very dark and has an elliptical shape. These features can be quickly identified, frame-by-frame, using a computer, which makes the pupil a good landmark for tracking. Another good landmark is a bright spot on the cornea produced by the reflection of a light source, the location of which also moves within the image frame in relation to eye rotation. The coordinates related to image pixels must be mapped to physical coordinates. Without this calibration process, the locations of landmarks in the acquired images have no physical meaning. With proper calibration, combined tracking of the pupil and corneal reflection can produce a very precise measurement. Ideally, the physical relationship between the head and the camera should be stable; however, there are head-free video-oculography techniques, which involve compensating for head movement artifacts with complex computational algorithms.

Visual tracking (also known as ocular pursuit) is a combination of smooth pursuit and saccadic eye movements and supports perceptual stability of the object of interest that is in motion. While saccadic tracking, i.e. an alternating sequence of saccades and fixations, allows only for temporally isolated observations of the target, the integration of smooth pursuit in tracking reduces slipping of the image to allow for a continuous observation of the target. Generating and maintaining time-based expectancies of specific sensory information requires allocation of neural resources. It is postulated that attention plays a key role in this process and that various cognitive impairments and symptoms of concussion patients may stem from a primary deficit in this mechanism [3]. Human visual tracking provides an ideal model for studying the ability to predict the temporal course of a stimulus and then dynamically synchronize the required action with it: were we to direct the gaze to where the target was detected, even a short delay could mean that the target had moved away from the location (fig. 2). Although visual sci-

228

Maruta · Ghajar Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Predictive Visual Tracking

The target being processed by the brain The actual target location + gaze

Fig. 2. Prediction and visuomotor synchronization. Maintaining the gaze on the target requires synchronization of the motor output with predicted target motion, rather than the incoming sensory information.

Target motion

ence has extensively exploited the relative constancy of the visual world, scientific inquiry into dynamic foveation, or visuomotor synchronization, has been limited. Predictive timing is critical for offsetting the constraint posed by the sensorimotor processing delay. This capacity prepares a motor outcome to intercept the forthcoming events. For example, predictive timing allows for balance to be maintained when turning a corner and for information to be derived from a continuously changing stimulus such as a visually moving object. In abnormal states of the brain, however, predictive timing may be compromised. We posit that the functional integrity of the predictive timing capacity can be assessed with quantification of predictive visual tracking performance.

The use of a circular target trajectory [4, 5] is exceptionally well suited for assessing the maintenance of predictive visual tracking. With a target traveling at a constant angular velocity with a fixed radius, the motion is highly predictable as it can be described with only two constants. In addition, unlike a one-dimensional tracking paradigm, there is no endpoint; thus, the periodic movement can continue indefinitely within the orbital range of the eye. Visual tracking performance may be quantified with parameters such as smooth pursuit velocity gain, phase error, and root-mean-square error (fig. 3). One metric for visuomotor synchronization is gaze positional error variability [6, 7]. Quantifying performance variability is essential since a dysfunction in predictive timing should increase performance variability. In a circular tracking paradigm, the gaze error variability in the direction parallel (tangential) to the target trajectory has a specific relevance to the precision of predictive timing. Although adult humans have normally developed extraordinary visuomotor coordination skills, there are interindividual variations in visual tracking performance. There is also stability of individuals’ performance over time, implying that interindi-

Detecting Eye Movement Abnormalities from Concussion Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

229

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Performance Assessment of Circular Visual Tracking

SD

Target Eye –4

a

–3

–2

–1

0

1

2

3 4 Direction of target movement

b

Fig. 3. Quantification of visual tracking performance. a Smooth pursuit velocity gain. The target velocity is shown with the smooth black trace, and the eye velocity with the jagged gray trace. Only the vertical components are shown in this example. The spikes in the eye velocity trace represent saccades, and the white trace superimposed on the eye velocity trace represents a sinusoidal fit of the smooth pursuit velocity. The smooth pursuit velocity gain is calculated as the ratio of the amplitudes of modulations of smooth pursuit eye and target velocities. b Position error variability. Top: histogram of instantaneous gaze positional error in the direction parallel (tangential) to the target movement, corresponding to the scattergram below; the variability of the distribution can be described with the standard deviation. Bottom: scattergram of the gaze positional error with the target trajectory straightened (arrow); each dot represents a sample taken at 500 Hz.

vidual variations characterize personal cognitive traits. Indeed, the performance indices of circular visual tracking show good test-retest reliability with intraclass correlation coefficients larger than 0.6 over a 2-week period [7]. Thus, the performance scores and their variations can provide insights into the spectrum of neurological differences and cognitive functioning, and each measurement allows for a single individual to be contrasted with a normative data set. In contrast, when the expected stability of the test outcome is not observed within an individual, a change in the person’s neurological or physiological state may be suspected. The ordinary stability of visual tracking indices hints at the possibility of monitoring of disease or cognitive states by means of comparisons to baseline standards.

Attention impairment is a key symptom of concussion [8, 9]. Although attention can be impaired in other conditions, and thus measures of attention cannot provide a biomarker specific for concussion, measuring and monitoring the attentional states of pa-

230

Maruta · Ghajar Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Concussion

Normal

Patient

tients are important for preventing reinjury and assessing recovery. Circular visual tracking measures are proposed as a biomarker for anticipatory neural activity [6, 7, 10]. Figure 4 illustrates disrupted visuomotor synchronization in a concussion patient indicated by large gaze position error variability and overall phase advancement. This pattern is associated with large anticipatory saccades [6, 7]. It is noteworthy that the gaze should land along the circular path ahead of the target. The behavior demonstrates preserved spatial prediction as well as precision of oculomotor control (as opposed to accuracy), and draws a distinction between the abilities to direct the gaze where the target will eventually be and to direct the gaze there in time for the target’s arrival. Similar behavior can be seen within the normal spectrum, in keeping with the often subtle nature of concussion consequences. Still, the visual tracking performance of a group of acute concussion patients with postconcussive symptoms was found to be worse than that of a normal control group (fig. 5) [11]. In addition, when the patients were retested 1 month after the injury, most showed improvement, contrasting with

Detecting Eye Movement Abnormalities from Concussion Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

231

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Fig. 4. Scattergram of gaze positions relative to the ­target. The coordinates are transformed so that the target is fixed at the 12 o’clock position. The dot-dashed line ­indicates the target trajectory with a 10° radius in visual ­angle. Each dot represents a sample taken at 500 Hz, and those displaced clockwise from the 12 o’clock position represent gaze positions ahead of the target. The ­patient was a 16-year-old ­student athlete who was ­injured 8 days prior to testing. Adapted from [11].

Fig. 5. Changes in a visual tracking index compared against a normative histogram. The change in the scores at the acute stage (⚪) and 1-month follow-up (×) for each concussion subject is indicated by horizontal dot-dashed lines. The concussion subjects are arranged based on their acute stage SD tangential error values. Two subjects dropped out of the study after the initial testing. The solid vertical line running through the histogram indicates the median, and the dotted vertical line separates the worst 5% of the normal population from the rest. Adapted from [11].

Direction of improvement

0

1

2

3

4

5

SD tangential errors (º)

performance stability observed in normal individuals [11]. Thus, both group and interindividual comparisons are useful for monitoring progress after concussion. Many factors, such as fatigue, intoxication, and psychiatric illnesses, can affect attention and interfere with visual tracking performance. However, attention is not a unitary cognitive process. Challenges lie in identifying how physiological, neurological, or psychiatric conditions influence specific components of attention. Different conditions may produce different patterns of changes in visual tracking indices, and these patterns may uniquely identify concussion.

The myriad cognitive impairments and symptoms associated with concussion are postulated to be secondary consequences of a primary deficit in attention, specifically in the ability to synchronize with external events [3, 10, 11]. Sensorimotor synchronization is quantified best by visual tracking of a predictably moving target; thus, we have focused on characterizing eye movements during predictive visual tracking. Experimental results indicate that the visual tracking indices allow for comparison of performance to a normative standard as well as for monitoring within-individual changes. In addition to predictive visual tracking, oculomotor anomalies associated with concussion also include deficits in vergence [12], antisaccade [13], and memoryguided saccades [13]. Further advances in analytics and eye-tracking technology will allow a broad application of oculomotor assessment for concussion and other cognitive impairments.

232

Maruta · Ghajar Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

Conclusion

References   7 Maruta J, Heaton KJ, Kryskow EM, Maule AL, Ghajar J: Dynamic visuomotor synchronization: quantification of predictive timing. Behav Res Methods 2013;45:289–300.   8 Binder LM, Rohling ML, Larrabee GJ: A review of mild head trauma. Part I: meta-analytic review of neuropsychological studies. J Clin Exp Neuropsychol 1997;19:421–431.  9 Bernstein DM: Recovery from mild head injury. Brain Inj 1999;13:151–172. 10 Maruta J, Lee SW, Jacobs EF, Ghajar J: A unified science of concussion. Ann NY Acad Sci 2010; 1208: 58–66. 11 Maruta J, Tong J, Lee SW, Iqbal Z, Schonberger A, Ghajar J: EYE-TRAC: monitoring attention and utility for mTBI. Proc SPIE 2012;8371:83710L-1 12 Thiagarajan P, Ciuffreda KJ, Ludlam DP: Vergence dysfunction in mild traumatic brain injury (mTBI): a review. Ophthalmic Physiol Opt 2011;31:456–468. 13 Heitger MH, Jones RD, Macleod AD, Snell DL, Frampton CM, Anderson TJ: Impaired eye movements in post-concussion syndrome indicate suboptimal brain function beyond the influence of depression, malingering or intellectual ability. Brain 2009; 132:2850–2870.

Jun Maruta, PhD Brain Trauma Foundation 7 World Trade Center, 34th Floor 250 Greenwich Street, New York, NY 10007 (USA) E-Mail [email protected]

Detecting Eye Movement Abnormalities from Concussion Niranjan A, Lunsford LD (eds): Concussion. Prog Neurol Surg. Basel, Karger, 2014, vol 28, pp 226–233 DOI: 10.1159/000358786

233

Downloaded by: Univ. of California Santa Barbara 198.143.33.33 - 7/31/2015 1:05:41 PM

  1 Hughes A: Topographical relationships between the anatomy and physiology of the rabbit visual system. Doc Ophthalmol 1971;30:33–159.   2 Corbetta M, Akbudak E, Conturo TE, Snyder AZ, Ollinger JM, Drury HA, Linenweber MR, Petersen SE, Raichle ME, Van Essen DC, Shulman GL: A common network of functional areas for attention and eye movements. Neuron 1998;21:761–773.   3 Ghajar J, Ivry RB; Cognitive and Neurobiological Research Consortium: The predictive brain state: timing deficiency in traumatic brain injury? Neurorehabil Neural Repair 2008;22:217–227.   4 Umeda Y, Sakata E: The circular eye-tracking test. I. Simultaneous recording of the horizontal and vertical component of eye movement in the eye-tracking test. ORL J Otorhinolaryngol Relat Spec 1975; 37: 290–298.   5 van der Steen J, Tamminga EP, Collewijn H: A comparison of oculomotor pursuit of a target in circular real, beta or sigma motion. Vision Res 1983; 23: 1655–1661.   6 Maruta J, Suh M, Niogi SN, Mukherjee P, Ghajar J: Visual tracking synchronization as a metric for concussion screening. J Head Trauma Rehabil 2010; 25: 293–305.

Detecting eye movement abnormalities from concussion.

An attention-based biomarker may be useful for concussion screening. A key role of attention is to generate time-based expectancies of specific sensor...
589KB Sizes 4 Downloads 2 Views